基于多方向光源的色纱织物密度图像检测

Density Image Detection of Yarn-Dyed Woven Fabric Based on Multi-Directional Light Source

针对色纱织物存在不同颜色,密度检测具有较大难度的问题,课题组提出了一种基于多方向光源的色纱织物密度图像检测算法。利用织物的3D表面结构来弱化色纱织物的色彩信号,使用正方形分布的4个定向光源在特定的照射角度下对色纱织物进行照明及采样,再把4个图像进行融合,使融合后的图像集成了每张图像中较清晰的部分。对于局部加权回归算法中宽度系数不能因为织物改变而自适应的问题,提出了自适应算法,让不同织物具有适合的宽度系数去平滑其投影曲线。实验结果表明提出的算法准确率可达99. 17%,适合在色纱织物密度检测方面进行推广。

Aiming at the difficulty of density detection of yarn-dyed woven fabrics with different color,a yarn-dyed woven fabric density detection algorithm based on multi-directional light source was proposed.The three-dimensional surface structure of fabric was used to weaken the color signals of yarn-dyed woven fabric,and the fabric was illuminated and sampled at a specific irradiation angle with four directional light sources with square distribution.The four sampling images were fused,so that the merged image integrates the sharper portions of each image.For the problem that the width coefficient in the local weighted regression algorithm cannot be adaptive to the fabric changes,an adaptive algorithm was proposed to make different fabrics have suitable width coefficients to smooth the projection curves.The experiments show that the accuracy of the algorithm is up to 99.17%,which is suitable for promotion of the yarn-dyed woven fabric density detection.