Detection of Material Defects Using Graph-Based Manifold Ranking and Heterogeneous Image Features

Abstract

The development of a method for detecting material defects using graph-based manifold ranking and heterogeneous image features is considered in the article. The aim of the research is to develop a method that allows high-precision detection of defect in images with low color contrast of detecting and background areas. The image is pre-segmented into regions to calculate the saliency map. The graph is based on regions. The defect is determined based on background area queries. The areas adjacent to the edges of the image belong to the background areas. Color features of the image are used in the existing approach based on the manifold ranking. Texture features are used in the proposed method to improve accuracy. Gabor's energy function is used to calculate texture features. The proposed method has shown good results for detection of material defect in images in which the background color and object color are in similar ranges. The experimental results are presented on test images. Precision-recall curves showing the advantage of the developed method are constructed.