A Study of the Identification System for Insect with Machine Vision

• Main Authors: Jia-Wei Hsu, 徐嘉偉
• Other Authors: Kuo-Yi Huang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/89441327122465772898

碩士 === 華梵大學 === 機電工程學系博碩專班 === 98 === In this paper, the classification system for insects has been developed with machine vision. The methods include the contour match principle, Bayes classifier. The system can detect the three insects using the proposed algorithm. The accuracy of detection is 93%. The algorithms are described as follows: 1.Contour searching principle (1)The insect contour is found using image process techniques and the active contour model. (2)Write a programming. 2.Color classification Algorithm (1)Use image process techniques to extract the color features. (2)The insects are classified using Bayes classifier according to the color features. (3)Write a classification programming.