Showing papers in "Precision Agriculture in 2020"

Economics of robots and automation in field crop production

Abstract: This study reviewed research published after 1990 on the economics of agricultural mechatronic automation and robotics, and identified research gaps. A systematic search was conducted from the following databases: ScienceDirect, Business Source Complete, Wiley, Emerald, CAB Abstract, Greenfile, Food Science Source and AgEcon Search. This identified 4817 documents. The screening of abstracts narrowed the range to a dataset of 119 full text documents. After eligibility assessment, 18 studies were subjected to a qualitative analysis, with ten focused on automation of specific horticultural operations and eight related to autonomous agricultural equipment. All of the studies found some scenarios in which automation and robotic technologies were profitable. Most studies employed partial budgeting considering only costs and revenues directly changed by the introduction of automation or robotics and assuming everything else constant. None examined cropping system changes, or regional and national impacts on markets, trade and labour demand. The review identified a need for in-depth research on the economic implications of the technology. Most of the studies reviewed estimated economic implications assuming that technology design parameters were achieved and/or based on data from prototypes. Data are needed on the benefits and problems with using automation and robotics on farm. All of the studies reviewed were in the context of agriculture in developed countries, but many of the world’s most pressing agricultural problems are in the developing world. Economic and social research is needed to understand those developing country problems, and guide the engineers and scientists creating automation and robotic solutions.

Color-, depth-, and shape-based 3D fruit detection

Abstract: A novel detection algorithm based on color, depth, and shape information is proposed for detecting spherical or cylindrical fruits on plants in natural environments and thus guiding harvesting robots to pick them automatically. A probabilistic image segmentation method is first presented to segment a red–green–blue image as a binary mask. Multiplied by this mask, a filtered depth image is obtained. Region growing, a region-based image segmentation method, is then applied to group the depth image into multiple clusters. Each cluster represents a fruit, leaf, or branch that is later transformed into a point cloud. Next, a 3D shape detection method based on M-estimator sample consensus, a model parameter estimator, is employed to detect potential fruits from each point cloud. Finally, an angle/color/shape-based global point cloud descriptor (GPCD) is developed to extract a feature vector for an entire point cloud, and a support vector machine classifier trained on the GPCD features is used to exclude false positives. Pepper, eggplant, and guava datasets were captured in the field. For the pepper, eggplant, and guava datasets, the detection precision was 0.864, 0.886, and 0.888, and the recall was 0.889, 0.762, and 0.812, respectively. Experiments revealed that the proposed algorithm was universal and robust and hence applicable to an agricultural harvesting robot.

Remote sensing and machine learning for crop water stress determination in various crops: a critical review

Abstract: The remote sensing (RS) technique is less cost- and labour- intensive than ground-based surveys for diverse applications in agriculture. Machine learning (ML), a branch of artificial intelligence (AI), provides an effective approach to construct a model for regression and classification of a multivariate and non-linear system. Without being explicitly programmed, machine learning models learn from training data, i.e., past experience. Machine learning, when applied to remotely sensed data, has the potential to evolve a real-time farm-specific management system to reinforce farmers' ability to make appropriate decisions. Recently, the use of machine learning techniques combined with RS data has reshaped precision agriculture in many ways, such as crop identification, yield prediction and crop water stress assessment, with better accuracy than conventional RS methods. As agriculture accounts for approximately 70% of the worldwide water withdrawals, it must be used in the most efficient way to obtain maximum yields and food production. The use of water management and irrigation based on plant water stress have been demonstrated to not only save water but also increase yield. To date, RS and ML-based results have encouraged farmers and decision-makers to adopt this technology to meet global food demands. This phenomenon has led to the much-needed interest of researchers in using ML to improve agriculture outcomes. However, the use of ML for the potential evaluation of water stress continues to be unexplored and the existing methods can still be greatly improved. This study aims to present an overall review of the widely used methods for crop water stress monitoring using remote sensing and machine learning and focuses on future directions for researchers.

Experience versus expectation: farmers’ perceptions of smart farming technologies for cropping systems across Europe

Abstract: Technological innovations are changing mechanisation in agriculture. The most recent wave of innovations referred to as smart farming technologies (SFT), promise to improve farming by responding to economic, ecological, and social challenges and thereby sustainably develop agriculture throughout Europe. To better understand the relevance of ongoing technological progress for farming systems across Europe, 287 farmers were surveyed in 7 EU countries and in 4 cropping systems, alongside 22 in-depth semi-structured interviews with experts from the agricultural knowledge and innovation system. Of the surveyed farmers, about 50% were SFT adopters and 50% were non-adopters. The number of adopters increased with farm size, and there were more adopters among arable cropping systems than in tree crops. Although all farmers broadly perceive SFT as useful to farming and generally expect SFT to continue to be so, when it comes to specific on-farm challenges, farmers are less convinced of SFT potential. Moreover, farmers’ perceptions of SFT vary according to SFT characteristics and farming context. Interestingly, both adopter and non-adopter groups are hesitant regarding SFT adoption, such that adopters are somewhat disillusioned about the SFT that they have experience with, and non-adopters because they are not convinced that the appropriate technologies are available and accessible. About 60% of all farmers surveyed have a number of suggestions for SFT to become more relevant to a broader range of farms. Both farmers and experts generally consider peer-to-peer communication as important sources of information and deplore a lack of impartial advice. Experts are generally more convinced of SFT advantages, and are positive regarding the long-term trends of technological development. The findings support previous findings on using farmers’ perceptions in innovation processes, and provide insight to the recent trends regarding SFT application to diverse cropping systems across Europe. This suggests that differences related to agricultural structures and farming systems across Europe have to be considered if SFT development and dissemination should be improved.

