Object Detection and Automatic Number / License Plate Recognition (ANPR)
Xen.ai is developing an Artificial Intelligence based Object Detection and Automatic Number/License Plate Recognition (ANPR / ALPR) service which could automatically detect objects such as a car license plate in a real time video stream and also be able to capture the registration number of all the cars present in the scene. This application could play an important role in intelligent transportation systems. The business application could vary from surveillance to traffic control. Before deployment of this application as a service Xen team has built the whole algorithm using Deep neural Nets, trained and tested it on a very real sophisticated imaging condition.
About the client
Engineering, Research and Consulting, Virginia U.S.
Client provides engineering oversight to acquisition programs such as next-generation space vehicles and weapons systems, including the development and operation of advanced test systems.
Client wants to Automate the License Plate Record System at scale to save the manual labor cost
Xen.AI developed a cloud based web platform which runs deep learning based Optical Character Recognizer at the backend core.
IR camera sensor
Google’s Tensorflow as ML framework, OpenCV for Image Processing and Spark as Big Data Engine.
IR camera sends the data in the form of images to the ML engine which recognizes the registration number of the vehicle and save it in the Database. We have built the cloud based ML Engine data pipeline process the data and neural network extracts the spatial features from the image and does the prediction using those features. The Characters Recognised will then passed to the loss a function which finds the relation between the characters for better accuracy of number sequence. The predicted result will be stored to the database and sends the report to the client.