Edge-preserving Filtering and Fuzzy Image Enhancement in Depth Images Captured by Realsense Cameras in Robotic Applications

Abstract

This paper presents both the use of depth cameras in robotic applications and effects of post-processing on the captured depth images. The performance of depth cameras and post-processing image enhancement are evaluated with the aim to improve the depth-based object detection. First, the edge-preserving exponential moving average (EMA) filter and the fuzzy contrast enhancement procedures are briefly introduced. Then, the use of depth cameras with post-processing methods is shown in the example of painting robots. The use of the stereo depth camera is essential in robotic applications, since it constitutes the initial steps in a series of robotic operations, where the goal is to both detect and extract obstacles on walls that are not intended to be painted.