VISION SYSTEM FOR ROBOTICS
Active In SP
Joined: Aug 2009
21-12-2009, 01:16 PM
Vision is the most valuable sense that the automation posses because the information that it contains is extremely rich and this is achieved without the need of physical contact The image extracted from the analog world is digitalized by using camera, the 'eye* of the robot, which is analyzed by the processor for the implementation of automation system. The whole process is synchronized.
The initial process of recognition is the segmentation, where the acquired image is 'broken up' into segments, as a part of edge detection. The final identified pixel is assigned co-ordinates and which is the input to the manipulator or the robot. In industrial view point, any machine vision system contributes to the efficiency of production, inspection or test function.
Morphological processing is efficient for the well fare of robotic in the highly competitive world of automation which provides a flexible programmability so that by using senses, which can perceive changes to the product to its environment and react accordingly thus vision system plays an important role in automation and control for robotic systems.
Vision is a powerful interface device for robots because of its potential for sensing body position, head orientation, direction of gaze, pointing commands and gestures. The vision-based interactions could make the machine more enjoyable or engaging or perhaps safer.
Vision systems may be used in different applications like operations that require information from the work environment and that include inspection navigation, identification of parts and assembly operations and communication image from the memory. If there is a match the part is accepted otherwise the part is either repaired or rejected. The image processing and analysis operation generally is made up of process like segmentation and edge detection. Most commercial vision systems for robotics have embedded routines that can be called from a macro, making it very easy to set up a system.
Image processing is the collection of routines and techniques that improve, simplify, enhance or otherwise alter an image analysis is the collection pf process in which a captured image that is prepared by image processing is analyzed in order to extract information about the image and to identify objects of facts about the object of its er- images arc used when the depth of the scene or its feature . do no, be determined. Three-dimensional image processing is dealt with operations that require motion detection ,depth measurement .remote sensing relative positioning and navigation .Computer tomography (CT) scan ,either X-ray or ultrasonic pulses are used 10 gel images of one slice of the object at a time ,and later .till of the images are put together Lo create a three-dimensional image of the internal characteristics of the object. Ail three dimensional vision shares the problem of coping with many to-one mapping of scenes to images.
Robotic Vision in context:
Robotic vision is used to describe any work which aims to provide a practical and affordable visual sense for any type of robot, they work in real Lime.' Exploitation of a priori knowledge about the working environment of the machine considerably eases the problem of understanding the environment. If robots are deployed in populated environments, it makes sense to base the perceptional skills used for localization on vision like humans do. Image retrieval system uses features that are invariant with respect to image translations, image rotations, and scale (up to a factor of two) in order to find the most similar matches. These features consist of histograms featuring of the local neighborhood of each pixel. This makes the localization system robust against occlusions and dynamics such as people walking by. During the filtering process the data of the samples are computed based on the similarity values generated by the retrieval system and according to the visibility area is computed for each reference image using a given map of the environment. The advantage of this approach is that the system is able to globally estimate the position of the robot and to recover from
possible localization failures. This system has been implemented and tested on a real robot system in a dynamic environment. In different experiments it has been shown to be able to globally estimate the position of she robot and to accurately keep track of it. The system is able to quickly determine the position of the robot and to reliably keep track of it despite of the dynamic aspects.