迭代式自适应法在Canny算子中的应用与研究

Application and Research of Iterative Adaptive Method in Canny Operator

为了增提高业机器人的智能水平,需要加装视觉系统,但因其工作环境复杂所以需要对现有机器视觉算法进行改进才能满足工业机器人的实际需要。边缘检测中广泛采用Canny算法,但因为其本身存在一些天然的缺陷不能直接应用于工业机器人上。第一,在进行高低阈值选取时需要人为设定,适应性差并且依赖人的经验。第二,在判断经非极大值处理后处于两阈值中间点是否为边缘点时,判断依据单一,容易混入更多的假边缘。通过研究现有机器视觉成果提出了相应的解决方案。第一,将迭代式自适应阈值法应用于Canny算法中,使算法有自主选择阈值的能力。第二,因为相邻边缘点处于相似的梯度环境中因此,使用梯度方向和(Sum of Gradient Direction简写为SGD)进行辅助判断边缘点。经仿真改进的方法,其结果表明改进的方法在检测精度和自适应方面要优于Canny算法,满足工业机器人对视觉方面的需求。

In order to improve the intelligence level of the robot,it is necessary to install the vision system,but because of the complex working environment,it is necessary to improve the existing machine vision algorithm to meet the actual needs of industrial robots.Canny algorithm is widely used in edge detection,but because of some natural defects,it can not be applied directly to industrial robots.First,in the choice of high-low threshold,artificial settings be needed,poor adaptability,and rely on human experience.Second,when judging whether the middle point of the two thresholds is the edge point after the treatment of the non maximum value,the judgment is simple,and more false edges are easy to be mixed in.Through the learning of existing machine vision results,the corresponding solutions are put forward.First,the iterative adaptive threshold method is applied to the Canny algorithm,so that the algorithm has the ability to select the threshold automatically.Second,the adjacent edge points should be in a similar gradient environment,so the Sum of Gradient can be used to help judge the edge points.Simulation results show that the improved method is superior to Canny algorithm in detection accuracy and self-adaptation,satisfy the visual needs of industrial robots.