基于一维最优Gabor滤波器的帘子布疵点检测算法

传统的二维Gabor滤波器用于帘子布疵点检测非常有效,但计算量巨大,为此提出一种快速自动检测帘子布疵点的方法。通过一维环形投影变换把一幅二维灰度图像压缩成一维形式,然后使用一维最优Gabor滤波器检测嵌入在帘子布均匀纹理中的疵点。采用粒子群优化算法进行一维Gabor滤波器最优参数的求解。假设图像大小为N×N,滤波器窗口为W×W,计算的复杂度可以显著地从二维Gabor空间的O(W2N2)到一维Gabor空间的O(WN2),并用实验验证了该方法的有效性。

Fabric defect detection based on 1D Optimal Gabor filter

Traditional 2D Gabor filtering schemes have been shown to be very effective for detecting local anomalies in textured surfaces of the fabric. However, they are computationally expensive. In order to alle-viate the limitations of 2D Gabor filtering, this paper used a 1D ring projection transformation to compress a 2D grey-level image into a 1D pattern, and then employed a 1D Optimal Gabor filter to detect defects embedded in a homogeneous texture. The optimal parameters of 1D Optimal Gabor filter are obtained by using particle swarm optimization algorithm. Given a problem with image size N×N and filter window O(W×W), the computational complexity can be reduced significantly from O(W2N2) in the 2D Gabor space to O(WN2) in the 1D Gabor space. The experiments on fabric have shown the efficiency and effectiveness of the method.