Plant Identification using Machine Learning

Identifying plants just by looking at them is difficult. You need to study every plant in order to identify them which is a tedious job and requires a lot of time investment. But what if you can just grab your smartphone and capture the photo of any plant and machine learning does your job for identifying that particular plant that you encountered? We used TensorFlow by google which provides important functionality of Machine Learning for creating image classification models and deploying them to embedded devices. It contains image processing that helps to identify the surrounding by comparing it with large datasets and train itself to increase its accuracy. A dataset is needed first which will contain tens of thousands of images of different plants taken from different angles. We use Mobilenet architecture with convolutional neural networks for this purpose. The trained model is then deployed to an android platform with the help of TensorFlow lite.