基于Lab颜色空间下的小麦赤霉病图像分割

Image segmentation of wheat scab based on Lab color space

关键词:  小麦赤霉病  图像处理  颜色空间  阈值分割  聚类算法

Key words:  wheat scab  image processing  color spaces  threshold segmentation  clustering algorithm

摘要:

针对智能识别小麦赤霉病方法中分割患病麦穗图像效果不佳的问题，运用中值滤波方法对患赤霉病麦穗图像进行降噪预处理，采用基于阈值的最大类间方差算法(OTSU)、基于聚类的k-means算法在RGB、HSV和Lab颜色空间中对小麦扬花期到黄熟期感染赤霉病的麦穗图像进行分割，提取出麦穗的病害部分。采用试验田环境下扬花期到黄熟期200张患赤霉病的麦穗图像进行分割试验，结果表明:将图像从RGB颜色空间转化为Lab颜色空间并对a分量采用最大类间方差算法(OTSU)进行分割的效果最佳，误分率仅有1.11%。

Abstract:

Arming at solving the problem of poor segmentation effect of diseased ear images in the method of wheat scab intelligent identification, the median filter method was used to denoise the image of wheat ear disease pre-processing, and the maximum between-class variance algorithm(OTSU)based on threshold and k-means algorithm based on clustering were adopted to segment the wheat ear images infected with wheat scab in the RGB, HSV and Lab color spaces from the flowering stage to the yellow ripening stage, and the diseased parts of the wheat ears were extracted. A total of 200 images of wheat ears affected by wheat scab from flowering stage to ripening stage were used for segmentation experiment. The results showed that the best segmentation effect was obtained by transforming the image from RGB color space to Lab color space and using OTSU algorithm to segment the a component, and the difference ratio was only 1. 11%.