Chili Ripeness Grading Simulation Using Machine Learning Approach

Sorting and grading red chilies following harvesting into different levels of quality and ripeness is an important part of perishable product quality control process. The current manual grading and sorting practice is time-consuming, costly, and requires a large knowledgeable and experienced workforce. Automating sorting and grading red chilies is much more efficient and cost effective than the manual method. This paper proposes a convolutional neural networks-based approach for efficiently sorting and grading red chilies into their respective grades and ripeness. The proposed approach is validated with huge chili images dataset that consists of three grades and different ripeness. The results confirm that the proposed approach can efficiently grade and sort red chilies into their respective grades and ripeness with respectable accuracy. We believe that the proposed model will greatly benefit farmers in efficiently sorting and grading red chilies and making them quickly available for marketing and processing industries.