1 research outputs found
Deep Learning-Based Rice Quality Evaluation using Image Processing for Physical Attribute Analysis
No full textThe quality check of rice grain was done manually by experienced inspectors, but their analysis was incorrect.This paper proposes an automated strategy for collecting data on various rice types and analysing them based ontheir physical properties. We used methods such as computer vision and digital image processing, which includedpre-processing, morphological analysis, edge and object detection, and object measurement. The system is trainedusing both manual and machine learning techniques. The findings of image processing are saved to a file, andhypotheses for manual and machine learning training are generated using SVM and manual approaches. Thequality is then examined to establish whether the two ways result in higher or lower marks, and the bestmethodology is chosen through observation. 
