摘要
在图像配准过程中,图像存在一定角度旋转是常见的问题,会造成后续视觉任务处理困难。同时,由于圆形目标图像不具有明显的外轮廓定位信息,故当其发生一定角度旋转时,检测难度有所增加。以多种圆形目标图像为研究对象,研究了常见的圆定位算法和旋转角度检测算法。实验结果表明,最小二乘法进行圆定位的误差最小,基于Zernike矩的旋转角度检测算法可以在兼顾速度的同时,达到较高的检测精度。
In the process of image registration,it is common for the image to have a certain angle rotation,which will cause difficulties in subsequent visual tasks.At the same time,since the circular target image does not have obvious outer contour positioning information,the detection difficulty increases when it is rotated at a certain angle.In this paper,a variety of circular target images are used as research objects,and the common circle positioning algorithms and rotation angle detection algorithms are studied.The experimental results show that the error of circle positioning by the least squares method is the smallest,the rotation angle detection algorithm based on Zernike moment can achieve high detection accuracy while taking into account speed.
出处
《工业控制计算机》
2024年第4期101-102,105,共3页
Industrial Control Computer
基金
大学生创新创业计划(S202210142024)
辽宁省科技研究计划(2023JH2/101300237)。
