The Research Of The Manipulator Self-recognition Target Based On Immune Ant Colony Algorithm

With the rapid development of science and technology, the research of mechanicalautomation and artificial intelligence control have become an important research direction forthe majority of scientists, more and more applications have been used to the industrialproduction and daily life. As an important part of the robot, visual image processingtechnology has great changes. Robot Vision makes robot have a self-sensing function. Robotcan obtain two-dimensional image of the environment through the sensors, which can betransformed to the computer symbols after analyzing, and the environment judgments andbehavioral decision are conducted by using artificial intelligence technique.Taking the MOTOMAN six-degrees freedom industrial robot in the automatic logisticslaboratory of XUST as the research platform, a manipulator target detection image processingprogram is designed for the robot working environment. According to the analysis andprocessing of the target image, the results of image filtering and image graying for themanipulator target are given. Compared the work environment and characteristics of everykinds of image segmentation methods, the advantages and disadvantages, and the scope ofapplication of segmentation methods are analyzed, which lays a foundation for study of imagesegmentation for the target image further.Next, according to the image segmentation problem of MOTOMAN robot, the antcolony algorithm and immune algorithm are studied systematically. The system characteristics,application scope, advantages and disadvantages of two algorithms are analysed andcompared based on studying the basic principle of algorithm. Various applications ofalgorithm in nonlinear fields are studied and analysed. thus, an idea of segmentation based onimmune ant colony algorithm is proposed.Based on the analysis of image processing technique related, according to the efficient image segmentation demand required by the target self-recognition of manipulator, a methodof image segmentation based on immune ant colony algorithm is proposed in this paper. Themethod that takes pheromone as standard adopts ant algorithm to traversing the whole imageand combines the immune algorithm to avoid segmentation errors by local optimal solutionand the stagnation of convergence while ants travel. Immunization selection throughvaccination optimizes the heuristic information further, which improves the efficiency of theergodic process and shortens the time of segmentation effectively. Large numbers ofsimulation experiments show the effectiveness and feasibility of this method in theapplication of manipulator target image processing. The relevant information of the targetimage is obtained accurately through the information processing to the target image.Finally, according to the experimental environment, programs of catching, handling,processing and releasing the goods of manipulator are developed based on the programminglanguage of MOTOMAN robot.