Artificial Intelligence for Indian Transport Infrastructure
We attempt to solve this problem by providing a framework based on image recognition and intelligent classification of images both static and moving. We start by building a Machine Learning [ML] model using labelled images. Our plan is to use GPU based TensorFlow on Azure Cloud services to build the ML model. Once the trained model is deployed, any image when fed to the model will identify the respective anomalies on the road. This will help/guide the concerned authority to take actions on anomalies. Also as explained in the attached presentation the model output consists of road health status with the worst roads displayed on top. The tool will guide the authorities to take action on priority basis.
https://drive.google.com/open?id=137LFud5o8LVDKNRMERR1Q-AHmx_X7oOJ
@misc{tensorflow, title={Tensorflow Object Detection API}, author={roffel, pkulzc}, year={2018}, publisher={Github}, journal={GitHub repository}, howpublished={\url{https://github.com/tensorflow/models/tree/master/research/object_detection}}, }
Reference,original code, : https://github.com/tensorflow/models/tree/master/research/object_detection This POC is based on Object Detection using Google's Tensorflow Object detection API.