Skill Gradient Descent Algorithm Based Convolutional Neural Network for Plant Growth and Crop Yield Prediction using Satellite Images

Agriculture is the backbone of all nations and it performs a major role in the entire Gross Domestic Product (GDP). Due to the weather conditions, geographical locations, land diversities, and diseases, monitoring crops becomes a demanding task. Former crop categorization and yield assessment modules are unsuccessful under adverse circumstances. Here, Skill Gradient Descent Algorithm based Convolutional Neural Network (SGDA-CNN) is developed for plant growth and Crop Yield Prediction (CYP) using satellite images. Initially, satellite is given to crop region segmentation. Then, crop region segmentation is conducted utilizing SegNet. Finally, predicting plant growth and crop yield is accomplished by employing a Convolutional Neural Network (CNN), which is trained utilizing the Skill Gradient Descent Algorithm (SGDA). The SGDA is designed by combining the Skill Optimization Algorithm (SOA) and the Gradient Descent (GD) optimization algorithm. The SGDA-CNN has gained error of 0.336, accuracy of 91% and Jaccard coefficient 0.901.