Identification of the Zigzag Leafhopper on a light trap using digital image processing

Abstract

This thesis presents the identification process of zigzag Leafhopper in paddy field on lights trap using machine vision. The objective is to identify the physical features of the Zigzag Leafhopper which can be used to distinguish it from other species of leafhopper such as Brown and Green hopper by its unique characteristic that is zigzag pattern. To solve this problem, machine vision will be used to capture zigzag leafhopper image which will be used for further analysis by developed algorithm in LabVIEW. This thesis determined the accuracy of the algorithm for the Zigzag Image. As a conclusion, all the objectives for this study have been achieved and the application of machine vision or vision assistant can be extended in other fields of agriculture in Malaysia. The threshold for Zigzag Leafhopper is 130 for minimum and 230 for maximum. The value of P value for all parameters lower than alpha value which in 0.05. This proposed method can identify the Zigzag Leafhopper with 95% accuracy of the time. The value of max and min area were 20576 and 3036, percent area per image were 14.89 and 2.44, convex hull area were 7.37 and 1.52, convex hull perimeter are 0.76 and 0.18 while hydraulic radius were 1541.92 and 338.81. From the value of mean, the value of max and min needed to detect and not detect the zigzag pattern. The max and min value was different from each sample due to feature. The value of max and min values area were 11161 and 527, percent area per image were 9.14 and 0.54, elongation were 58547 and 2319, compactness were 1072.19 and 361.01 while perimeter were 8.09 and 1.21. When the image in this range so result be as Zigzag Leafhopper.