论文部分内容阅读
为实现机器人幼果期自动化果实作业,以果园幼小青苹果为对象,研究了自然环境下幼果期苹果的机器侦测方法. 首先,采用自适应G-B色差法对初始图像计算,获得色差灰度图,使用迭代阈值分割法提取果实兴趣区;其次,对经形态学处理后的兴趣区图像进行Blob分析,计算每个Blob的离心率和像素面积,去除明显偏离果实形状特点的Blob;最后,应用改进圆形Hough变换算法检测潜在类圆形果实目标,最终采用融合方向梯度直方图特征和网格搜索优化支持向量机的判别模型进一步去除虚假果实目标,提升苹果目标的侦测精确度. 试验结果显示,该方法对果园自然环境下幼小青苹果的侦测正确率为88.51％,漏报率和误报率分别为11.49％和4.84％,算法模型综合性能指标为90.29％,表明该方法对幼果期苹果目标具有较强的侦测能力和较好的鲁棒性,该结果为果实作业机器人幼果期的自动化果实侦测提供参考.“,”In order to realize automatically managing fruit production by robot during young fruit period, this paper took young green apples in orchard as object and studied the detection method of young green apples by machine under natural environment. Firstly, adaptive green and blue chromatic aberration (AGBCA) map was designed and combined with the iterative threshold segmentation (ITS) algorithm to detect region of interest (ROI) contained potential apple fruits pixels. Then, potential fruits were identified by an improved circular hough transformation (CHT) after morphological operation and Blob analysis of the results obtained from AGBCA and ITS, which kept many potential apple fruits pixels as possible. Finally, a kernel support vector machine(SVM) classifier, optimized by grid search optimal algorithm, was built to remove false fruit objects based on histogram of oriented gradient (HOG) feature descriptor. The experimental results showed that the true positive rate of proposed method was 88.51％, false negative rate and false positive rate were 11.49％ and 4.84％, respectively. And the F1-Measure of proposed model was 90.29％, indicating the proposed method had better detection ability and robustness for young green apples detection. The results provided references to fruit robot for automatic detection during young fruit stage.