基于RGB线性组合模型的柑橘果实为害状识别

为了准确识别柑橘果实病虫害，提高柑橘生产信息化水平，本研究提出了最优RGB线性组合颜色模型来进行目标识别。首先在具有代表性的几幅图像中选取为害状区域与正常区域的点若干，统计这些样本点的R、G、B值及其均值，设计为害状区域与正常区域灰度差最大线性规划目标函数并求解，建立识别模型。最后利用识别模型结合阈值分割法对采集的96幅柑橘果实图像进行处理，发现识别正确率、误检率、漏检率分别81.25%，14.58%，4.17%，识别效果良好。与2R-G-B模型,G-B模型和R-B模型相比，本方法为害状区域与正常区域灰度差最大，故识别正确率最高，误检率、漏检率最低。试验结果表明：RGB线性组合模型法可用于柑橘果实病虫害的识别。

Citrus Fruit Disease and Insect Pest Recognition Based on Linear Combination of RGB Color Model

The investigation proposed the optimal linear combination of RGB color model which could recognize citrus fruit disease and insect pest accurately and improve the informationalized level of citrus fruit production. First, several points from the damage pattern region and normal region in the representative image were selected and these points of sampling of red, green, blue value and their average values were recorded. Then linear programming objective function was designed and solved based on larger gray difference in damage pattern region and normal region to establish recognition model. Finally, combined recognition model with threshold segmentation method was used to process 96 citrus fruits images, and the recognition accuracy, false detection rate and missing detection rate were found to be 81.25%, 14.58% and 4.17%, respectively, the result was good. Compared with 2R-G-B, G-B and R-B, this method had larger gray difference in damage pattern region and normal region, showed the highest recognition accuracy and lowest false detection rate and missing detection rate. The thesis concludes that the linear combination of RGB color model can be used in the citrus fruit disease and insect pest recognition.