基于计算机视觉的纸浆纤维特性检测与研究

Measurement of the Characteristics of Pulp Fibers Based on Computer Vision

设计了基于计算机视觉的纸浆纤维特征在线检测装置,并提出了用于在线特征提取的纤维图像处理算法,采集的在线纤维图像经图像去噪、图像增强、细化后便可以提取纤维的特征。采用纤维长度、宽度、弯曲度以及统计分布来描述纤维特征,除了传统方法中曲度的方法外,还利用拐角检测算法来描述纤维的弯曲特征。

An on-line measurement device of pulp fiber based on computer vision is designed according to the characteristics of pulp fiber. The image-processing algorithm is proposed to extract the characteristics of the fiber, which can be obtained after fiber image denoising, image enhancement and corner detection. In the paper, the fiber length, coarseness, curl degree and distribution are used to describe the fiber characteristics. The comer detection algorithm is applied to identify the curl characteristics of die fiber. Experimental results prove the validity and effectiveness of this method.