The Research Of Manipulator Grasping Based On Image Recognition Of Artificial Immune Algorithm

In modern society, the robot plays an important role in the automated and intellectualized process of production and living; it was studied and applied widely in the fields of the society, meanwhile, the machine vision technology provides a bionic data acquisition method for robots. Because of its comprehensiveness and intuition in data acquisition, the machine vision provides the information foundation for intelligent robot. Thus, the study of machine vision is an important branch of intelligent robot.In the research background that mechanical arm is applied in automatic logistic sorting system, this article researched the image segmentation in manipulator vision system, which is based on the manipulator MOTOMAN in robot lab of Xi’an University of Science and Techology. Aiming at the limitations of Otsu method in embedded machine vision system, after the analysis and research of a variety of image segmentation methods and image color model space, we proposed the H-S two-dimensional adaptive threshold segmentation method of artificial immune algorithm in HSV model space, so can improve the image segmentation efficiency of manipulator vision system. The major research in this article as follows:Firstly, we carried research on automatic logistic sorting system, and established the structure model of logistics sorting system based on machine vision. According to the requirements of automatic logistic sorting system, we designed each subsystem, on this basis, a systematic study has been made on manipulator MOTOMAN, based on the analysis of characteristic parameters, motion parameters and the principle of two time development, we provided the control program development scheme by calling the dynamic link library of Motocom32 in integrated development environment Visual C++.Secondly, we studied the manipulator vision system. In the application that using the embedded system as the manipulator vision processor, we presented the H-S two-dimensional adaptive threshold segmentation method by artificial immune algorithm. After the analysis of visual image segmentation method and image color space model, we proposed the method of using modified H-S two-dimensional adaptive threshold segmentation in HSV color space, so it can adapt the characteristics of embedded system. On the basis of Otsu segmentation method, the model of adaptive threshold segmentation method by artificial immune algorithm was established. According to the H-S two-dimensional histogram of manipulator vision system, we presented the H-S two-dimensional adaptive threshold segmentation method by artificial immune algorithm.Finally, the image segmentation method was verified by simulation, which was proposed in this article, and the automatic logistic sorting system was built; By the Matlab simulation, we got the test results of this image segmentation method in embedded vision system; After built the automatic logistic sorting system based on MOTOMAN, it validated the practicability and reliability of image segmentation method further. Through the experiment, the logistic sorting system runs reliably, it shows this research has important reference significance in practical application, and it can provide further value in the research of field of machine vision.