Dirt With Flask: Image Processing for Soil Color

Civil and geotechnical engineers benefit greatly from an image-processing tool that is field-deployable. This study will discuss the development process of creating an image processing application that can find the most prominent colors in an object and display the results in both RGB and Munsell Color System. The application is built with React Native as its frontend, Python as its backend, and Flask for integration. The application uses a Median Cut color quantization algorithm to find the most prominent color and uses a Munsell Renotation Inversion algorithm to convert to the Munsell Color System. Using the Munsell Soil Color Chart and soil samples from the field, the application's image processing components are tested to have great accuracy. In the future, the applications' functionalities should be expanded and the potential usage of the phone's flashlight as a controllable external lighting should be explored.