Development of a Data Science Solution for Clay Pigeon Shooting Scoring

Industry

Entertainment, Information Technology

Technologies

AI, C/C++, Computer vision, Python

About Our Client

The Customer is a European IT startup developing software solutions for shooting sports.

Need to Fix Flaws in a Partially Developed Software Product

The Client had partially developed image analysis software for clay pigeon shooting that was to automatically determine shooting results. The software product had certain inaccuracies (e.g., failed to detect target hit/miss) and the Client needed to review the software and fix the issues. In data science since 1989, ScienceSoft was chosen to implement the changes.

________________________________________________________________________

Clay pigeon shooting (clay target shooting) – a shooting sport that involves shooting a firearm at flying targets (clay pigeons). The sport has more than 20 different forms of regulated competition; the most popular of them are trap, skeet, and sporting.

Fixing Inefficiencies and Enhancing the Product

ScienceSoft’s data science team started with the audit of the existing software.

The Client’s software was powered by two convolutional neural networks (CNNs) that classified objects either as a clay target or as background and localized the target’s position. The CNNs had been trained on historical data (shooting data elicited from different periods, weather conditions, locations).

As a result of the assessment, ScienceSoft’s team identified a number of errors which hindered efficient software operation (e.g., synchronization failures, inaccurate target localization, low frame rate).

After we outlined the issues and inefficiencies, we took the following measures to rectify them:

Enhanced the system’s algorithms so that it can distinguish clay targets from the target fragments lying on the ground after previous launches.
Added 2 CNNs (the main CNN and the supplementary one) to enable simultaneous target detection and tracking of multiple targets. The new CNNs didn’t require image cropping and were efficient even when trained on small datasets.
Optimized computation resource consumption by 50%.

After the improvements, the system could function in the real-life outdoor environment and could be used for scoring in different types of shooting games, which involve more than one flying target.

Optimized Software Attracted the Attention of Prominent Sports Organizations

The Client obtained fully functioning software for automated results detection in clay pigeon shooting, which can be employed in major international competitions. The software product is already considered by reputable sports organizations as a solution for clay pigeon shooting competitions.