关键词: 计算机图形学, 化工污水, 流量检测, 数学模型, Laplace算子, 准确度, 识别平均时间

Abstract:

The main problem of sewage treatment of chemical industry along the river is that the traditional methods can not effectively detect the flow and other sewage information due to the oil-bearing characteristics of chemical sewage.This paper proposes a non-contact, low-cost, high stability sewage flow detection method. The research content includes three parts: sewage flow expression in computer graphics, sewage flow detection model and water target identification algorithm. Construct a double coordinate system to establish the mathematical expression of sewage flow and camera image in computer graphics and prove the feasibility of sewage flow detection, build a mathematical model between sewage flow and water area, and use the difference of texture features of water objects and walls and other objects under natural light to realize the identification of sewage water objects in natural images; The actual sewage flow is calculated by using the constructed mathematical model and the identification results of water body. Based on this method, a robust wastewater flow detection system is developed and tested in real scene. The results indicate that: 1) the system can still get a high correct rate more than 95% by shooting images in different weather; 2) the non-contact detection mode greatly reduces the contact between the equipment and the chemical wastewater, and increases the service life of the equipment; 3) under the current experimental hardware configuration, the average time for the system to recognize a single image is less than 100 ms.

Key words: computer graphics, chemical sewage, flow detection, mathematical model, Laplacian, accuracy, average identification time