Picking Robot Image Of The Stream Processors To Speed Up The System

In order to improve production efficiency and make farmers get rid of simple and heavy labor, we need to develop picking robot of high-speed and high precision to solve these problems. The principle of harvesting robot is the use of vision technology and mechanical arm robot composed of visual servo control system for target identification and picking. However, because of the actual picking operation mostly in the unstructured environment and job object affected by various factors, it is easy to produce irregular movement so that picking the precision and accuracy rate are low. To solve this problem, using the stream processor constructs image acceleration system for farm harvesting robot. The stream processors can speed up large amount of image data processing. Not only speeds picking robot targets on processing speed, but also improve the efficiency of picking and accuracy.In summary, the image acceleration system of the picking robot can improve the rate of harvesting and avoid the staff repetitive picking. To more in-depth research in this area has important theoretical and practical significance.The main content and contributions of this thesis are summarized as follows:1. Introduces the research background, purpose and significance. It analysis in detail current development status of agricultural picking robot and the stream processors, and discusses the major problems. Finally, the main content and the technical routes are presented.2. This thesis analysis architecture characteristic of the harvesting robot,and is analyzed the robot system composed of the main control unit, the amount of information input unit, and the robot executing agencies. It progresses image processing part of the detailed structural analysis, and discusses the issue of the existence of massive information, the computer control unit architecture issues as well as image processing methods of these three aspects. Finally, according to the three questions it presents a set of speed picking robot image system solutions based on the stream processor. 3. First of all, this thesis discusses handle ideas of stream processors, and describes in detail with the calculation of its decomposition, decoupling computation and data access, data sub-block processing, explicit communication, and parallel processing.Constructed a stream processor hardware architecture model can be better prepared for the algorithm flow overtaken. According to chapter II, section 4 it provides stream-based processors solutions to execute the image speed picking robot system design, and designs the new system model as well as describes its working principle and system functionality. For SP16HP-G220 stream digital signal processor it progresses functional description. Finally, core instruction set provided the calculation of SPI and the Pentium family of processors instruction set is compared, and the algorithm streaming for the next chapter provides quantitative data support.4. After the stream processor hardware system is completed, it is the preparation of the system software. First we must understand the stream of applications and basic concepts of programming stream, so as to understand the data stream in the process, and we can better organize the data stream. This thesis summarizes the mapping method of stream applications, and for the four steps processes detailed description and characteristics. According to this theory it sets up the stream processor programming model, and for convection-level procedures and the core-level process model process the structural characteristics of the model. According to the second and third chapters designed picking robot of speed up the system image as the application platform, it processes the parallel algorithm analysis and Algorithm fluid based on the actual needs of selected representative image processing gray-scale transformation and target recognition systems to the Sobel edge detection operator.5. According to the previous chapter established stream-based image processor speed picking robot processes simulation test related. Used streaming through the fourth chapter of the gray transform algorithm with Sobel operator, and Chapter III Section 4 of the data in Table 3.1 establish the simulation model, after the end of the simulation test results are analyzed and discussed. For the simulation results show the problems and make corresponding adjustments to the model so that the established model of stream processor is better. By summing up the experience and the system design, a set of image accelerating system of the picking robot based on stream processor.Although the major research project has been completed in this thesis, these still exists much research works to be done to make the agriculture robot system suitable for practical application. At the end of this thesis, the author makes a summary of the main research works and contributions, and give some suggests for further research.