In the present modern world, humans are being replaced by computers in almost every field for various beneficial reasons. One such field, where intervention of computers is inevitable and making significant progress is “computer vision”, which replicates human vision. In computer vision system, similar to that of an eye, a sensor (a camera) is used to capture the images and different algorithms that usually run on a computer acts as brain for analyzing and understanding features in those captured images. Generally, image is visual representation of information. For better analyzing the information that an image has, many image processing techniques are being used. In real, complete view of an area is always challenging. Here, we are achieving monitoring surroundings of a robot by detecting moving objects from a sequence of panoramic stitched images, which is typically a feature extraction problem in computer vision. By knowing the surrounding environment of a robot, we can enhance its navigation and control. A camera setup, which consists of a single camera or multiple cameras, mounted on the robot is used for acquiring about twenty images, each at an angle of 20 degrees, so as to cover entire 360 degrees Images overlapping with the previous image are stitched to get a single panoramic image. Many of such stitched panoramic images can be created for the same scene at different intervals of time and these complete set of stitched images are then processed at the base station for detecting moving objects over that particular scene. So for analyzing the environment by the images acquired, they have to be transferred to the base station by means of a communication channel. Stitching can be done either at the base station or at the on-site. In the earlier case, we need to transfer all the twenty images per interval to the base station but in the latter we just have to send a single stitched image, which reduces the bandwidth usage. Further, using image c- mpression technique for compressing stitched image enhances the efficient usage of channel bandwidth.
C. Raghavachari and Dr. Shanmugha Sundaram G. A., “Efficient use of bandwidth by image compression for vision-based robotic navigation and control”, in Communications and Signal Processing (ICCSP), 2014 International Conference on, Melmaruvathur, 2014.