Fruit detection in natural environment using partial shape matching and probabilistic Hough transform

Abstract: This paper proposes a novel technique for fruit detection in natural environments which is applicable in automatic harvesting robots, yield estimation systems and quality monitoring systems. As most color-based techniques are highly sensitive to illumination changes and low contrasts between fruits and leaves, the proposed technique, conversely, is based on contour information. Firstly, a discriminative shape descriptor is derived to represent geometrical properties of arbitrary fragment, and applied to a bidirectional partial shape matching to detect sub-fragments of interest that match parts of a reference contour. Then, a novel probabilistic Hough transform is developed to aggregate these sub-fragments for obtaining fruit candidates. Finally, all fruit candidates are verified by a support vector machine classifier trained on color and texture features. Citrus, tomato, pumpkin, bitter gourd, towel gourd and mango datasets were provided. Experiments on these datasets demonstrated that the proposed approach was competitive for detecting most type of fruits, such as green, orange, circular and non-circular, in natural environments.

Smartphone adoption and use in agriculture: empirical evidence from Germany

Abstract: Smartphone technology is promising for the future development of agriculture, as it can facilitate and improve many operational procedures and can also be combined with precision agriculture technologies. Yet, existing research on smartphone adoption in agriculture is scarce. Therefore, this paper empirically explores the factors influencing smartphone adoption by German farmers. The relationship between farmers, farm characteristics and smartphone adoption was analysed using a binomial logit model. The dataset, collected in 2016, consists of 817 German farmers and is representative in terms of age, farm size and diversification as well as regional distribution across the study area. The results indicate that, among other factors, farmers’ age, education, and farm size are determinants of smartphone adoption. Furthermore, the paper provides descriptive information about the usage of smartphone functions and agriculture-related app functions. Thus, this paper contributes to the literature by identifying key determinants of smartphone adoption in agriculture. The findings may be of interest for policy makers, researchers in the field of precision agriculture technologies as well as developers and providers of farm equipment and precision agriculture technologies that integrate with smartphones, since the paper includes information concerning smartphone use and key factors influencing smartphone adoption.

Detection of target spot and bacterial spot diseases in tomato using UAV-based and benchtop-based hyperspectral imaging techniques

Abstract: Early and accurate diagnosis is a critical first step in mitigating losses caused by plant diseases. An incorrect diagnosis can lead to improper management decisions, such as selection of the wrong chemical application that could potentially result in further reduced crop health and yield. In tomato, initial disease symptoms may be similar even if caused by different pathogens, for example early lesions of target spot (TS) caused by the fungus Corynespora cassicola and bacterial spot (BS) caused by Xanthomonas perforans. In this study, hyperspectral imaging (380–1020 nm) was utilized in laboratory and field (collected by an unmanned aerial vehicle; UAV) settings to detect both diseases. Tomato leaves were classified into four categories: healthy, asymptomatic, early and late disease development stages. Thirty-five spectral vegetation indices (VIs) were calculated to select an optimum set of indices for disease detection and identification. Two classification methods were utilized: (i) multilayer perceptron neural network (MLP), and (ii) stepwise discriminant analysis (STDA). Best wavebands selection was considered in blue (408–420 nm), red (630–650 nm) and red edge (730–750 nm). The most significant VIs that could distinguish between healthy leaves and diseased leaves were the photochemical reflectance index (PRI) for both diseases, the normalized difference vegetation index (NDVI850) for BS in all stages, and the triangular vegetation index (TVI), NDVI850 and chlorophyll index green (Chl green) for TS asymptomatic, TS early and TS late disease stage respectively. The MLP classification method had an accuracy of 99%, for both BS and TS, under field (UAV-based) and laboratory conditions.

Passion fruit detection and counting based on multiple scale faster R-CNN using RGB-D images

Abstract: The accurate and reliable fruit detection in orchards is one of the most crucial tasks for supporting higher level agriculture tasks such as yield mapping and robotic harvesting. However, detecting and counting small fruit is a very challenging task under variable lighting conditions, low-resolutions and heavy occlusion by neighboring fruits or foliage. To robustly detect small fruits, an improved method is proposed based on multiple scale faster region-based convolutional neural networks (MS-FRCNN) approach using the color and depth images acquired with an RGB-D camera. The architecture of MS-FRCNN is improved to detect lower-level features by incorporating feature maps from shallower convolution feature maps for regions of interest (ROI) pooling. The detection framework consists of three phases. Firstly, multiple scale feature extractors are used to extract low and high features from RGB and depth images respectively. Then, RGB-detector and depth-detector are trained separately using MS-FRCNN. Finally, late fusion methods are explored for combining the RGB and depth detector. The detection framework was demonstrated and evaluated on two datasets that include passion fruit images under variable illumination conditions and occlusion. Compared with the faster R-CNN detector of RGB-D images, the recall, the precision and F1-score of MS-FRCNN method increased from 0.922 to 0.962, 0.850 to 0.931 and 0.885 to 0.946, respectively. Furthermore, the MS-FRCNN method effectively improves small passion fruit detection by achieving 0.909 of the F1 score. It is concluded that the detector based on MS-FRCNN can be applied practically in the actual orchard environment.

