The business goal
Our client is one of the largest sporting goods retailers in the world. They intend to mix augmented reality and Computer Vision techniques in order to create a new product in the sporting goods field.
The goal of the project is to create a CV application for an augmented reality version of the darts game. Using a real dartboard, darts, and a video camera, a user can create a virtual game that analyzes the dartboard and dart positions and calculates the scores. Initially, the camera is oriented to the dart board without darts (start of the game) and then players start throwing darts to the dartboard and after each throw the system detects the segment of the dart board where the dart sticks and calculates the score gained by the player. The process of the dartboard segmentation, darts detection, and score management is shown over the video by the camera.
We use the latest advancements on a technical side in the Computer Vision and Deep Learning and engaged OpenCV and TensorFlow libraries written in C++ and Python respectively.
As a result, our client got a working POC with high accuracy. The corporation plans to use this for further development to make it a commercial product in official competitions.