The Visual Saliency Detection Algorithm And Application Based On Contrast And Point Cloud Segmentation

The rapid development of video display technology has led to the explosive growth of image information data, how to deal with these information efficiently has become an urgent problem to be solved. Visual saliency detection technologies emerge as the times require, which greatly improve the processing efficiency of visual data.The algorithm of color contrast saliency detection is the mainly method in the detection of saliency. But it has trouble to deal with images with complex background and images contain high contrast regions in background. Focusing on the problem mentioned above, proposed algorithm consider both the color contrast and texture contrast, as well as point cloud segmentation using SURF feature.First, the principle of visual attention mechanism and contrast is introduced in this paper. And then, using the classical color contrast algorithm to divided super-pixel regions as arithmetic unit; Using the concept of global contrast combining with Euclidean distance, area weight and spatial relationships to calculate the color contrast.Second, the texture feature of the image is introduced, and using the Gabor filter to get the texture feature vector of the image. Using the global contrast to get the texture contrast and combined with the principal component analysis(PCA) to get the better detection results.Third, using the feature points detected by SURF operator as seed points, and segment image by using cloud model of region growing; It's not only solves the location detection of the salient region, but also can accurately detect the shape characteristics of the salient target.Finally, the three methods are combined together to obtain final saliency map. Experimental results show that this method has a higher precision and recall rate, which can not only resist the complex texture and the noise but also can effectively distinguish the region with high contrast in the background. And in this paper, we combine the saliency detection algorithm with the practical application, using the saliency detection algorithm to detect the surface defect of strip steel in industrial production. The practical application have achieved very good detection results, and further verify the practicality of the salient detection.