Research On Level Detection System Of Spillway Launder Based On Machine Vision

The demand gap of domestic potash fertilizer is big and the resources reserves of soluble potash are comparatively rare, while the existing process route has been lagging behind in foreign technology. This paper deeply understands the important process parameters of the potash production, improves the concentrate grade of potash and reduces the potassium waste recovery. Aiming at the existing measurement methods of the spill launder level detection’s failure detection accuracy and fault prone to problems, this article further study the camera calibration and level regional fluctuations’locating and tracking, and several other aspects of the reverse floatation’image measurement system. Meanwhile, it also designs a set of reverse flotation liquid level measurement system based on machine vision, to realize high accuracy fluctuant reverse floatation liquid level detection and real-time tracking. The research mainly includes the following aspects:According to the working principle of machine vision measuring system and accuracy requirement of reverse floatation liquid level measurement, selecting the best matching optical lens, CCD camera and other hardware parts put up an experimental simulation platform. Then, it achieves the whole overflow liquid level measurement system design and software modularization design.In order to improve the detection accuracy of the overflow tank liquid level, the distortion parameters of industrial camera model should be taken an in-depth study, and the two-step Tsai calibration algorithm with circular target plate achieve high-precision camera calibration.The complex detection scenes of reverse flotation solution are large differences in level and bubble interference. Firstly, several common region segmentation algorithms have been study. Choosing OSTU plus redundancy threshold as adaptive threshold and Surendra background modeling locate the level fluctuations positioning area.According to the complexity of spillway launder’s different areas, firstly, this paper introduces several common surface extraction algorithms. Secondly, it uses the adaptive background modeling to locate the level fluctuation’s region, then uses LSD edge detection algorithm to extract different regions’liquid level and fit the optimal interface between the slurry and bubble. Finally, this paper research the image-liquid measurement system’s detection precision and error analysis. In the last part of the paper, the system is realized by C++programming, and simulation tests are done to the spillway launder’s fluctuant liquid level. The experimental results show that the scheme design in this paper can achieve the real-time liquid level measurement with high precision.