摘要
在苹果种植面积和产量日益增长的情况下,果实的各种病变如黑点病、斑点落叶病等也日益增多,严重影响了果农的经济收益。针对传统人工肉眼检测方法,受人为主观判断容易产生错误,降低病变识别准确度的缺陷,提出了基于最大类间方差法(OTSU)的苹果果实病斑图像分割方法。设计开发了苹果病斑图像分割处理系统,系统基于MATLAB GUI开发界面,将读取的苹果彩色病斑图像分别进行灰度处理、直方图均衡化、滤波增强、模糊增强、图像分割、识别病斑区域等一系列操作,分别选取苹果黑点病、斑点落叶病、苦痘病、红点病、痘斑病、日灼病等六种病斑果实图像进行采样处理,利用OTSU算法对六种苹果病斑图像进行分割识别。分割实验结果显示,对于病斑适中、颜色较深的苹果病变区域分割后识出率和识别成功率较高。
With the increasing of apple planting area and yield,the diseases of apple fruits such as black spot and blotch defoliation are increasing,which seriously affects the economic profit of fruit farmers.Aiming at the defects of traditional artificial eye detection method,which is subject to subjective judgment and easy to make mistakes and reduce the accuracy of disease detection,an image segmentation method of apple fruit disease spot based on the maximum inter-class variance(OTSU)method is proposed.Apple disease spot image segmentation and processing system is designed and developed based on Matlab GUI development interface,the apple color spot image is read by a series of operations such as grayscale processing,histogram equalization,filter enhancement,mold and enhancement,image segmentation,recognition of the spot area,etc.The images of apple black spot,apple spot falling,apple bitter spot,apple red spot,apple spot and apple sunburn are sampled and processed,and the images of apple black spot are segmented and recognized by OTSU Algorithm.The results of segmentation experiment show that the recognition rate and the success rate are higher for the apple lesion regions with moderate and dark color.
作者
薛飞
刘立群
XUE Fei;LIU Li-qun(School of Information Science and Technology,Gansu Agricultural University,Lanzhou 730070,China)
出处
《计算机技术与发展》
2020年第12期181-186,共6页
Computer Technology and Development
基金
甘肃农业大学信息科学技术学院发展基金(GAU-XKFZJJ-2020-11)
甘肃省高等学校科研项目(2019B-086)
甘肃农业大学学科建设专项基金(GAU-XKJS-2018-255,GAU-XKJS-2018-253,GAU-XKJS-2018-251)
国家自然科学基金(61741201)
甘肃省教育科学“十三五”规划2019年度一般规划课题(GS[2019]GHB2152)。
作者简介
薛飞(1989-),男,硕士研究生,助教,研究方向为农业图像处理;通信作者:刘立群(1982-),女,硕士,副教授,研究方向为智能计算、图像处理。
