WANG Shouting, LU Zhaoquan, MA Huili. Visual-based pose detection method for specific objects[J]. 2021, (4): 150-153+160. DOI: 10.13873/J.1000-9787(2021)04-0150-04.
基于视觉的特定目标物位姿检测方法
摘要
针对工业机器人基于位置的视觉伺服控制需要获取目标物位姿信息的问题
提出了一种简便的检测方法。将相机拍摄的图像从三原色(RGB)空间转换到HSV空间
利用目标物与背景的HSV信息有显著差异的特点分离出目标物所在区域。对该区域进行边缘检测
根据霍夫变换的结果和一些先验知识筛选出目标物的特征边缘及特征点。通过相机标定提前获取相机的内外参数
用这些参数将特征点的像素坐标转化为用户坐标
进而逐步求出目标物在用户坐标系下的位姿信息。以检测方形物体为例进行了多组实验
结果显示单次检测时间在0.4 s左右
平均检测误差不超过2 mm。
Abstract
Aiming at problem that position-based visual servo control of industrial robots needs to acquire position and pose information of object
a simple detection method is proposed.The image captured by camera is transformed from RGB space to HSV space
and the region where the object is located is separated by the feature that the HSV information of the object and the background is significantly different.Edge detection is carried out in this region
and feature edges and feature points of the target are selected according to the results of Hough transform and some prior knowledge.Internal and external parameters of camerain is obtained in advance.With these parameters
the pixel coordinates of the feature points are transformed into user coordinates
and then the position and pose information of the object in user coordinates is obtained step by step.Taking the detection of square objects as an example
a number of experiments are carried out.The results show that the single detection time is about 0.4 s
and the average error of detection is less than 2 mm.