Soybean Seeds Visual Classification System Based On DSP And ARM

The rapid development of modern information technology makes real-time image processing attracted wider attention. In recent years, the continuous development of DSP technology, the theoretical research in the field of digital signal processing applied to the actual system, and promote the development of new theories and applications, technological developments in the field of image processing has also played a very important role. Meanwhile, with the rapid development of Internet, mobile communications and multimedia technologies, embedded systems have also been more widely used. This paper analyzes the latest developments in embedded systems, DSP, real-time image processing and other fields, and combined with the laboratory results of previous studies and practical experience in the development of image processing systems (especially the license plate recognition system) is proposed based TMS320DM6437 (DSP) and the overall design scheme TMS320DM355 (ARM) soybean seed embedded visual grading system and designing an embedded real-time image processing system based on DSP + ARM. The paper reviewed photovoltaic grid system islanding phenomenon and resulting hazards, expounded the phenomenon occurrence mechanisms, islanding detection standards, islanding detection models, in-depth analyzed and discussed the existing islanding detection methods, focused on the advantages, disadvantages and range of applications of passive and active detections based on inverter side. Introduced the effectiveness evaluations of detection methods in various coordinate systems.Currently, the use of machine vision technology to detect the quality of food and agricultural products is gradually popularized. Soybeans as an important agricultural products, increasing attention has been paid, but the current is detected on soybean quality, grading remained at artificial levels. Therefore, the application of computer vision technology to detect the appearance of grain crop characteristics and identify the species, the level of automation to improve agricultural production and agricultural added value, strengthen our competitive advantage in the international agricultural market has important significance. This paper describes the working principle of the system, hardware configuration, software systems and grading tests. Soybean image set, after extraction of the beans background segmentation parameters, the use of statistical methods for beans region boundary feature, feature extraction area determining roundness and smoothness of optimal classification signatures. The system will speed DSP and StrongARM processors have unique advantages in communications, networking and real-time control, embedded real-time environment for the next several complex algorithms implementation issues opened up a new way. In TMS320DM6437 and TMS320DM355 DaVinci technology processors as the core processing unit, embedded image processing algorithms to achieve visual grading soybean seed. Select four different varieties of soybean in 2000 as a test sample, the system repeat the test, grading screening accuracy of 95%. Achieve a good classification results, and to solve the low efficiency of conventional soybean grain grading system, major damage shortcomings.In subsequent chapters, the basis of the actual needs of this paper on the subject, combining DSP and embedded system development process, describes the system functions and the overall design scheme, which covers the system hardware design, software platform and related development environment and tools and image algorithm design and grading tests.Finally, he summarized the achievements and shortcomings of the Institute made the subject of proposed topics for further research and application prospect outlook.