基于颜色特征的筒纱分拣机器人识别定位方法

Identification and Positioning Method of Cheese Sorting Robot Based on Color Feature

针对筒纱分拣机器人作业目标的表面纹理复杂、位置随机摆放等干扰因素的问题,课题组提出了基于颜色特征的筒纱识别定位方法。构建了具有视觉感知的4自由度DOBOT筒纱分拣机器人系统,通过视觉系统获取筒纱多目标的图像,采用对图像进行预处理的算法来提高分拣目标的对比度;将作业目标由RGB空间转换到HSV颜色空间,提取各分量的颜色特征,采用区域生长法对不同颜色的多目标区域进行提取;建立基于图像信息的形心坐标,对各目标区域的连通域进行定位。实验结果表明:该方法能够实现对不同颜色的作业目标的识别与定位,并在分拣机器人的手眼标定的基础上,实现了对不同颜色的作业目标进行分拣。

Aiming at the problems of complex surface textures and random placement of the targets in the cheese sorting robot,research group proposed a cheese recognition and positioning method based on color features. A four degrees of freedom DOBOT cheese sorting robot system with visual perception was constructed. The cheese multi-target images were obtained through the visual system,and the image was preprocessed to improve the contrast of the sorted targets. The job target was converted from RGB space to HSV color space,the color characteristics of each component were extracted,and multi-target regions of different colors were extracted by using the region growing method. The centroid coordinates based on image information were established to locate the connected domains of each target area. The experimental results show that the method can realize the identification and positioning of the work targets of different colors,and based on the hand-eye calibration of the sorting robot,it can sort the work targets of different colors.