协作机器人关节预成型体表面编织角视觉检测

Visual detection of braiding angle on the surface of joint preforms of collaborative robot

针对协作机器人90°关节碳纤维预成型体表面编织角测量难度大、准确度低等问题,提出一种离散角均值灰度编织角检测算法用于预成型体表面纤维平均编织角的测量。首先通过图像采集设备进行预成型体表面图像的采集,进而通过插值法补偿机器人关节表面图像,使用边缘检测和二维离散傅里叶变换等方法进行图像处理,并且使用离散角均值灰度编织角检测算法进行编织角角度的具体计算。试验结果表明:运用此方法检测编织角的平均误差为0.186°,平均检测时间为0.856 s。该算法能够在实际生产中提高编织角检测速度和准确率。

Aiming at the difficulty,low accuracy and so on of measuring the surface braiding angle of carbon fiber preforms with 90°joints of collaborative robots,a discrete angle mean grayscale braiding angle detection algorithm was proposed to quantitatively measure the average braiding angle of preformed surface fibers.Firstly,the surface image of the preformed fabric was collected by the image acquisition device,then the surface image of the robot joint was compensated by the interpolation method.Image treatment was performed by edge detection and two⁃dimensional discrete Fourier Transform method and so on.Detecting algorithm of walk⁃off angle grayscale mean value for braiding angle was adopted for the specific calculation of the braiding angle.The test results showed that the average error of detecting braiding angle by this method was 0.186°,and the average detection time was 0.856 s.This algorithm could improve the speed and accuracy of braiding angle detection in actual production.