In the past decade, with researchers investing a lot of research work, computer vision, machine vision and other theories continue to develop and improve, mobile robot vision system has involved image acquisition, compression coding and transmission, image enhancement, edge detection, threshold segmentation, target recognition, three-dimensional reconstruction, etc., covering almost all aspects of machine vision. Mobile robot vision systems are mainly used in the following three aspects:
1. Use visual inspection of products to replace human visual inspection. Including: shape inspection, that is, checking and measuring the geometric dimensions, shape and position of parts; Defect inspection, that is, checking whether parts are damaged and scratched; Complete inspection, that is, checking whether the parts on the assembly are complete.
2. Identify the parts to be assembled one by one, determine their spatial position and direction, guide the robot's hand to accurately grasp the required parts, and put them in the designated position to complete the classification, handling and assembly tasks.
3. Navigate the mobile robot, use the vision system to provide the mobile robot with the external information of its environment, so that the robot can independently plan its travel route, avoid obstacles, safely reach the destination and complete the specified work task.
The first two belong to the category of industrial robots, and with the development of technology, researchers have also proposed the concept of visual servos. Visual navigation, whether autonomous, semi-autonomous, or the earliest remote control method, is an interactive process for the actor (subject) and the environment (object) to complete a certain task.
Nevertheless, the vision of mobile robots has not yet reached the point where vision is so important to humans for mobile robots, in large part because of the imperfection of visual information processing theories and methods. The camera can scan millions of pixels of the environment in a very short time, even exceeding the information collection ability of the human eye now, but the processing method and processing speed are far from being comparable to humans.
However, it can be believed that with the advancement of microelectronics technology and the development of computer vision, the visual function of mobile robots will become more and more powerful, and machine vision will account for an increasing proportion in the information perception of mobile robots.