微零件图像亚像素边缘定位算法

The Sub-Pixel Edge Location Algorithm of Micro-Part Image

  • 摘要: 为快速而精确地检测微零件图像边缘,提出了一种新的图像亚像素边缘定位算法. 此算法与常用亚像素边缘定位算法区别在于不需要进行边缘初定位,而直接提取亚像素边缘点. 应用一种新的图像坐标排序方法,使无规律的边缘点能够按照一定时针顺序链接成闭合曲线. 以微齿轮图像作为实验对象测试了算法的精度与速度,结果表明该算法在边缘定位精度满足亚像素的情况下,能够快速提取边缘曲线,处理时间不超过1 s.

     

    Abstract: In order to speedily get the edge curve of micro-part image with high accuracy, a new algorithm based on sub-pixel edge location is proposed. Differing from the usual algorithm of sub-pixel edge detection, the proposed algorithm extracts the sub-pixel edge points directly without the need for fixing the edge location aforehand, so the less time is spent and the processing speed is improved. In addition, a new sorting method of image coordinates is applied to link the disordered edge points according to a certain clockwise order and a two-step image denoising method is used to consist with the micro-part image structure. To show the effect of the proposed algorithm, a few of experiments has been done to test the processing speed and accuracy. The experimental data show that the processing time is generally less than 1 s and the edge location accuracy is better than half-pixel.

     

/

返回文章
返回
Baidu
map