Spray and economics assessment of a UAV-based ultra-low-volume application in olive and citrus orchards

Abstract: Automation is a new frontier in specialty agriculture equipment. Specifically, unmanned aerial vehicles (UAV), machine vision and robotics will increasingly appear in sustainable agricultural systems. The use of small UAVs retrofitted with spraying systems allows precision aerial applications on small targets. These precision applications can result in significant cost savings and reductions in risk to operators during treatments. This paper presents a novel and practical design and development of a small application system capable of being mounted on an unmanned aerial vehicle for agrochemical spraying tasks and an analysis of the quality of the application and economic costs in olive and citrus orchards compared with those of a conventional treatment. Once the equipment had been developed, field trials in super-high-density olive and citrus orchards were undertaken to evaluate the spray deposition efficiency. For comparison with a conventional hydro-pneumatic sprayer, the field tests took into account parameters such as the applied volume rate, spray drift, application time and equipment costs and depreciation. The results obtained indicate that there was a 7 €/ha difference in the application costs between the aerial vehicle and conventional equipment. It is hoped that the conclusions of this work will serve as the basis for a debate about the existing legislation governing this type of aerial work, which can be beneficial in specific cases and should be carried out in a safe and legal manner.

Leaf Area Index evaluation in vineyards using 3D point clouds from UAV imagery

Abstract: The Leaf Area Index (LAI) is an ecophysiology key parameter characterising the canopy-atmosphere interface where most of the energy fluxes are exchanged. However, producing maps for managing the spatial and temporal variability of LAI in large croplands with traditional techniques is typically laborious and expensive. The objective of this paper is to evaluate the reliability of LAI estimation by processing dense 3D point clouds as a cost-effective alternative to traditional LAI assessments. This would allow for high resolution, extensive and fast mapping of the index, even in hilly and not easily accessible regions. In this setting, the 3D point clouds were generated from UAV-based multispectral imagery and processed by using an innovative methodology presented here. The LAI was estimated by a multivariate linear regression model using crop canopy descriptors derived from the 3D point cloud, which account for canopy thickness, height and leaf density distribution along the wall. For the validation of the estimated LAI, an experiment was conducted in a vineyard in Piedmont: the leaf area of 704 vines was manually measured by the inclined point quadrant approach and six UAV flights were contextually performed to acquire the aerial images. The vineyard LAI estimated by the proposed methodology showed to be correlated with the ones obtained by the traditional manual method. Indeed, the obtained R2 value of 0.82 can be considered fully adequate, compatible to the accuracy of the reference LAI manual measurement.

Assessment of maize yield and phenology by drone-mounted superspectral camera

Abstract: The capability of unmanned aerial vehicle (UAV) spectral imagery to assess maize yield under full and deficit irrigation is demonstrated by a Tetracam MiniMCA12 11 bands camera. The MiniMCA12 was used to image an experimental field of 19 maize hybrids. Yield prediction models were explored for different maize development stages, with the best model found using maize plant development stage reproductive 2 (R2) for both maize grain yield and ear weight (respective R2 values of 0.73 and 0.49, and root mean square error of validation (RMSEV) values of 2.07 and 3.41 metric tons per hectare using partial least squares regression (PLS-R) validation models). Models using vegetation indices for inputs rather than superspectral data showed similar R2 but higher RMSEV values, and produced best results for the R4 development stage. In addition to being able to predict yield, spectral models were able to distinguish between different development stages and irrigation treatments. These abilities potentially allow for yield prediction of maize plants whose development stage and water status are unknown.

Detection and counting of flowers on apple trees for better chemical thinning decisions

Abstract: Accurate chemical thinning of apple trees requires estimation of their blooming intensity, and determination of the blooming peak date. Performing this task, as of today, requires human experts to be present in the orchards for the entire blossom period or extrapolate using a single observation. Since experts are rare and in high demand, there is a need to automate this process. The system presented in this paper is able to estimate the blooming intensity and the blooming peak date from a sequence of tree images, with close-to-human accuracy. For this purpose, a two years dataset was collected in 2014–2015, partially tagged for the flowers location and completely annotated for blooming intensity. Using this dataset, an algorithm was developed and trained with three stages: a visual flower detector based on a deep convolutional neural network, followed by a blooming level estimator, and a peak blooming day finding algorithm. Despite the challenging conditions, the trained detector was able to detect flowers on trees with an Average Precision (AP) score of 0.68, which is on a par with contemporary results of other objects in detection benchmarks. The blooming estimator was based on a linear regression component, which used the number of flowers detected and related statistics to estimate the blooming intensity. The Pearson correlation between the algorithm blooming estimation and human judgments of several experts indicated high agreement levels (0.78–0.93) which were similar to the correlations measured among the human experts. Moreover, the developed estimator was relatively stable across multiple years. The developed peak date finding algorithm identified correctly the orchard’s blooming peak date, which was used to determine the thinning date in the current practice (the entire orchard is thinned in the same day). Experiments testing the algorithm’s ability to find a blooming peak date for each tree independently showed encouraging results, which may lead upon refinement to a more precise practice for tree-specific thinning.

