基于目标保护的林业机器人视觉系统的研究

摘要

According to the demand of environment protection, the picking robot vision system to distinguish the shape features of Camelliaoleijera's fruits and leaves was impoved. First, was taken image was semeted by gray-threshold, then the figure parameters was picked, and the SVM to classification was used, the object samples was taken as antibody of the immune algorithm, and the artificial immune algorithm was improved to promote the precision of recognition. At last, by control ling the picking range based on the numbers of flowers, the aims of object protection were realized.%根据环境保护的需要，针对油茶林花瓣及果实的形状特征改进了采摘机器人视觉系统。首先通过灰度阈值法进行图像分割，再提取其形态参数，用支持向量机算法将参数分类，得到目标样本作为免疫算法的抗体，并进一步改进了人工免疫算法，提高了识别精度。最后根据最终图像中花数量多少来控制采摘，达到了目标保护的目的。