An image processing and computer vision framework for efficient robotic sketching

摘要

The development of computer vision field has attracted researchers to use it in the robotics field. The computer vision field provides a useful tool for decision making and environment detection during the automation process. This is done by extracting useful information from an image or video. However, extraction of useful/desired information from an image or video is not an easy task. This is because sometimes the imaging device is placed under unfavorable environmental conditions such as poor or extreme illumination in object or lightning source, hazy environment, the relative motion of in between object and camera during the image/video capturing process etc. Hence the image/video must be pre-processed before sending to the computer vision application for further post-processing work. In this work, we have proposed a method that at first removes the unfavorable environmental conditions defects present in the image/video and then applies suitable computer vision method for extraction of useful/desired information. After that the proposed method generates an sketch of the object under choice which can be further sent to any robotic application for sketching. The proposed method is efficient and it can be used for robotics sketching applications.