Performance Evaluation of Pest Detection Techniques via Image Processing in Agriculture

Abstract

In India 75% of population is dependent on agriculture. It provides not only food but also playing a key role in the economy of a country. Due to the attacks of bio aggressors or pests, agricultural byproduct lost heavily in every year. This paper is based on image processing methods to detect pest. Some of the segmentation techniques like region of interest (ROI), relative difference in intensities (RDI) and k-mean (KMEAN) clustering are implemented and tested for the detection of pests and extracting from its background. Performance of three above techniques can be analyzed by image quality assessment parameters such as structural content (SC), normalized absolute error (NAE), normalized correlation coefficient (NCC), average differences (AD) and peak signal to noise ratio (PSNR). ROI achieves better performance among all the techniques. This algorithms was developed and implemented by MATLAB 9.1 2017a version.