基于图像识别的动物纤维细度快速测量算法

Animal fiber diameter measuring algorithm using image recognition

以图像识别、图像分析为基础,研究动物纤维平均细度测量算法。采用图像二值化、边缘跟踪算法提取纤维边缘,用边缘突变点提取算法、纤维微段识别算法确定纤维微段,以边缘亚像素定位算法获取纤维边缘的准确位置,计算纤维细度。结果表明:基于该算法的动物纤维细度自动测量系统的测量精度达到±0.135μm,该算法能够准确、快速地测量动物纤维的细度,满足自动测量的要求。

The fiber diameter measuring algorithm is a key part of an automatic animal fiber diameter measurement system. The algorithm described here used image processing methods with the fiber images first digitized so the edges could be detected. Then the edges were subdivided into short curved segments by identifying critical points so fiber snippets could be extracted from the images. The fiber diameters were then calculated after the fiber edges were repositioned by a subpixel positioning algorithm. The measuring process is fast and simple; the result is reliable and accurate and the measuring precision can reach ±0.135 μm.