AI BASED ONLINE QUALITY EVALUATION OF SOYBEAN

Abstract

This research paper is about system, developed for quality evaluation of soybean seed by applying AI-ML with minimum time and cost. Images of sample soybean seed with all possible anomalies were captured. Data sets have been prepared to train machine learning model. Tensor flow based framework has been used to build faster RCNN model (Region-based Convolution Neural network). The trained model used to perform fast image analysis and provide counts of soybean seeds and non-soybean objects. The quality grade of sample soybean seed has been derived based on these counts. Results obtained during laboratory experimentation were encouraging to use this system for soybean processing industry