| Summary: | 碩士 === 國立臺南大學 === 電機工程學系碩博士班 === 107 === In recent years, the issue of food safety has been frequently valued and discussed. How to confirm food safety and non-toxicity is the key to protect human health. In this paper, a special kind of fluorescent proteins are to test whether the food sample is toxic or not .This is because the above-mentioned fluorescent expression, toxic substances can be added. The feature can be further exploited to determine the level of toxicity of the tested food sample by checking the fluorescence intensity. Since the current professional biosensor is expensive and most of them are not portable, the detection process is quite inconvenient and time consuming. This paper proposes to integrate mobile devices spectral analysis, IoTtalk (an IoT platform) and special fluorescent proteins to develop a portable toxic compound screening system. Users use their personal smart phones to capture the spectral power distribution of special fluorescent proteins which is mixed with food samples. Through IoTtalk platform, the pictures are sent to the back-end toxic compound screening server. The server analysis and determines whether the food sample is toxic or not and then return the result to users. In order to realize the screening system for toxic compounds and get reliable detection result, this paper proposes a two-stage learning and analysis method to accomplish spectral analysis and toxic compound detection. In the first stage, spectral power distribution images of different fluorescence concentration samples are captured by the smart phone and then sent to the screening server for analysis and learning. After the learning stage the second stage uses the learning result to identify and detect the fluorescent performance of the testing object. This article proposes to integrating mobile phone with spectrum analysis and fluorescent protein to accomplish a portable detection system for toxic compounds, which allows users to do inspections at any time and any place. A prototyping system is shown in the paper. Experimental results show that the system is easy to operate and can get reliable results even when the fluorescent protein solution is mixed with colors because of food.
|
|---|
