Research And Implementation On The Industrial Robot Vision Assembly System Based On IoT

With the implementation of reform and opening policy and the development of economic, decades’ high-speed economic development makes China’s demographic dividend becoming vanished. And to replace those labor working with simple action by the industrial robot is considered as the most suitable way, because robots can work continually for 24 hours a day without employee benefits, self-assurance, injury, social insurance, and so on. Besides, industrial robots are more adapted to work in a harsh, harmful environment. Therefore, the study of industrial robotic assembly system, to allow it to complete a variety of production tasks better, and, to meet the expansion and upgrading of various types of labor-intensive enterprises’ needs, is not only able to improve production efficiency, reduce work-related injuries and alleviate the burden on business costs, also has great practical significance for the promotion of Chinese social development and economic construction.Currently, in the world, traditional robotic assembly system and technology, teaching and playback, are applied in industrial processes with a large scale. But this is obviously not suitable for the current era of rapid updating of product categories. As each change in product needs all equipment necessary to be reconfigured in the whole production line in a traditional assembly system, resulting in a tremendous waste of resources. Therefore, a stronger ability of environmental awareness and data processing is needed for industrial robots assembly system. Only in that way, industrial robot can truly liberate mankind in the heavy work of manufacturing.As the visual system is the main source of information for human beings to perceive the world and obtain knowledge, which points out the orientation of research for this thesis to realize industrial robot assembly system and related technology. In this thesis, the study of industrial robot vision system is mainly based on traditional stereo vision system, including the study of recognition algorithm for production parts, and the study of key steps to build stereo vision system, camera calibration and stereo matching. Bringing in visual selective attention mechanism is to simulate the human visual system to quickly deal with critical visual information, and finally, a novel parts’ recognition model based on HC(Histogrambased Contrast) saliency algorithm model and BP(Back Propagation) neural network is proposed and implemented. At the same time, through the analysis of various types of stereo matching algorithm, a novel matching algorithm based on the integration of visual attention characteristics matching and regional matching is proposed, which combines the robust property of feature matching algorithm and the matching density property of region matching algorithm. At last, the efficiency of a successful match to the target is improved.Moreover, as human beings get the data via the Internet, the industrial robot also needs "surfing the Internet ". Production-related data can be obtained by robots in real-time through IoT(Internet of Things). In this thesis, RFID(Radio Frequency Identification) technology is the main source of components’ real-time distribution information for the IoT platform. In order to fulfill customers’ needs, components and industrial robotic assembly system being interconnected, the communication gateway to access the Internet is built with the embedded system and mobile communication network. Information management center is the core part for data processing. Assembly tasks are distributed by the center based on the integration and processing of various information from all parts, which allows to reasonably allocate production resources. Finally, the feasibility and the effectiveness of the industrial robot vision assembly system based on IoT to assemble LED lamp are demonstrated through several experiments in the last parts of this thesis.