Improved Segmentation Approach for Plant Disease Detection

The Agricultural sector plays a vital role in sustainable economic growth and food security. However, crop diseases often cause a great threat in achieving this goal. As such, a successful outcome depends entirely on proper detection and classification of plant diseases. This has created many opportunities of new possibilities for researchers. Nowadays, a lot of work is being done to recognize and classify plant diseases more precisely using computer vision. The objective of this research is to create a methodology that will provide a better solution to classify plant diseases. This work mainly focuses on implementing an improved segmentation technique using a combination of thresholding and morphological operations. For classification, we have used the deep neural network. Our proposed method has achieved 99.25% accuracy in Plant Village database.