Pseudo Color Region Features for Plant Disease Detection

This study reports a novel pseudo color region features for a computer vision system for the identification of diseases in Tomato Plants. The HSV based algorithm identifies eccentric and non- eccentric dots, spots, patches and region of different pseudo colors. Proposed method uses region properties and creates an enhanced and effective feature vector for plant disease detection. The features are more intuitive for humans to understand and help in tuning the underlying Artificial Intelligence Model better. The algorithm uses a scalable data structure to store regions counts using a hash function. It has wide application in the Computer Vision domain.