Vineyard yield estimation by combining remote sensing, computer vision and artificial neural network techniques

Abstract: In viticulture, it is critical to predict productivity levels of the different vineyard zones to undertake appropriate cropping practices. To overcome this challenge, the final yield was predicted by combining vegetation indices (VIs) to sense the health status of the crop and by computer vision to obtain the vegetated fraction cover (Fc) as a measure of plant vigour. Multispectral imagery obtained from an unmanned aerial vehicle (UAV) is used to obtain VIs and Fc, which are used together with artificial neural networks (ANN) to model the relationship between VIs, Fc and yield. The proposed methodology was applied in a vineyard, where different irrigation and fertilisation doses were applied. The results showed that using computer vision techniques to differentiate between canopy and soil is necessary in precision viticulture to obtain accurate results. In addition, the combination of VIs (reflectance approach) and Fc (geometric approach) to predict vineyard yield results in higher accuracy (root mean square error (RMSE) = 0.9 kg vine−1 and relative error (RE) = 21.8% for the image when close to harvest) compared to the simple use of VIs (RMSE = 1.2 kg vine−1 and RE = 28.7%). The implementation of machine learning techniques resulted in much more accurate results than linear models (RMSE = 0.5 kg vine−1 and RE = 12.1%). More precise yield predictions were obtained when images were taken close to the harvest date, although promising results were obtained at earlier stages. Given the perennial nature of grapevines and the multiple environmental and endogenous factors determining yield, seasonal calibration for yield prediction is required.

Automated crop plant counting from very high-resolution aerial imagery

Abstract: Knowing before harvesting how many plants have emerged and how they are growing is key in optimizing labour and efficient use of resources. Unmanned aerial vehicles (UAV) are a useful tool for fast and cost efficient data acquisition. However, imagery need to be converted into operational spatial products that can be further used by crop producers to have insight in the spatial distribution of the number of plants in the field. In this research, an automated method for counting plants from very high-resolution UAV imagery is addressed. The proposed method uses machine vision—Excess Green Index and Otsu’s method—and transfer learning using convolutional neural networks to identify and count plants. The integrated methods have been implemented to count 10 weeks old spinach plants in an experimental field with a surface area of 3.2 ha. Validation data of plant counts were available for 1/8 of the surface area. The results showed that the proposed methodology can count plants with an accuracy of 95% for a spatial resolution of 8 mm/pixel in an area up to 172 m2. Moreover, when the spatial resolution decreases with 50%, the maximum additional counting error achieved is 0.7%. Finally, a total amount of 170 000 plants in an area of 3.5 ha with an error of 42.5% was computed. The study shows that it is feasible to count individual plants using UAV-based off-the-shelf products and that via machine vision/learning algorithms it is possible to translate image data in non-expert practical information.

Status quo of adoption of precision agriculture enabling technologies in Swiss plant production

Abstract: This paper presents the state of application of Precision Agricultural enabling Technology (PAT) in Swiss farms as an example for small-scale, highly mechanised Central European agriculture. Furthermore, correlations between farm and farmers’ characteristics and technology adoption were evaluated. Being part of a comprehensive and representative study assessing the state of mechanisation and automation in Swiss agriculture, this paper focuses on the adoption of Driver Assistance Systems (DAS) and activities in which Electronic Measuring Systems (EMS) are used. The adoption rate of DAS was markedly higher compared to EMS in all agricultural enterprises. The adoption rate was highest for high-value enterprise vegetables and surprisingly low for the high-value enterprise grapes. The results of a binary logistic regression showed that farmers located in the mountain zone were less likely to adopt PAT compared to farmers in the valley. Small farm size correlated with low adoption rates and vice versa showing adoption happens country-specific in the upper farm size distribution. The results show the potential for novel technologies to be adopted by farmers of high-value products. Furthermore, technologies have been partially used to reduce physical workload but not yet to evaluate crop or management performance to support decisions. However, automatic collection and forwarding of data is a fundamental step towards Smart Farming realizing its full potential in the future.

Estimation and mapping of nitrogen content in apple trees at leaf and canopy levels using hyperspectral imaging

Abstract: Accurate and rapid diagnosis of nitrogen status in fruit trees on an individual tree basis is a prerequisite for precision orchard nutrient management. This study presents a rapid and non-destructive approach for estimation and mapping of nitrogen content in apple trees at both leaf and canopy levels. An ImSpector V10 system was used to collect hyperspectral images (400–1000 nm) for both apple leaves and canopies. Nitrogen content in apple leaves was measured by Vario EL cube. Raw reflectance and first derivative reflectance were used to relate to leaf nitrogen content. Partial least squares (PLS) regression and multiple linear regression (MLR) analyses were performed to estimate nitrogen content from reflectance. The results showed that both PLS and MLR models achieved reasonable predictive accuracy (PLS and MLR models based on raw reflectance: R2 = 0.7728 and 0.7843 (p < 0.001); PLS and MLR models based on first derivative reflectance: R2 = 0.7745 and 0.774 (p < 0.001)). However, the MLR model based on raw reflectance demonstrated its advantage over the PLS models as well as the MLR model based on first derivative reflectance, because it only used 4 key wavelengths (505, 560, 675 and 705 nm) while the other models were based on either the full wavelengths (132 wavelengths) or more narrowband wavelengths adjacent to the selected key wavelengths. Furthermore, nitrogen distribution maps at both leaf and canopy levels were generated based on the nitrogen contents estimated by the MLR model based on raw reflectance. This new approach may be potentially applied to precision apple orchard nutrient management.

