基于多光谱成像技术的大麦赤霉病识别

该文提出了一种根据大麦多光谱图像实时识别大麦赤霉病害的方法。首先利用阈值分割以及形态学的处理算法去除大麦穗图像背景和麦芒干扰信息；其次从预处理后的多光谱图像中提取图像的颜色统计特征；最后将这些颜色统计特征数据经过预处理后应用偏最小二乘法（principal component analysis, PLS）进行模式特征分析，经过交互验证法判别选取最佳的主成分数，输入到最小二乘-支持向量机模型(least square-support vector machine, LS-SVM)，建立病害识别模型。经过比较发现多元散射校正处理后，最佳主成分为1的最小二乘支持向量机模型对病害的识别准确率最高，达到93.9%。表明利用多光谱成像信息可对大麦赤霉病进行准确识别，为植物病害监测与防治提供了一条新方法。

Identification of barley scab based on multi-spectral imaging technology

Site-specific variable pesticide application is one of the major precision crop production management operations. Barley scab identification and classification by human sight need special crop protect knowledge with the lower efficiency. A method for real-time and reliable detection of barley disease was developed. The samples of the infected barley and the healthy barley were collected. The backgrounds of images were removed by using the Nir channel image and the threshold segmentation algorithm, and the barley awn was removed by imopen function. Then statistical characteristics of images were captured, including the mean values and variances of the gray values of the image. After the statistical characteristics were preprocessed, Partial Least Squares (PLS) analysis was applied as calibration method as well as a way to extract the new eigenvectors which could be used to represent the information of original image data. The selected new eigenvectors were used as the input data matrix of least squares-support vector machine (LS-SVM) to develop LS-SVM identifying models and the barley, which was infected by barley scab or not, were used to be the outputs of the LS-SVM model. It was found that the the LS-SVM model was the best method with the predicting accuracy of 93.9%. The results indicateed that the method of identifying barley scab based on multi-spectral images was feasible. Thus, it is concluded that multi-spectral imaging technique is available for the detection of barley scab on the barley spike.