Texture Defect Detection in Gradient Space

In this paper, we propose a machine vision algorithm for automatically detecting defects in patterned textures with the help of gradient space and its energy. Gradient space image is obtained from the input defective image and is split into several blocks of size same as that of the periodic unit of the input defective image. Energy of the gradient space image is used as feature space for identifying defective and nondefective periodic blocks using Ward’s hierarchical clustering. Experiments on real fabric images with defects show that the proposed method can be used for automatic detection of fabric defects in textile industries. Keywords—cluster, defect, energy, gradient space, periodicity