Understanding the adoption of smartphone apps in crop protection

Abstract: There is a steady increase in smartphone apps available to improve farmers’ decision making with respect to crop protection. While current studies have focused on smartphone adoption in general and farmers’ general willingness to pay for crop protection smartphone apps in particular, none have focused on the initial adoption decision. Furthermore, it has not been studied yet which app functions are perceived as useful and which are actually used by farmers. Based on an online survey conducted in 2019 with 207 German farmers, this study investigated latent factors affecting farmers’ adoption decision for crop protection smartphone apps based on the Unified Theory of Acceptance and Use of Technology (UTAUT) framework applying partial least squares equation modelling and a binary logit model. Descriptive results show that 95% of the surveyed farmers use a smartphone, but only 71% use a crop protection smartphone app. Apps providing information about weather, pest scouting and infestations forecasts are perceived as most useful by the majority of farmers. However, reported use fell short of reported usefulness. With respect to the model for the UTAUT, 73% of the variation in the behavioral intention to use a crop protection smartphone app is explained by the model. The results are of interest for policy makers in the field of digitization in agriculture as well as providers and developers of crop protection smartphone apps since the results could be used for further development of apps and policies regarding digitization.

Brown rice planthopper ( Nilaparvata lugens Stal) detection based on deep learning

Abstract: The brown rice planthopper (Nilaparvata lugens Stal) is one of the main pests of rice. The rapid and accurate detection of brown rice planthoppers (BRPH) can help treat rice in time. Due to the small size, large number and complex background of BRPHs, image detection of them is challenging. In this paper, a two-layer detection algorithm based on deep learning technology is proposed to detect them. The algorithm for both layers is the Faster RCNN (regions with CNN features). To effectively utilize the computing resources, different feature extraction networks have been selected for each layer. In addition, the second layer detection network was optimized to improve the final detection performance. The detection results of the two-layer detection algorithm were compared with the detection results of the single-layer detection algorithm. The detection results of the two-layer detection algorithm for detecting different populations and numbers of BRPHs were tested, and the test results were compared with YOLO v3, a deep learning target detection network. The test results show that the detection results of the two-layer detection algorithm were significantly better than those of the single-layer detection algorithm. In the tests for different numbers of BRPHs, the average recall rate of this algorithm was 81.92%, and the average accuracy was 94.64%; meanwhile, the average recall rate of YOLO v3 was 57.12%, and the average accuracy rate was 97.36%. In the experiment with different ages of BRPHs, the average recall rate of the algorithm was 87.67%, and the average accuracy rate was 92.92%. In comparison, for the YOLO v3, the average recall rate was 49.60%, and the average accuracy rate was 96.48%.

Monitoring leaf potassium content using hyperspectral vegetation indices in rice leaves

Abstract: Potassium (K) is one of three main crop nutrients, and the high rate of potash fertilizer utilization (second only to nitrogen) leads to high prices. Therefore, efficient application, as well as rapid and time monitoring of K in crops is essential. Several turnover box and field experiments were conducted across multiple years and cultivation factors (i.e., potassium levels and plant varieties) yielding 340 groups of leaf samples with different K contents; these samples were used to examine the relationship between reflectance spectra (350–2500 nm) and leaf K content (LKC). The correlation between LKC and the two-band spectral indices computed with random two bands from 350 to 2500 nm were determined for the published K vegetation indices in rice. Results showed that the spectral reflectance, R, of the shortwave infrared (1300–2000 nm) region was sensitive to the K levels and significantly correlated with rice LKC. New shortwave infrared two-band spectral indices, Normalized difference spectral index [NDSI (R1705, R1385)], Ratio spectral index [RSI (R1385, R1705)], and Difference spectral index [DSI (R1705, R1385)], showed good correlations with LKC (R2 up to 0.68). Moreover, the three-band spectral indices (R1705 − R700)/(R1385 − R700) and (R1705 − R1385)/(R1705 + R1385 − 2 × R704) were developed by adding red edge bands to improve accuracy. Three-band spectral indices had an improved prediction accuracy for rice LKC (R2 up to 0.74). However, several previously published K-sensitive vegetation indices did not yield good results in this study. Validation with independent samples showed that the indices (R1705 − R700)/(R1385 − R700) and (R1705 − R1385)/(R1705 + R1385 − 2 × R704) had higher accuracies and stabilities than two-band indices and are suitable for quantitatively estimating rice LKC. The widescale application of these proposed vegetation indices in this paper still needs to be verified in different environmental conditions. This study provides a technical basis for LKC monitoring using spectral remote sensing in rice.

