基于工业视觉的水体表面污染监测系统

Water surface pollution monitoring system based on Industrial Vision

摘要:

针对工矿企业突发性水污染事故具有明显悬浮物或明显颜色变化，且多数污染可通过表层水体视觉方式直观判断等特点，本文基于突发性水污染事故监测需求分析，研究了一种基于工业视觉的水体表面污染监测系统。针对连续多帧图像信息受水流流速、颗粒悬浮物和系统噪声等影响问题，提出了随机扰动信号滤波方法和基于颜色信息评价的多级图像信息调节方法，进而采用优化神经网络建立污染分析模型，并开发了监测装置对水体图像进行实时分析，实现污染状态快速判断。在某冶炼企业应急站投用该系统后，降低了劳动强度，缩短判断时间，提高判断准确度，降低环保事故发生率、 减少岗位人员数量、降低成本，取得了显著的应用效果。

Abstract:

Sudden water pollution accidents in industrial and mining enterprises have obvious suspended solids or obvious color changes, and most pollution can be judged intuitively through the visual way of surface water, based on the monitoring demand analysis of sudden water pollution accidents, a water surface pollution monitoring system based on industrial vision is studied in this paper. Aiming at the problem that continuous multi frame image information is affected by water flow velocity, particle suspended solids and system noise, a random disturbance signal filtering method and a multi-level image information adjustment method based on color information evaluation are proposed. Then, the pollution analysis model is established by using optimized neural network, and a monitoring device is developed to analyze the water image in real time to realize the rapid judgment of pollution state. After the system is used in the emergency station of a smelting enterprise, it reduces the labor intensity, shortens the judgment time, improves the judgment accuracy, reduces the incidence of environmental protection accidents, reduces the number of post personnel and reduces the cost, and has achieved remarkable application results.