Processing and analyzing object images with the view of object grasping

Résumé

This paper presents a solution based on image processing techniques used for analyzing images of objects which will be grasped, using existing software technologies completed with specific software modules. Digital image processing, as part of the grasping strategy, is based on a number of Microsoft technologies used to manage the devices attached to a computer, and a series of image processing techniques. The analysis is made directly on the video sequence taken from a video device (web cam). From the video, sequence frames are extracted which are then converted into RGB format and transformed into digital images used for analysis. Based on the existing technologies review, several necessary adjustments are identified to the proposed objective, and several specific software modules which can be conceived are described in details, and are then used to facilitate the implementation. The sequence presented in this paper is a part of an ample process, that aims to optimize the grasping strategy used for anthropomorphic prehensors with fingers from robot equipment.