Fighting against COVID-19: A novel deep learning model based on YOLO-v2 with ResNet-50 for medical face mask detection.

The Coronavirus disease 2019 (COVID-19) is the fastest transmittable virus caused by severe acute respiratory syndrome Coronavirus 2 (SARS-CoV-2). The detection of COVID-19 using artificial intelligence techniques and especially deep learning will help to detect this virus in early stages which will reflect in increasing the opportunities of fast recovery of patients worldwide. This will lead to release the pressure off the healthcare system around the world. In this research, classical data augmentation techniques along with Conditional Generative Adversarial Nets (CGAN) based on a deep transfer learning model for COVID-19 detection in chest CT scan images will be presented. The limited benchmark datasets for COVID-19 especially in chest CT images are the main motivation of this research. The main idea is to collect all the possible images for COVID-19 that exists until the very writing of this research and use the classical data augmentations along with CGAN to generate more images to help in the detection of the COVID-19. In this study, five different deep convolutional neural network-based models (AlexNet, VGGNet16, VGGNet19, GoogleNet, and ResNet50) have been selected for the investigation to detect the Coronavirus-infected patient using chest CT radiographs digital images. The classical data augmentations along with CGAN improve the performance of classification in all selected deep transfer models. The outcomes show that ResNet50 is the most appropriate deep learning model to detect the COVID-19 from limited chest CT dataset using the classical data augmentation with testing accuracy of 82.91%, sensitivity 77.66%, and specificity of 87.62%.

/pdf/a-deep-transfer-learning-model-with-classical-data-38suaw232h.pdf

A deep transfer learning model with classical data augmentation and CGAN to detect COVID-19 from chest CT radiography digital images.

Crop disease diagnosis is an essential step in crop disease treatment and is a hot issue in agricultural research. However, in agricultural production, identifying only coarse-grained diseases of crops is insufficient because treatment methods are different in different grades of even the same disease. Inappropriate treatments are not only ineffective in treating diseases but also affect crop yield and food safety. We combine IoT technology with deep learning to build an IoT system for crop fine-grained disease identification. This system can automatically detect crop diseases and send diagnostic results to farmers. We propose a multidimensional feature compensation residual neural network (MDFC-ResNet) model for fine-grained disease identification in the system. MDFC-ResNet identifies from three dimensions, namely, species, coarse-grained disease, and fine-grained disease and sets up a compensation layer that uses a compensation algorithm to fuse multidimensional recognition results. Experiments show that the MDFC-ResNet neural network has better recognition effect and is more instructive in actual agricultural production activities than other popular deep learning models.

/pdf/mdfc-resnet-an-agricultural-iot-system-to-accurately-aixg5t6b06.pdf

MDFC–ResNet: An Agricultural IoT System to Accurately Recognize Crop Diseases

Plant diseases are the primary issue that reduces agricultural yield and production, causing significant economic losses and instability in the food supply. In plants, citrus is a fruit crop of great economic importance, produced and typically grown in about 140 countries. However, citrus cultivation is widely affected by various factors, including pests and diseases, resulted in significant yield and quality losses. In recent years, computer vision and machine learning have been widely used in plant disease detection and classification, which present opportunities for early disease detection and bring improvements in the field of agriculture. Early and accurate detection of plant diseases is crucial to reducing the disease’s spread and damage to the crop. Therefore, this paper employs a two-stage deep CNN model for plant disease detection and citrus diseases classification using leaf images. The proposed model consists of two main stages; (a) proposing the potential target diseased areas using a region proposal network; (b) classification of the most likely target area to the corresponding disease class using a classifier. The proposed model delivers 94.37% accuracy in detection and an average precision of 95.8%. The findings demonstrate that the proposed model identifies and distinguishes between the three different citrus diseases, namely citrus black spot, citrus bacterial canker and Huanglongbing. The proposed model serves as a useful decision support tool for growers and farmers to recognize and classify citrus diseases.

Citrus disease detection and classification using end-to-end anchor-based deep learning model

Leaf and spike wheat disease detection & classification using an improved deep convolutional architecture

Deep learning has brought a huge improvement in the area of machine learning in general and most particularly in computer vision. The advancements of deep learning have been applied to various domains leading to tremendous achievements in the areas of machine learning and computer vision. Only recent works have introduced applying deep learning to the field of using computers in agriculture. The need for food production and food plants is of utmost importance for human society to meet the growing demands of an increased population. Automatic plant disease detection using plant images was originally tackled using traditional machine learning and image processing approaches resulting in limited accuracy results and a limited scope. Using deep learning in plant disease detection made it possible to produce higher prediction accuracies as well as broadened the scope of detected diseases and plant species considered. This article presents a survey of research papers that presented the use of deep learning in plant disease detection, and analyzes them in terms of the dataset used, models employed, and overall performance achieved.