Geostatistical and fuzzy clustering approach for delineation of site-specific management zones and yield-limiting factors in irrigated hot arid environment of India

Abstract: Delineation of management zones (MZs) are needed to manage fields in order to maximize economic return, minimize environmental impact, and improve soil and crop management. The MZs of uniform production potential may offer an effective solution to nutrient management. In this study, a total of 122 geo-referenced representative surface (0–250 mm depth) soil samples were collected from the study area covering an area of 6296 ha. Soil samples were analysed for pH, EC, CaCO3, organic carbon (SOC), available nitrogen (AN), available phosphorus (AP), available potassium (AK) and micronutrients (Fe, Mn, Zn and Cu). Their spatial variability was analyzed and spatial distribution maps were constructed using geostatistical techniques. Geostatistical analysis showed that exponential, rational quadratic, tetraspherical, pentaspherical and circular models were the best-fit models for soil properties and available nutrients. Further, geographical weighted principal component analysis (GWPCA) and possibilistic fuzzy C-means (PFCM) clustering algorithm were carried out to delineate the management zones based on optimum clusters identified using fuzzy performance index (FPI) and normalized classification entropy (NCE). The results revealed that the optimum number of MZs for this study area was four and there was heterogeneity in soil nutrients in four MZs. The study indicated that MZ-based soil test crop response recommendation reduces the application quantity of fertilizer significantly at a large extent. Therefore, the management zone concept can reduce agricultural inputs and environmental pollution, and maximize crop production.

Techno-economic impacts of using a laser-guided variable-rate spraying system to retrofit conventional constant-rate sprayers

Abstract: Specialty crops, such as apples, are vulnerable to insects and pathogens, and require higher pesticide input than row crops, a significant fraction of which is off-target loss, causing adverse environmental and socio-economic impacts. An advanced laser-guided variable-rate sprayer (VRS) could improve spray deposition uniformity and minimize pesticide waste, while maintaining efficacy against insects and pathogens. Despite these merits, retrofitting a conventional sprayer with laser-guided variable-rate spraying functions adds to its cost. Thus, the objective of this study was to analyze the techno-economics of a conventional pesticide sprayer retrofitted with VRS, in comparison to a conventional constant-rate sprayer (CRS) for pesticide application during apple production. A techno-economic model was developed for the apple orchards covering areas of 4 and 20 ha, which are common orchard sizes in the USA. The model incorporated cost for operation, equipment, fuel use and labor during pesticide application. The data were obtained from field tests in orchards in Ohio, USA in years 2016 and 2017, literature, and the original VRS development team at USDA-ARS and Ohio State University. The results indicated that VRS can reduce pesticide costs by 60–67%, pesticide application time by 27–32% and labor and fuel by 28% compared to CRS. For larger orchards, VRS also reduced equipment requirement. Compared to CRS, overall annual pesticide application cost savings by using VRS were between $1420 and $1750 ha−1. The payback time for using VRS was estimated to be between 1.1 and 3.8 years for apple orchards between 4 and 20 ha, respectively, in Ohio.

Vegetation detection and discrimination within vegetable plasticulture row-middles using a convolutional neural network

Abstract: Weed control between plastic covered, raised beds in Florida vegetable crops relies predominantly on herbicides. Broadcast applications of post-emergence herbicides are unnecessary due to the general patchy distribution of weed populations. Development of precision herbicide sprayers to apply herbicides where weeds occur would result in input reductions. The objective of the study was to test a state-of-the-art object detection convolutional neural network, You Only Look Once 3 (YOLOV3), to detect vegetation both indiscriminately (1-class network) and to detect and discriminate three classes of vegetation commonly found within Florida vegetable plasticulture row-middles (3-class network). Vegetation was discriminated into three categories: broadleaves, sedges and grasses. The 3-class network (Fscore = 0.95) outperformed the 1-class network (Fscore = 0.93) in overall vegetation detection. The increase in target variability when combining classes increased and potentially negated benefits from pooling classes into a single target (and increasing the available data per class). The 3-class network Fscores for grasses, sedges and broadleaves were 0.96, 0.96 and 0.93 respectively. Recall was the limiting factor for all classes. With consideration to how much of the plant was identified (broadleaves and grasses), the 3-class network (Fscore = 0.93) outperformed the 1-class network (Fscore = 0.79). The 1-class network struggled to detect grassy weed species (recall = 0.59). Use of YOLOV3 as an object detector for discrimination of vegetation classes is a feasible option for incorporation into precision applicators.

Comparison between vegetation indices for detecting spatial and temporal variabilities in soybean crop using canopy sensors

Abstract: Crop monitoring through remote sensing techniques enable greater knowledge of average variability in crop growth. Canopy sensors help provide information on the variability of crop through the use of vegetation indices. The objective of this work was to compare the potential and performance of three vegetation indices used for monitoring soybean variability with canopy sensors was compared. The optimal time for sensor readings was determined during the soybean crop development stages. Also, the quality of the readings between vegetation indices [the normalized difference vegetation index (NDVI), normalized difference red-edge (NDRE), and inverse ratio (IRVI)] was compared through control charts and the saturation detection index. The experimental design was based on statistical quality control and comprised 65 sampling points within a 30 × 30 m grid. At 30, 45, 60, 75, and 90 days after sowing (DAS), the parameters used as quality indicators, such as fresh and dry biomass, canopy width, chlorophyll index, plant height, yield, and the vegetation indices were assessed using canopy sensors. The optimal time for canopy sensor readings, based mainly on the NDRE, was at 45 and 60 DAS. The lower variability exhibited by NDRE led to higher process quality when compared with those for NDVI and IRVI. The control charts proved to be promising in identifying the moment when saturation occurs for the indices more susceptible to saturation, such as the NDVI.

