Flotation Level Measurement System Based On The Binocular Stereo Vision Technology

With the rapid development of computer technology and machine vision, stereo vision technology has been more and more applied to the reconstruction of three-dimensional geometry information, this paper in-depth understand the low measurement accuracy and poor anti-interference problem of the existing traditional liquid level detection technology in the practical application. This paper studies level detection technology of reverse flotation tank in potash production process using the binocular vision technology and design the binocular vision technology based reverse flotation tank level detection system. This paper realizes the camera calibration and the high-precision real-time measurement of reverse flotation liquid level, the research mainly includes the following aspects:(1This paper in-depth study measuring principle of liquid level detection technology based on binocular stereo visual, according to the working principle of the measuring system and the accuracy requirements of reverse flotation liquid level, selecting hardware of best matching lianr, CCD camera and image acquisition card simulation to build the laboratory platform, realize overall scheme design of reverse flotation tank liquid level detection system and modular software process design of level detection.(2) This paper research the transformation of camera coordinate, image coordinate and world coordinate system in the model for stereo vision. The theoretical calculation is made to find internal and external parameters of the camera calibration. This paper discusses the distortion parameters and its correction of the stereo vision nonlinear model, and put forward the calibration method of the camera calibration.(3) This paper presents some basic assumptions and basic constraints which stereo matching need to follow; in-depth studies the image matching algorithm based on gray scale, and studies how to combine the epipolar geometry constraint and image matching algorithm based on gray correlation apply to image feature matching.(4) This paper study the image preprocessing technology which is needed in extraction of image features; mainly research and use the image preprocessing means of image enhancement, filtering, gray and feature point extraction; introduces the transformation of the space point coordinates in the camera coordinate system and the world coordinate system.Finally, the based on Halcon plane calibration method realize camera calibration, the calibration results were analyzed and summarized; the epipolar geometry constraint and gray correlation image matching algorithm is applied to reverse flotation tank liquid level detection system. Based on the measured 10 group level results, the paper makes a detailed analysis of the error distribution and accuracy.