POTENTIAL OF NIR COMPUTER VISION TECHNIQUE FOR THE DETECTION OF MANGO’S DEFECTS

KRISHNA KUMAR PATEL; ABHIJIT KAR; KRISHNA KUMAR PATEL

KRISHNA KUMAR PATEL Division of Food Science & Post Harvest Technology, IARI, New Delhi – 110 012 (India)
ABHIJIT KAR Department of Post Harvest Engineering and Technology, Aligarh Muslim University, Aligarh 202002 (India)
KRISHNA KUMAR PATEL Assistant Professor, Department of Agricultural Engineering, F/O Agricultural Sciences, P.G.C., Ghazipur, U.P., India

Keywords: Defect, Detection, Mango, NIR imaging, etc

Abstract

Currently, research on computer vision technique is most prevalent for quality assessment of agricultural produce. This technique assists in achieving the specified objectives prior to sale of agricultural products in the global market. Defect detection is an important task during sorting and grading of mangoes. But, the colour camera based computer vision system (CVS) has been found unable to detect the hidden defects just below the peel of mangoes. The near infra-red (NIR) camera, in contrast, can be used for identification of such defects. A semi-automatic NIR CVS was developed for inspection of hidden defects just underneath of the peel of mangoes. Image acquisition, image processing and image analysis are the main functions of developed NIR CVS. The NIR camera was used to capture standard image from the samples of Chausa fruits. Algorithm steps were developed for image processing and analysis of hidden defects. Water core, fungal infection and the defects on the surface of mangoes, which were hidden just below the peel, identified and quantified. The present study, thus, showed that the NIR CVS has potential for detection of hidden defects of mangoes. The semi-automatic CVS can be upgraded to fully automatic for commercial application.