Image processing techniques for insect shape detection in field crops

Abstract

In agriculture, crop pest detection is considered as one of the challenging tasks for the farmers. An automatic insect detection system using machine vision and image analysis provides better identification of crop insects on early stage with reduced time and greater accuracy which helps farmers to increase the crop yield. In the present work, digital image processing techniques were applied for crop insects images to perform preprocessing, segmentation and feature extraction to detect the shape of insects in the sugarcane crop. Sobel edge detection is applied to segment the insect image against background. In feature extraction, shape of the insect can be recognized by nine geometric shape features. This insect shape identification method performs well and achieves high accuracy for sugarcane crop insects with round (circle), oval, triangle and rectangle shapes. The present work was implemented in MATLAB 2015b using Image Processing Toolbox.