Improving the Quality of the Disparity Map

In computer vision tasks that are related to traffic, a method such as disparity mapping is widely used. With their help, you can estimate the distance to other objects and obstacles. The accuracy of subsequent work on determining distances to other objects depends on the quality of the calculation of the disparity map. When low-quality custom cameras are used in such tasks, the accuracy of calculating the disparity map sharply decreases. To increase the quality of disparity maps obtained in this way, it is proposed to apply the method described in this article. The advantage of this method is its simplicity with a satisfactory result relative to more accurate methods. The efficiency of the calculation is ensured by preliminary processing of the original image and subsequent filtering of the map, which makes it possible to obtain more visual and informative disparity maps.