Classification of Chili Plant Growth using Deep Learning

Chili is among top grown crops in Malaysia. Nevertheless, chili plant growth monitoring in Malaysia is still performed manually by human labor, which consumed lots of time, energy and the plant growth can only be monitored at site. Hence, this paper discussed the potential of deep learning algorithm in classifying chili plant growth images from a small dataset. The experiment is performed on 256 chili plant images that under various conditions, where the images were captured using a 12-megapixel and f/1.8 aperture camera. Experimented using ResNet-101 and ResNet-50 of Mask R-CNN models with 75% of the dataset for training and 25% for testing, the results showed that both models were able to detect the correct age of chili plants with accuracy of 96% for Mask R-CNN ResNe-50 which is lower than Mask R-CNN ResNe-101 by 1%.