基於光學技術和影像處理應用於泡沫污染物監測與分析

Abstract

以往污水處理廠水樣發生濃度過高的時候，再進行水質的檢測時，往往都要將水樣抽載並送檢，如此將花費大量心力，因此我們透過污水集濃產生的泡沫污染物，採用光學和影像技術應用於泡沫污染物型態智慧分析，以自動化視覺感知，進一步增進處理水污染檢測之效率和精準度。 本研究目標在開發一種高效的泡沫識別技術，通過使用影像辨識技術結合光學儀器輔助，以非接觸的方式來檢測水中的泡沫濃度和種類，實現對水中泡沫的自動監測和分析。樣品方面選擇了在污水中容易起泡的種類，生物性蛋白代表牛血清蛋白和界面活性劑代表十二烷基苯磺酸鹽，並去探討了其泡沫生成型態的變化。 在光學檢測部分，對其成份做基本分析，獲得吸收光譜並進行譜線特徵的比較，利用飛時測距儀器去檢測其液面泡沫堆疊高度。在影像處理部分，通過相機擷取泡沫特徵後，藉由灰度化、銳化、對比度強化、中值濾波等一系列影像處理的方法來分析擷取之泡沫影像，最終進行二值化處理，計算並框選泡沫覆蓋面積。 研究結果顯示，牛血清蛋白泡沫較十二烷基苯磺酸鹽泡沫更容易向上堆疊，而十二烷基苯磺酸鹽泡沫較牛血清蛋白泡沫容易向水平面擴散，這是因為蛋白質在界面處會吸附並與靜電、疏水鍵和共價鍵相互作用，導致泡沫形成多層膜並減緩擴散速率，而界面活性劑泡沫因為其親水端的特性，界面活性劑泡沫在水中擴散速度會較快。 這些技術的開發將為廢水處理廠操作人員提供寶貴的訊息，使他們能夠即時採取措施，有效處理持久性泡沫，從而提高系統的運行效率。

In the past, when wastewater treatment plant samples showed excessively high concentrations, water quality testing often required the extraction and transportation of samples for analysis, entailing significant effort. To deal with this challenge, we have leveraged the foam pollutants generated during wastewater concentration, employing optical and imaging technologies for intelligent analysis of foam pollutant patterns. This approach introduces automated visual perception, thereby enhancing the efficiency and precision of water pollution detection. The primary aim of this study is to develop an efficient foam identification technique. We achieve this by combining image recognition techniques with optical instrument assistance to non-invasively detect foam concentration and types in water. We accomplish automated monitoring and analysis of foam in water through this approach. For sample selection, we have chosen easily foaming substances typically found in wastewater - bovine serum albumin (BSA) representing biological proteins, and sodium dodecyl benzene sulfonate (SDBS) representing surfactants. We further investigate the variations in their foam generation characteristics. In the optical inspection phase, we conduct fundamental analysis of the foam components, obtaining absorption spectra and comparing spectral features. Additionally, we utilize time-of-flight (ToF) distance measurement instruments to assess the stacking height of liquid surface foam. In the image processing phase, we capture foam characteristics using a camera. Subsequently, through a series of image processing techniques such as grayscale conversion, sharpening, contrast enhancement, and median filtering, we analyze the captured foam images. This culminates in binary processing to calculate and delineate the foam coverage area. Research outcomes reveal that foam generated by bovine serum albumin tends to stack upward, while foam produced by sodium dodecyl benzene sulfonate exhibits a greater tendency to spread horizontally. This disparity arises from the interaction of proteins at interfaces, involving electrostatic, hydrophobic, and covalent bonds, resulting in the formation of multi-layered membranes that slow down diffusion rates. Conversely, surfactant foam, due to its hydrophilic end, exhibits faster diffusion in water. The development of these techniques furnishes wastewater treatment plant operators with valuable insights, enabling them to promptly implement measures and effectively address persistent foam issues, thereby enhancing overall system operational efficiency.