Machine-Learning-Application

This repositories is made for machine learning part of SoilIt development.

Collect Data for Dataset

The data is collected by using Bing Image Downloader library. Each class get 120 picture collected.
Code example of 'Tanah Laterit':

Preprocessing Data

To remove undesired data, the data of soil image and plant are processed. Panda processes the Plant dataset, whereas Tensorflow processes the Image dataset.

Training model

The model is created with mobilenetv2. Then it converted by tensorflow lite for deployment to android.

TF Lite

It consists of Soil Class labels and the model for deployment.