改进Canny算法在码垛机器人视觉图像处理中的应用研究

Application of improved Canny algorithm in vision image processing of palletizing robot

针对实际应用过程中码垛机器人视觉图像处理易受车间光照不足、噪声干扰等因素影响的问题,对码垛机器人视觉系统视觉图像检测原理进行了研究。通过对比常用边缘检测算法的检测性能,提出了一种改进的Canny算法;通过改进梯度幅值的计算方法,提高了去噪效果和边缘定位精度,并通过双阈值的自适应提取的方法,实现了自适应设定阈值对图像进行边缘检测;通过实验,对边缘检测算法的检测效果进行了测试分析。研究结果表明:改进的Canny算法能有效克服光照不足、噪声干扰等因素影响,避免边缘检测过程中出现的断边和虚假边缘,其检测效果优于传统的Canny算法以及其他几种边缘检测方法。

In order to solve the problems of the visual image processing of palletizing robot was easily infected by insufficient illumination and noise interference,the principle of visual image detection of palletizing robot vision system was investigated. Compared with the detection performance of edge detection algorithm,an improved Canny algorithm was proposed. The denoising effect and edge positioning accuracy were improved by improving the calculation method of gradient amplitude,and the adaptive setting threshold was used to detect the edge of the image. The test results of the edge detection algorithm were tested and analyzed by experiment. The results indicate that the improved Canny algorithm can effectively overcome the influence of illumination,noise and other factors,and avoid the broken edge and false edges in the edge detection process. The detection result is superior to the traditional Canny algorithm and several other edge detection methods.