为了快速准确地获取黄瓜叶片的含氮量和叶面积指数等生长信息, 提出了采用多光谱图像技术对黄瓜生长信息进行检测的新方法。利用标定板建立黄瓜叶片光谱反射率同图像灰度值之间的线性公式。通过多光谱相机对样本在绿光、红光和近红外三个通道的图像进行处理, 获得叶片样本在每一通道的灰度值, 然后根据标定板所建立的灰度值与反射率间的经验线性公式将对应的灰度值转为反射率值, 并由反射率值计算出黄瓜的植被指数。采用最小二乘-支持向量机(LS-SVM)建立植被指数同叶片含氮量以及叶面积指数间的拟合模型。结果表明植被指数同叶片含氮量和叶面积指数的拟合相关系数分别为0.8665和0.8553。表明植被指数与黄瓜的叶片含氮量和叶面积指数具有紧密的相关性, 也为快速采集黄瓜生长信息提供了一种新方法。

In order to rapidly and accurately acquire cucumber growing information, such as the nitrogen content and leaf area index (LAI), a multi-spectral imaging technique was investigated. The linear relation between reflectance and image gray value was developed using the calibration board. The gray value of leaf sample was achieved by image processing of green, red and near infrared channels obtained by a three-channel CCD camera. Then the gray value of the leaf sample was transferred into reflectance value by aforementioned experiential linear function. The reflectance value was used for the calculation of vegetation index. Least squares-support vector machines (LS-SVM) model was developed for the relation between vegetation index and nitrogen content, vegetation index and leaf area index. The results indicate that the correlation coefficients of vegetation index and nitrogen content, vegetation index and LAI are 0.8665 and 0.8553, respectively. The overall results demonstrate that there is a close relation between the vegetation index and growing information of cucumber, and the multi-spectral imaging technique is a new powerful method for the acquisition of cucumber growing information.