An artificial neural network model for estimating Mentha crop biomass yield using Landsat 8 OLI

Abstract: Yield forecasting is essential for management of the food and agriculture economic growth of a country. Artificial Neural Network (ANN) based models have been used widely to make precise and realistic forecasts, especially for the nonlinear and complicated problems like crop yield prediction, biomass change detection and crop evapo-transpiration examination. In the present study, various parameters viz. spectral bands of Landsat 8 OLI (Operational Land Imager) satellite data and derived spectral indices along with field inventory data were evaluated for Mentha crop biomass estimation using ANN technique of Multilayer Perceptron. The estimated biomass showed a good relationship (R2 = 0.762 and root mean square error (RMSE) = 2.74 t/ha) with field-measured biomass.

Integrating machine vision-based row guidance with GPS and compass-based routing to achieve autonomous navigation for a rice field weeding robot

Abstract: Autonomous weeding robots are a productive and more sustainable solution over traditional, labor-intensive weed control practices such as chemical weeding that are harmful to the environment when used excessively. To achieve a fully autonomous weed control operation, the robots need to be precisely guided through the crop rows without damaging rice plants and they should be able to detect the end of the crop row and make turns to change rows. This research attempted to integrate GNSS, compass and machine vision into a rice field weeding robot to achieve fully autonomous navigation for the weeding operation. A novel crop row detection algorithm was developed to extract the four immediate rows spanned by a camera mounted at the front of the robot. The extracted rows were used to determine a guide-line that was used to precisely maneuver the robot along the crop rows with an accuracy of less than a hundred millimeters in variable circumstances such as weed intensity, growth stage of plants and weather conditions. The GNSS and compass were used for locating the robot within the field. A state-based control system was implemented to integrate the GNSS, compass and vision guidance to efficiently navigate the weeding robot through a pre-determined route that covers the entire field without damaging rice plants. The proposed system was experimentally determined to deliver good performance in low weed concentrations with an accuracy of less than 2.5° in heading compensation and an average deviation from the ideal path of 45.9 mm. Though this accuracy dropped when the weed concentration increased, the system was still able to navigate the robot without inflicting any serious damage to the plants.

Mapping within-field leaf chlorophyll content in agricultural crops for nitrogen management using Landsat-8 imagery

Abstract: Spatial information on crop nutrient status is central for monitoring vegetation health, plant productivity and managing nutrient optimization programs in agricultural systems. This study maps the spatial variability of leaf chlorophyll content within fields with differing quantities of nitrogen fertilizer application, using multispectral Landsat-8 OLI data (30 m). Leaf chlorophyll content and leaf area index measurements were collected at 15 wheat (Triticum aestivum) sites and 13 corn (Zea mays) sites approximately every 10 days during the growing season between May and September 2013 near Stratford, Ontario. Of the 28 sites, 9 sites were within controlled areas of zero nitrogen fertilizer application. Hyperspectral leaf reflectance measurements were also sampled using an Analytical Spectral Devices FieldSpecPro spectroradiometer (400–2500 nm). A two-step inversion process was developed to estimate leaf chlorophyll content from Landsat-8 satellite data at the sub-field scale, using linked canopy and leaf radiative transfer models. Firstly, at the leaf-level, leaf chlorophyll content was modelled using the PROSPECT model, using both hyperspectral and simulated mulitspectral Landsat-8 bands from the same leaf sample. Hyperspectral and multispectral validation results were both strong (R2 = 0.79, RMSE = 13.62 μg/cm2 and R2 = 0.81, RMSE = 9.45 μg/cm2, respectively). Secondly, leaf chlorophyll content was estimated from Landsat-8 satellite imagery for 7 dates within the growing season, using PROSPECT linked to the 4-Scale canopy model. The Landsat-8 derived estimates of leaf chlorophyll content demonstrated a strong relationship with measured leaf chlorophyll values (R2 = 0.64, RMSE = 16.18 μg/cm2), and compared favourably to correlations between leaf chlorophyll and the best performing tested spectral vegetation index (Green Normalised Difference Vegetation Index, GNDVI; R2 = 0.59). This research provides an operational basis for modelling within-field variations in leaf chlorophyll content as an indicator of plant nitrogen stress, using a physically-based modelling approach, and opens up the possibility of exploiting a wealth of multispectral satellite data and UAV-mounted multispectral imaging systems.

Predicting plant available phosphorus using infrared spectroscopy with consideration for future mobile sensing applications in precision farming

