A glimpse of how Computer vision is used in Autonomous driving systems

Computer vision is the field of machine learning where we are able to train a model to classify, cluster or predict output using images as the input data. Using a deep learning technique called Convolutional Neural Network(CNN) is one of the most widely used methods in computer vision.

In an autonomous vehicle, the general pipeline in making decisions is by collecting data/telemetry using sensors, cameras and radars, processing them using real time systems implementing cutting edge analytics, machine learning , and data processing, and finally, a system which can use the predictions and insights into actionable decisions.

In particular, the field of computer vision is of keen interest in the autonomous vehicle world. Most of the decisions are based on image/video data. Computer vision along with deep learning models help the driver system to process data and convert them into decisions. Some of the systems which use computer vision in a vehicle are crash avoidance systems, auto pilot/self drive systems, pedestrian detection systems etc.

This type of model helps the driving system to make real time decisions based on a pre-trained model. Further there is a possibility to capture interesting scenarios and retrain the model.