摘要
为了提高黄瓜叶部病害的识别效果及实用性,增强特征描述能力,提出一种新的黄瓜病害图像分割及颜色、纹理特征提取算法。首先,对自然环境中采集的病害图片进行不同颜色空间的预处理和分割,对分割结果图像进行融合处理;然后,对病斑区域提取模糊量化直方图、颜色聚合度作为颜色特征,利用颜色相似性度量函数计算颜色共生矩阵,提取病斑的纹理特征;最后,通过核主成分分析对颜色和纹理特征进行融合,去除冗余成分,利用支持向量机对病害进行分类识别。在采集的黄瓜病害实验库上进行实验,取得了94.11%的识别率。通过充分利用病害识别中占主要识别依据的颜色信息,并结合纹理信息,较好地表达了病斑特征。和其他方法的实验对比结果表明,本文方法能有效提高黄瓜病害的识别率。
In order to enhance the recognition effect and practicability of diseases in cucumber leaf,heighten the capacity of character description,a novel method segmenting the cucumber disease images and then extracting color features and texture features of these images is presented. Firstly,the disease images collected in natural environment in different color spaces are preprocessed and segmented. After that,the segmented images are fused. Secondly,fuzzy quantization histogram and polymerization degree of color of lesion area are extracted as the color features,then the texture features of lesion are extracted via the color co-occurrence matrix calculated by color similarity measurement function. Finally,kernel principal component analysis method is utilized to fuse the color features and texture features so that the redundant components can be removed,and support vector machine is used for classification. We do some experiments on the collected cucumber disease images database and achieve the recognition rate of 94. 11%. Though making the most use of the color features which play the leading role in cucumber disease recognition and combining with the texture information,lesion characteristics are preferably expressed. The comparison of experiment results of the proposed method with other algorithms show that our method can effectively improve the recognition rate of cucumber diseases.
出处
《电子测量与仪器学报》
CSCD
北大核心
2015年第7期970-977,共8页
Journal of Electronic Measurement and Instrumentation
基金
国家"863"计划(2012AA011103)
安徽省科技计划(1206c0805039)
国家自然科学基金(61432004)资助项目
关键词
叶部病害
模糊量化直方图
颜色聚合度
颜色共生矩阵
核主成分分析
leaf disease
fuzzy quantization histogram
polymerization degree of color
color co-occurrence matrix
kernel principal component analysis
