基于计算机视觉的菜青虫检测系统的研究

A Cabbage Caterpillar Detection Method Based on Computer Vision

针对传统人工识别及计数受人为主观性等因素的影响,致使虫识别效果与计数准确性不佳的问题,本文采用Hog特征及SVM分类器实现菜青虫的识别,通过对图片进行模块划分和梯度计算获取HOG特征,通过Cell复用获取样本,运用SVM分类器进行菜青虫的分类识别,实验结果表明SVM+HOG分类识别方法正确识别率高,识别速度快。

In view of the influence of traditional artificial recognition and counting subjectivity and other factors,the insect recognition effect and counting accuracy are not good.This paper adopts HOG feature and SVM classifier to realize the identification of cabbage caterpillar.The HOG feature is obtained by partitioning and gradient calculation.The samples are obtained by Cell multiplexing.The SVM classifier is used to classify and identify the cabbage caterpillar.The experimental results show that the SVM+HOG classification and recognition method has high recognition rate and fast recognition speed.