Research On The Key Of Preliminary Processing System And Technic For Wolframit Based On Machine Vision

As the machine vision inspection is more and more widly used in industry, it’s urgent to improve the preliminary processing method on on wolframite. To solve the problem of the selection by person, a preliminary processing system on wolframite based on the DSP for the preliminary processing craft on wolframite is built in this paper. Combined with machine vision, a preliminary processing algorithm on wolframite is designed on the basis of invert selection. In the paper, this algorithm is also transplanted into the DSP object system and the technical requirements related to the preliminary process are listed.The main contributions of this paper are in the following aspects:(1) A brief summarize of wolframite primary process and the application situation of machine vision in the industry. To show the strong advantages of machine vision compared to human eyes in the primaries on wolframite process by analyzing the the unique texture features and color features of black tungsten.(2) According to the development of DSP, combined with the requirements of the high performance to price ratio in industry, select a DSP chip which is suitable for wolframite primary system, and the chip component of hardware interface circuit are briefly described. Then it will build a DSP object system which is competent in industry and providing the corresponding technical target and development process.(3) The algorithm was set with the combination of hardware and process. Firstly, ore image motion image restoration. A pixel labeled algorithm based on HSV color space was presented on the optical and texture features of the wolframite ore. The algorithm converted the color space of the original ore image from RGB to HSV and binarized it. On the basis of pixel labeled algorithm in connected domain the target ore could be detected by the invert selection. Extensive literature research shows that the mainstream edge extraction algorithm contain Sobel operator, Robert operator and Laplpcian operatoe. The experimental results prove that the system can not only efficiently identify the correct object ore but also reduce the cost. It is practical and of high value to the real-time preliminary processing system. There will finally be a way to optimize the algorithm with the help of the DSP object system.(4) Based on the algorithm and the DSP object system, study each part of the process technology of black tungsten primary system, such as the gravel control, belt color control, the the conveyor belt speed control on the basis of the processing time in per imagine, ore image acquisition environment and acquisition time control, gripper choosing amount control and so on.Theoretical study and practical experiments have proved that the system can not only efficiently identify the correct object ore but also reduce the cost. It’s of high pertinence,stability and accuracy rate which reached 93.2%. The algorithm applicated in the target system will be translated into linear assembly language from C code in this paper and by this kind of transformation, it’s speed has dropped from 494ms to 42ms. In conclusion, the primary sorting system applied with machine vision for the wolframite ore is practical and of high value to the real-time preliminary processing system.