Author: Hao Fu
Fu, Hao, 2019 Feasibility study on developing embedded system for PFAS detection in drinking water, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
In this article, I explain the difficulty of PFAS detection using HPLC-MS, including a long detection time and expensive inspection. Alternatively, image processing is used to collect images of and a coloration reaction of PFOA, which is much simple and easier than HPLC-MS. However, the accuracy of the mobile image recognition system has limitations. That is because that different mobile phone cameras will generate different RGB values. Furthermore, after the collection of RGB, there needs to design an appropriate image interpolation algorithm towards the image recognition system. In my project, the embedded system is used to drive the camera module. The OV2640 is a 1/4" CMOS UXGA (1632*1232) image sensor from OmniVision. The sensor is small and low in operating voltage. Users have complete control over image quality, data format and transmission. For example, RGB565 format, YUV422 format or Bayer format. All image processing functions including Brightness and contrast can be programmed via the SCCB interface. The output pixels data transformation is passing through the DCMI (Digital camera Interface). Firstly, analyse image acquisition results for different brightness according to image output formats RGB565 and YUV422. Through the results, it can be found that the best image quality can be obtained when the brightness of the camera module is adjusted to the maximum using the internal registers of the camera module. Secondly, the accuracy of the whole recognition system is compared by linear algorithm, exponential algorithm and polynomial algorithm. Firstly, the influence of different colours on the whole solution is analysed by calculating the coefficient of determination, and then selected the function expression that best matches the data. The results of the image recognition system show that polynomial algorithm has the highest recognition. Linear algorithm has the lowest recognition. By using PFAS detection colour recognition with embedded systems, the accuracy of PFAS solution detection can be effectively improved, from 30 ppb of mobile phone recognition system to 5 ppb of embedded system, which can effectively reduce development time and development cost.
Keywords: PFAS detection,Image Processing,Embedded system,STM32F407,OV2640,RGB565 and YUV422 format, Algorithm design,linear algorithm,Exponential algorithm,Polynomial algorithm
Subject: Engineering thesis
Thesis type: Masters
Completed: 2019
School: College of Science and Engineering
Supervisor: Professor Youhong Tang