Development of Portable Photo Spectrometer for Classification of Biomedical Fluids using Neural Network

Abstract

Spectrometer is one of the powerful technologies in testing, discovering, and measuring spectral components of different fluids. It is commonly used is the field of Physics, Material science, Chemistry and Biochemistry. Different fluids such as blood, milk, petrol, and Human body fluids can be classified using the spectrometer on the base of the light absorbance of different wavelengths. Commercially various smart spectrometers are available for different fluids classification, but no one is classifying the fluids at real time having such a low price and high accuracy. In this study we have manufactured a light weight, portable, cost effective and standalone Artificial Intelligence based spectrometer device for the fluid classification. The device uses small quantity of the fluid as 3 milliliters in a test tube and predict the results by pressing the push button. The fluids such as milk and petrol were tested by the device and classified with 96% and 97% accuracy respectively. Initially data samples of different quality of fluids were taken using the light sources and camera, and the data was randomized and divided into testing and training data. After that Neural Network is trained to predict the results and microcontrollers are used for model deployment. The casing of the device was designed, and 3D printed having the size of 70 x 70 x 45 mm and weight of 200 grams. An OLED is used to see the results on the screen. The device can be used for the quality analysis for different fluids. It is easy to use, rechargeable, accurate and cost effective as compared to other spectrometers.