摘要
针对规模化种植农作物的病变判断问题,本文利用机器视觉技术,提出了表面多特征决策融合的农作物病变判断算法。采集多张农作物叶片的图像,对每一张叶片分别通过不同算法提取表面特征;设置特征权重,将不同特征结合起来判断该叶片出现病变的可能性;并对所有的叶片的结果采用大多数投票决策法对该区域病变情况进行判断。与人工判断方式相比,本文算法减少了工作量,同时可以实现农作物病变的有效预警。
For the judgment of the lesions of large-scale planted crops,this paper uses machine vision technology to propose an agricultural disease judgment algorithm based on surface multiple feature decision fusion.First the images of multiple agricultural leaves are collected that different surface feature extraction algorithms are applied to each leaf. Second,the feature weights are set and different features are combined to determine the likelihood of lesions in the leaf.Finally,the method of majority voting decision is used'to judge the lesions in the region of each leaf. Compare with the artificial judgment method,this algorithm reduces the workload and can provide effective early warning of crop diseases.
作者
刘恩泽
吴文福
LIU En-ze;WU Wen-fu(College of Biological and Agricultural Engineering,Jilin University,Changchun 130022,China)
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2018年第6期1873-1878,共6页
Journal of Jilin University:Engineering and Technology Edition
基金
国家重点研发计划项目(2016YFD0401001)
关键词
计算机应用
机器视觉
农作物病变判断
表面多特征
特征权重
大多数投票决策
computer application
machine vision
agricultural disease judgment
surface multiple feature
feature weight
majority voting decision
