在视频序列检测织物折皱回复角中,对单帧织物回复翼图像的倾斜度求法由Hough变换向细化后取四等分点发展。尽管相对Hough变换在实时性上有所提高,然而连续单帧四等分点的不连续性所产生的奇异值限制其进一步发展。通过分析织物回复翼的形态及变化,本文采用最小投影长度法与极小惯量法来测量其倾斜度,并在测试平台上对二者与传统方法进行精度与实时性比较,结果表明极小惯量法在精度允许的范围内实时性最好。
In the measurement of fabric wrinkle recovery angle by video sequence, it has developed from hough transform to quartering point after thinning to measure fabric wrinkle recovery angle. Alought the real-time of quartering point after thinning is better than Hough transform, it is hard to further develop for discontinuity in sequence images. In this paper, we measure fabric wrinkle recovery angle by minimal length of projection and minimal value of moment of inertia after analyzing the shape of fabric wrinkle recovery wing. The results show that minimal value of moment of inertia is best measurement in precision and real-time comparing among three measurement.