CMSC818W: Introduction to IoT & Mobile Computing (Semester Project)
In this report, we propose a lightweight liquid testing system for categorizing different liquids and detecting adulterants by analysis through signal passing through liquids. Because different liquids have very different chemistry compositions, the light transmitted and scattered through the liquids will usually produce noticeably spectral signatures. By passing light of different wavelength through different types of liquids, our framework is able to measure the attenuation in the signals created by light transmitting and scattering, and then do analysis on the measurement by training neural network models for characterizing and therefore achieving liquid detection for future liquid inputs. Our system has four important advantages: the system requires very few customized hardwares and hardwares being used are very affordable. Also, our system is highly generalized and can be possibly used for additional liquids with proper training.Moreover our system requires no direct contact to measure the signal attenuation.Lastly, our system provides an identification of these liquids with decent accuracy.
├── Cicuit Schematics
│ ├── All schematic files
│ ├── 818W_Code_arduino
│ │ ├── 818W_Code.ino
│ └── Adafruit_SI1145
│ │ └── all library files
├── Data
│ ├── Data.zip
│ ├── Data_1.zip
│ ├── Data 2.zip
│ ├── Data 3.zip
│ └── All data files extracted are here
├── DNN_Model_&_Scripts
│ ├── 818W_Project_Classification.ipynb
│ └── serial_comm_data_collection.py
├── Reference_Papers
│ └── All reference papers are here
├── Report_&_Presentation
│ ├── 818W_Project.pdf
│ └── Identifying Liquids through Radiometric Effect and Deep Networks.pptx
├── Results_Distributions
│ └── Heatmaps & data distribution charts are here
For more details, please find:
- Project Report here!