Development of a Distance Measurement System Based on Stereoscopic Vision for the NAO Humanoid Robot

This work presents the development and implementation of a distance measurement system based on stereoscopicvision using the upper and lower cameras of the NAO humanoid robot. The proposed methodology incorporates image processing techniques, including temporal stabilization filters, texture quality analysis, and adaptive algorithms to enhance measurement accuracy. The implemented system utilizes the OpenCV library for image processing and vertical stereo matching techniques, taking into account the specific configuration of the NAO robot’s cameras. Experimental results demonstrated that the system is capable of performing distance measurements with improved stability through the implementation of temporal stabilization buffers and adaptive quality filters. Although the obtained measurements exhibited high variability due to the challenging vertical camera configuration, the use of smoothing filters and texture analysis allowed common stereoscopic vision issues to be identified and mitigated. This work contributes to the advancement of computer vision applications in humanoid robotics, specifically in the areas of spatial perception and autonomous navigation.