Abstract: Phosphorus (P) fertilisation recommendations rely primarily on soil content of plant available P (Pavl) that vary spatially within farm fields. Spatially optimized P fertilisation for precision farming requires reliable, rapid and non-invasive Pavl determination. This laboratory study aimed to test and to compare visible-near infrared (Vis–NIR) and mid-infrared (MIR) spectroscopy for Pavl prediction with emphasis on future application in precision agriculture. After calibration with the conventional calcium acetate lactate (CAL) extraction method, limitations of Vis–NIRS and MIRS to predict Pavl were evaluated in loess topsoil samples from different fields at six localities. Overall calibration with 477 (Vis–NIRS) and 586 (MIRS) samples yielded satisfactory model performance (R2 0.70 and 0.72; RPD 1.8 and 1.9, respectively). Local Vis–NIRS models yielded better results with R2 up to 0.93 and RPD up to 3.8. For MIRS, results were comparable. However, an overall model to predict Pavl on independent test data partly failed. Sampling date, pre-crop harvest residues and fertilising regime affected model transferability. Varying transferability could partly be explained after deriving the cellulose absorption index from the Vis–NIR spectra. In 62 (Vis–NIRS) and 67% (MIRS) of all samples, prediction matched the correct Pavl content class. Rapid discrimination between high, optimal and low P classes could be carried out on many samples from single fields thus marking an improvement over the common practice. However, Pavl determination by means of IR spectroscopy is not yet satisfactory for determination of precision fertilizer dosage. For introduction into agricultural practice, a standardized sampling protocol is recommended to help achieve reliable spectroscopic Pavl prediction.

Biophysical parameters of coffee crop estimated by UAV RGB images

Abstract: The advance of digital agriculture combined with computational tools and Unmanned Aerial Vehicles (UAVs) has enabled the collection of data for reliably extracting vegetation indices and biophysical parameters derived from the Structure from Motion (SfM) algorithm. This work aimed to evaluate the accuracy of the photogrammetry technique using an SfM point cloud for the estimation of the height (h) and crown diameter (d) of coffee trees from aerial images obtained by UAV with an RGB (Red, Green, Blue) camera and compared the results with data measured in situ for 12 months. The experiment was carried out in a coffee plantation, Lavras, Minas Gerais, Brazil. A rotary-wing UAV was used in autonomous flight mode and coupled to a conventional camera, flying at a height of 30 m with an image overlap of 80% and a speed of 3 m/s. The images were processed using PhotoScan software, and the analyses were performed in Qgis. A correlation of 87% was obtained between the h values in the field and h values obtained by the UAV, and there was a 95% correlation between the d values obtained in the field and the values obtained by the UAV. It was possible to obtain significant estimates of the attributes, such as the h and d of coffee trees, using UAV–SfM images acquired with an RGB digital camera.

Row-crop planter performance to support variable-rate seeding of maize

Abstract: Current planting technology possesses the ability to increase crop productivity and improve field efficiency by precisely metering and placing crop seeds. Planter performance depends on determining and utilizing optimal settings for different planting variables such as seed depth, down pressure, and seed metering unit. The evolution of “Big Data” in agriculture today brings focus on the need for quality as-planted and yield mapping data. Therefore, an investigation was conducted to evaluate the performance of current planting technology for accurate placement of seeds while understanding the accuracy of as-planted data. Two studies consisting of two different setups on a 6-row, John Deere planter for seeding of maize (Zea mays L.) were conducted. The first study aimed at assessing planter performance at 2 depth settings (25 and 51 mm) and four different down pressure settings (varying from none to high), while the second study focused on evaluating planter performance during variable-rate seeding with treatments consisting of two seed metering units (John Deere Standard and Precision Planting’s eSet setups) with five different seeding rates and four ground speed treatments which provided a combination of 20 different meter speeds. Field data collection consisted of measuring plant emergence, plant population and seed depth whereas plant spacing, plant population after emergence along with distance and location for rate changes within the field were also recorded for the variable-rate seeding study. Results indicated that both depth setting and downforce affected final seeding depth. Measured seed depth was significantly different from the target depth even though time was spent adjusting the units to achieve the desired prior to planting. Crop emergence did not vary significantly for the different depth and downforce settings except for target depth in Field 1. Results from the variable-rate study indicated that seeding rate changes were accomplished within a quick response time (< 1 s) at all ground speeds regardless of magnitude of rate change. Data showed that planter performance in terms of emergence and plant spacing CV was comparable for most of the meter speeds (17.4–33.5 rpm) among the two seed meters utilized in the study. Plant spacing CV increased with an increase in meter speed, however no significant differences existed among meter speeds in the range of 17.4–33.5 rpm. Results implied that correct seed metering unit setup is very critical to obtain expected performance of today’s planting technology. A concerning find was that the quality of as-applied maps from the commercial variable-rate display was not reflective of the actual planter performance in the field. The study recommended that operators need to ensure the correct planter and display setups in order to achieve needed seed placement performance to support variable-rate seeding.

Examining the social and biophysical determinants of U.S. Midwestern corn farmers’ adoption of precision agriculture

Abstract: Precision agricultural technologies (PA) such as global positioning system tools have been commercially available since the early 1990s and they are widely thought to have environmental and economic benefit; however, adoption studies show uneven adoption among farmers in the U.S. and Europe. This study aims to tackle a lingering puzzle regarding why some farmers adopt precision agriculture as an approach to food production and why others do not. The specific objective of this study is to examine the social and biophysical determinants of farmers’ adoption of PA. This paper fills a research gap by including measurements of farmer identity—specifically their own conceptions of their role in the food system—as well as their perceptions of biophysical risks as these relate to the adoption of PA among a large sample of Midwestern U.S. farmers. The study has identified that farmer identity and perceptions of environmental risk do indeed influence PA adoption and that these considerations ought to be incorporated into further studies of PA adoption in other jurisdictions. The findings also appear to highlight the social force of policy and industry efforts to frame PA as not only good for productivity and efficiency but also as an ecologically beneficial technology.