Above-ground biomass (AGB) is a basic agronomic parameter for field investigation and is frequently used to indicate crop growth status, the effects of agricultural management practices, and the ability to sequester carbon above and below ground. The conventional way to obtain AGB is to use destructive sampling methods that require manual harvesting of crops, weighing, and recording, which makes large-area, long-term measurements challenging and time consuming. However, with the diversity of platforms and sensors and the improvements in spatial and spectral resolution, remote sensing is now regarded as the best technical means for monitoring and estimating AGB over large areas. In this study, we used structural and spectral information provided by remote sensing from an unmanned aerial vehicle (UAV) in combination with machine learning to estimate maize biomass. Of the 14 predictor variables, six were selected to create a model by using a recursive feature elimination algorithm. Four machine-learning regression algorithms (multiple linear regression, support vector machine, artificial neural network, and random forest) were evaluated and compared to create a suitable model, following which we tested whether the two sampling methods influence the training model. To estimate the AGB of maize, we propose an improved method for extracting plant height from UAV images and a volumetric indicator (i.e., BIOVP). The results show that (1) the random forest model gave the most balanced results, with low error and a high ratio of the explained variance for both the training set and the test set. (2) BIOVP can retain the largest strength effect on the AGB estimate in four different machine learning models by using importance analysis of predictors. (3) Comparing the plant heights calculated by the three methods with manual ground-based measurements shows that the proposed method increased the ratio of the explained variance and reduced errors. These results lead us to conclude that the combination of machine learning with UAV remote sensing is a promising alternative for estimating AGB. This work suggests that structural and spectral information can be considered simultaneously rather than separately when estimating biophysical crop parameters.

/pdf/modeling-maize-above-ground-biomass-based-on-machine-2jgr7spham.pdf

Modeling maize above-ground biomass based on machine learning approaches using UAV remote-sensing data

Monitoring hourly night-time light by an unmanned aerial vehicle and its implications to satellite remote sensing

In order to overcome the insufficiency of single remote sensing data in change detection, synthetic aperture radar (SAR) and optical image data can be used together for supplementation. However, conventional image fusion methods fail to address the differences in imaging mechanisms and cannot overcome some practical limitations such as usage in change detection or temporal requirement of the optical image. This study proposes a new method to fuse SAR and optical images, which is expected to be visually helpful and minimize the differences between two imaging mechanisms. The algorithm performs the fusion by establishing relationships between SAR and multispectral (MS) images by using a random forest (RF) regression, which creates a fused SAR image containing the surface roughness characteristics of the SAR image and the spectral characteristics of the MS image. The fused SAR image is evaluated by comparing it to those obtained using conventional image fusion methods and the proposed method shows that the spectral qualities and spatial qualities are improved significantly. Furthermore, for verification, other ensemble approaches such as stochastic gradient boosting regression and adaptive boosting regression are compared and overall it is confirmed that the performance of RF regression is superior. Then, change detection between the fused SAR and MS images is performed and compared with the results of change detection between MS images and between SAR images and the result using fused SAR images is similar to the result with MS images and is improved when compared to the result between SAR images. Lastly, the proposed method is confirmed to be applicable to change detection.

https://www.mdpi.com/2220-9964/7/10/401/pdf

Fusion of SAR and Multispectral Images Using Random Forest Regression for Change Detection

Quantifying the characteristic of urban expansion is the prerequisite of understanding its social and ecological outcomes. Here, using Shenzhen, a city that grew from a fishery village to a megacity in 40 years as a case study, we quantified the urban expansion at two scales, over the entire city and within the city, at district level. Twelve land cover maps from 1979 to 2017 were classified using Landsat TM/ETM+/OLI images. Based on the long-term map series, we examined the impacts of urban expansion on land cover change, and the directional variations of those change trajectories. Our results demonstrated that: (1) Shenzhen experienced rapid and dramatic urban expansion from 1979 to 2017. The developed land increased by 3400%, from 27 km2 to 946 km2 with an average annual growth of 24 km2. While the developed land continued to increase, its expansion rate exhibited a general decreasing trend, except for the time period of 1990–1995. (2) The urban expansion primarily consumed forest, water, and farmland. Forest experienced a 44% loss from 1979 to 2017, equals 659 km2 of land. (3) The urban expansion exhibited distinctive directional variations from the southwest to the northeast, following a “fan” model. The within city differences are potentially related to changes of regional geographical conditions. In the western coastal area, water experienced major changes, and in the northern region forest had the most lost. The understanding of urban expansion variations within the city will help local government in designing and implementing effective plans that balancing future urban development and ecological conservation.

Urban expansion in Shenzhen since 1970s: A retrospect of change from a village to a megacity from the space

Multi-sensor fusion using random forests for daily fractional snow cover at 30 m

Most bamboo forests grow in humid climates in low-latitude tropical or subtropical monsoon areas, and they are generally located in hilly areas. Bamboo trunks are very straight and smooth, which means that bamboo forests have low structural diversity. These features are beneficial to synthetic aperture radar (SAR) microwave penetration and they provide special information in SAR imagery. However, some factors (e.g., foreshortening) can compromise the interpretation of SAR imagery. The fusion of SAR and optical imagery is considered an effective method with which to obtain information on ground objects. However, most relevant research has been based on two types of remote sensing image. This paper proposes a new fusion scheme, which combines three types of image simultaneously, based on two fusion methods: bidimensional empirical mode decomposition (BEMD) and the Gram-Schmidt transform. The fusion of panchromatic and multispectral images based on the Gram-Schmidt transform can enhance spatial resolution while retaining multispectral information. BEMD is an adaptive decomposition method that has been applied widely in the analysis of nonlinear signals and to the nonstable signal of SAR. The fusion of SAR imagery with fused panchromatic and multispectral imagery using BEMD is based on the frequency information of the images. It was established that the proposed fusion scheme is an effective remote sensing image interpretation method, and that the value of entropy and the spatial frequency of the fused images were improved in comparison with other techniques such as the discrete wavelet, a-trous, and non-subsampled contourlet transform methods. Compared with the original image, information entropy of the fusion image based on BEMD improves about 0.13–0.38. Compared with the other three methods it improves about 0.06–0.12. The average gradient of BEMD is 4%–6% greater than for other methods. BEMD maintains spatial frequency 3.2–4.0 higher than other methods. The experimental results showed the proposed fusion scheme could improve the accuracy of bamboo forest classification. Accuracy increased by 12.1%, and inaccuracy was reduced by 11.0%.

Multisource Remote Sensing Imagery Fusion Scheme Based on Bidimensional Empirical Mode Decomposition (BEMD) and Its Application to the Extraction of Bamboo Forest

Yun Chen

Papers

Multisource Remote Sensing Imagery Fusion Scheme Based on Bidimensional Empirical Mode Decomposition (BEMD) and Its Application to the Extraction of Bamboo Forest