基于轮廓分析的双串叠贴葡萄目标识别方法
发布时间：2019-08-09 15:53
【摘要】：为准确定位叠贴情况下的葡萄目标,提出了一种基于轮廓分析的双串叠贴葡萄目标识别方法。首先提取最能突显夏黑葡萄的HSV颜色空间中的H分量,通过改进K-means聚类方法对葡萄图像进行分割,运用形态学去噪等处理获取葡萄图像区域,再提取该区域边缘轮廓和左右轮廓的类圆中心。然后以该中心点为原点建立基于轮廓分析的叠贴葡萄串分界线几何求解与计算模型,分别在逆时针方向45°~135°和225°~315°区域内沿葡萄轮廓搜索距离原点最近的点,进而确立两叠贴葡萄轮廓拐点及其分界线,最终实现对叠贴葡萄目标的分别提取。对从果园采集的27幅双串叠贴葡萄图像进行试验,结果显示:24幅图像中的叠贴葡萄串被正确识别和提取,成功率达88.89%,目标像素区域的识别精准度为87.63%~96.12%,算法处理时间在0.59~0.68 s之间。将算法移植到自主研制的机器人上进行视觉定位试验,结果表明所提方法可很好地用于两叠贴葡萄目标的识别与定位。
[Abstract]:In order to accurately locate grape targets under overlapping conditions, a double-string overlapping grape target recognition method based on profile analysis is proposed. Firstly, the H component in the HSV color space of summer black grape is extracted, the grape image is segmented by improved K-means clustering method, and the grape image region is obtained by morphological denoising, and then the edge outline and the center of the left and right contours of the region are extracted. Then, taking the center point as the origin, the geometric solution and calculation model of overlapping grape string dividing line based on outline analysis are established. The nearest point to the origin is searched along the grape outline in the counterclockwise direction of 45 掳~ 135 掳and 225 掳~ 315 掳, respectively, and then the inflection point and dividing line of the two stacked grape contours are established, and finally the target of stacked grape is extracted respectively. 27 double-string overlapping grape images collected from orchards were tested. The results showed that the success rate was 88.89%, the recognition accuracy of target pixel region was 87.63% and 96.12%, and the processing time of the algorithm was between 0.59 and 0.68 s, the success rate was 88.89%, the accuracy of target pixel region was 87.63% and 96.12%, and the processing time of the algorithm was 0.59 鈮，
本文编号：2524875
[Abstract]:In order to accurately locate grape targets under overlapping conditions, a double-string overlapping grape target recognition method based on profile analysis is proposed. Firstly, the H component in the HSV color space of summer black grape is extracted, the grape image is segmented by improved K-means clustering method, and the grape image region is obtained by morphological denoising, and then the edge outline and the center of the left and right contours of the region are extracted. Then, taking the center point as the origin, the geometric solution and calculation model of overlapping grape string dividing line based on outline analysis are established. The nearest point to the origin is searched along the grape outline in the counterclockwise direction of 45 掳~ 135 掳and 225 掳~ 315 掳, respectively, and then the inflection point and dividing line of the two stacked grape contours are established, and finally the target of stacked grape is extracted respectively. 27 double-string overlapping grape images collected from orchards were tested. The results showed that the success rate was 88.89%, the recognition accuracy of target pixel region was 87.63% and 96.12%, and the processing time of the algorithm was between 0.59 and 0.68 s, the success rate was 88.89%, the accuracy of target pixel region was 87.63% and 96.12%, and the processing time of the algorithm was 0.59 鈮，
本文编号：2524875
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2524875.html