Deep Learning in Plant Diseases Detection for Agricultural Crops: A Survey

The main objective of this paper was to develop a mathematical model for predicting the optimum vacuum pressure of a precision vacuum seeder using onion seed properties, vacuum characteristics and the hole geometry of seed plate.  The onion seed properties were  linear dimensions (length (L), width (W) and thickness (T)), one thousand seed mass (m), seed cross sectional area (As), projected area (Ap), geometrical mean diameter (Dg), sphericity (f), seed density (ρs), repose angle (q), the dynamic coefficient of friction (m), terminal velocity (Vt) and the drag coefficient, (Cd).  The vacuum characteristics were the vacuum velocity in general flow (Vv), vacuum density at normal temperature (rv), vacuum pressure at normal temperature (Pv) and the Reynolds¢s number (NRe).  The hole geometry of seed plate were the effective hole diameter (Do), the radius in sink flow (R) and the conical entrance angle (a).
Experiments were carried out under laboratory and soil bin conditions.  Laboratory experiments were conducted to determine the onion seed properties.  Engineering calculations were performed for estimating the vacuum characteristics and also for calculating the hole geometry of seed plate.  On the other hand, soil bin tests were carried out to verify the accuracy of developed model at various levels of blower (vacuum pump) speeds using three seed plates with different hole diameters.  The blower speeds (BS) used were 4000, 4500, 5000 and 5500-rpm and the different hole diameters of seed plates were 0.8 mm, 1 mm and 1.2 mm.
Results indicated that the developed model could satisfactorily describe the parameters affecting on determining the vacuum pressure in the hole with correlation coefficients ranged from 0.97 to 0.99.  It also showed that the final model could be used for predicting the optimum vacuum pressure of a precision vacuum seeder of onion seeds with an efficiency of 0.99

https://mjae.journals.ekb.eg/article_107382_fdd60950f9735a5fb2b5350f3f965878.pdf

Mathematical model for predicting vacuum pressure of onion seeds precision seeder

Furrow openers using single disc coulter operated at certain disc and tilt angles are
becoming common for single pass seeding operation. Concerns have been expressed as to the
efficiency of disc to place seed and fertilizer at a proper depth, and resulting compaction of the
furrow wall. Tests were conducted in an indoor soil bin using 460 mm (18-in.), and 610 mm (24-in.)
diameter disc coulters to determine the effect of combined disc and tilt angles on soil forces,
and resulting furrow wall strength. Experiments were carried out using various levels of
combined disc angles, moisture content, soil compaction, and depth of operations.
Measurements were taken for draft, vertical force, side force, soil strength of the furrow wall,
and furrow profile. Results indicated that the most favorable condition in terms of the lowest
values of draft, vertical force, side force, and soil strength of the furrow wall were obtained with
610 mm diameter disc at 5°&15° combined angle. However, for 460 mm diameter disc, the
lowest values of draft and side forces were obtained with 5°&15° and 5°&25° combined angles,
respectively.

Effect of Combined Disc Angles on Soil Forces of Coulter Discs

An onion peeling machine was designed, manufactured, developed and evaluated to suit the small and medium processing units. The onion peeling machine consists of seven parts, main frame, peeling drum, inlet and outlet openings, collection basin, water and air supplying systems, and power transmission. The machine was evaluated using different onion bulb sizes (small, medium & large). A mixed sample of different sizes was also tested. The evaluation process was conducted under three different drum rotational speeds (30, 40 and 50 rpm), three different peeling residence times (1, 2 and 3mins) and three different batch loads (18, 24 and 30kg). The optimum peeling efficiency of 74.9, 65.24, 80.08 and 85.45% were obtained at 24kg batch load (0.36 ton/h.), 2min peeling residence time and 40 rpm for small, medium, mixed and large sizes respectively.Water pump and air compressor were added to the peeler to improve the peeling efficiency, and it was tested under the previous conditions. The corresponding values of peeling efficiency for the machine with pump water at water pressure of 400kPa were 76.73, 83.06, 99.20 and 87.49 %, respectively. While they were 76.33, 72.87, 87.530 and 88.37%, respectively for air pumping at pressure of 500kPa. The estimated costs of onions peeling machine were 28.47, 29.56 and33.75 L.E/ton for the machine only, machine with water pump and machine with air compressor respectively. These values of estimated costs using the developed machine are very competitive with the cost of manual peeling process which approached about 75 L.E/ton.

https://mjae.journals.ekb.eg/article_102323_5dbdeecd837fbcd134eaf3a2d1f5f6e1.pdf

M. T. Afify

Papers

Deep Learning in Plant Diseases Detection for Agricultural Crops: A Survey

Mathematical model for predicting vacuum pressure of onion seeds precision seeder

Effect of Combined Disc Angles on Soil Forces of Coulter Discs

Development and evaluation of an onion peeling machine

A Single Unit Till-Planting System for Egyptian Soil