基于亚像素边缘检测的高速摄影下枪机运动分析

Bolt motion analysis in high-speed photographybased on sub-pixel edge detection

提出一种对高速摄影下枪机的运动图像进行分析,进而估算枪机位移和运动速度的测试方法。在图像预处理阶段使用了自适应的双边滤波算法,在保证对噪点滤除的同时对图像的边缘细节做到了尽可能的保留。使用Zernike矩亚像素边缘检测算法提取目标边缘,将检测精度提高到了一个像素以内。并通过Hough变换对边缘进行直线拟合,通过直线方程计算目标运动的像素距离。这种测试方法在测试空间狭小,且不接触目标的前提下,对枪机运动参数实现了准确预测,具有较高的实用价值。

A test method is presented for bolt motion analysis under high-speed photography, hence to estimate the bolt distance and motion velocity. During the image preprocessing, the adaptive bilateral filtering algorithm is adopted, which filters the noise as well as remain edge details.Using Zernike moment sub-pixel edge detection method to extract the target boundaries thus improve the detection accuracy to within one pixel. Hough transform is used for linear edgefit, and pixel distance of target movements is calculated by linear equations. Under the non-contact testing condition withinnarrow test space, this proposed method can accurately predict the bolt motion with a practical application prospect.