中文摘要：

在利用计算机视觉技术进行番茄缺素识别研究中,提出利用多分辨率小波分析算法提取番茄叶片的G体特征。结果表明：小波变换各分解子图的均值,方差,熵可区分正常与缺氮叶片;随着叶片含氮量的减少,小波分解子图各特征值随之变大。为利用特征值的范围,诊断叶片的缺素程度提供了技术支持。

英文摘要：

In this paper, green weight features （ RGB model） of tomato leaves are extracted with multi-resolution wavelet analysis algo- rithm in order to recognize the nutrient deficiency by using computer vision technology. The result shows that the Means, Variance, and En- tropy of the sub-images based on the wavelet decomposition can distinguish deficiency nutrient leaves from normal leaves, and with the reduc- tion of the nitrogen content in tomato leaves, the features＇ value of the sub-images increase. It provides support to diagnose the degree of nutrient deficiency by using the range of the features.