水处理过程中图像识别方法的应用与实现

发布时间：2018-04-21 07:33

本文选题：絮凝体 + 图像识别　；参考：《北京化工大学》2017年硕士论文

【摘要】：洗煤废水具有浊度高(大于1000)、强负电(-28mv)、高稳定性的特点。洗煤废水的处理循环利用,能够降低采矿行业工厂的用水成本,是工厂节能环节的重要组成部分。絮凝反应是通过向废水中投入絮凝剂,形成絮凝体的过程,是废水处理的核心环节。将图像识别技术与废水处理中絮凝反应控制环节相结合,提高反应效果,保证系统稳定性是目前需要研究的技术方案。文章根据现有废水处理中絮凝反应所需要的硬件和控制流程,提出基于絮凝体图像和废水流量检测双反馈信号的控制算法。对絮凝体图像采集系统的硬件设计、硬件选型与电路实现做了介绍。对前端采集的絮凝体图像数据的转换、絮凝体图像数据的处理算法进行了优化和实验验证。对絮凝体原始图像的降噪处理算法、絮凝体灰度图像的二值化处理算法、絮凝体二值图像的降噪处理算法、絮凝体图像中独立目标物体的标记算法、独立目标物体的边缘检测算法和目标物体的等效直径获取算法进行了算法优化与验证。文章结合实际采煤工厂的洗煤废水,在实验室内环境下搭建絮凝反应平台,对文章中提出的图像采集、絮凝体图像的识别与目标物体的特征提取进行实验,基本实现了水处理过程中图像识别的应用。
[Abstract]:The coal washing wastewater has the characteristics of high turbidity (> 1000 渭 m), strong negative charge (-28 MV) and high stability. The treatment and recycling of coal washing wastewater can reduce the cost of water in the mining industry, and it is an important part of energy saving in the plant. Flocculation reaction is the process of forming flocculant by injecting flocculant into wastewater and is the core link of wastewater treatment. Combining the image recognition technology with the flocculation reaction control link in wastewater treatment to improve the reaction effect and ensure the stability of the system is the technical scheme that needs to be studied at present. According to the hardware and control flow of flocculation reaction in wastewater treatment, a new control algorithm based on double feedback signals of flocculant image and wastewater flow detection is proposed in this paper. The hardware design, hardware selection and circuit realization of the flocculant image acquisition system are introduced. The conversion of flocculant image data and the processing algorithm of flocculant image data are optimized and verified by experiments. The denoising algorithm of flocculant original image, the binarization processing algorithm of flocculant gray image, the denoising processing algorithm of flocculant binary image, the marking algorithm of independent object in flocculant image are studied. The edge detection algorithm of independent target object and the equivalent diameter acquisition algorithm of target object are optimized and validated. Combined with coal washing wastewater from actual coal mining plant, a flocculation reaction platform was set up in the laboratory environment. Experiments were carried out on the image collection, the recognition of flocculation image and the feature extraction of the target object proposed in the paper. The application of image recognition in the process of water treatment is basically realized.
【学位授予单位】：北京化工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X752;TP391.41

【相似文献】