Identification and counting the Green Leafhopper (GLH) on a light trap by machine vision

Abstract

There are almost more than 200 organic pesticides and more than thousand formulation used to solve the pest problems in this world. Generally, the method that protected crops from diseases and control the pests by biological means instead of pesticides. For the diseases of plant or some non-chemical control methods for pest contains a strong demand presented in few cuontries now. In fact, greenhouse staffs will periodically observe the plants or search for pests in production condition. However, this manual method is time consuming and waste the human resource. An autonomous system which use in disease classification and analyzation of crops is able to build up some advancements for pattern recognition techniques in machine vision. In this study, the machine vision will be used in order to detect, recognize, and counting green leafhopper which can cause the disease of rice dwarf virus (RDV) to the rice plant in the light trap. The purpose of the machine vision created is to replace the old version of counting process which reduce the time taken on counting green leafhopper which can destroy the farm in short period cause a huge amount of loss. Machine Vision is a system which build up with a camera as the sensor and extract the output image of the green leafhopper after proceed with the algorithm proposed. The accuracy and persistency of the machine vision will be evaluated successfully in this study.