We have recently successfully led and completed a research project funded by InnovateUK to develop a new technique for training machine learning models for the detection of objects and events in CCTV images.
To complete the project Zest Consult, Ltd. partnered with:
Costain - A UK leading tier-1 construction & smart infrastructure company. 3D models Collection & Video feeds. System development, integration, and evaluation
University of West of England, Bristol (UWE)-A business-facing university. Synthetic data generation, including various neural networks models, and computer vision system development
We brought our considerable expertise in CCTV and video to develop new CCTV Video analytics for health and safety, security, and productivity use-cases in the construction industry. Leveraging our relationships with UWE, we brought the latest cutting-edge techniques in artificial intelligence to develop methods that are far less expensive, much quicker, and create more robust video analytic machine learning models than currently exist in the market.
The COVID-19 contingency has highlighted the urgent need for site monitoring systems that enable measuring relative distances, effectively and accurately, among workers and plant equipment to ensure safe working conditions.
For example, construction sites and manufacturing facilities have been shut down indefinitely due to large numbers of contagions among workers. Government guidelines have been useful, but they are not sufficient. This underlines the need for accurate and inexpensive monitor systems that contribute to provide safe working conditions while maximizing throughput and productivity.
Common camera-based site monitoring solutions focus on detecting damage due to crime and environmental hazards, such as intruders and fires. Other more capable systems enable limited object and change detection. Monitoring approaches that make use of Bluetooth devices as mobile phones and wearables have been proposed as well, but they reliability has not been proven yet (e.g., mobile tracking apps). Moreover, they require additional equipment, which increases costs and makes adoption more difficult.
We developed an innovative approach to generate the datasets necessary to enhance camera-based monitoring systems using 3D models to train AI Machine Learning as opposed to obtaining real-world data which significantly improved their current capabilities.
This project provided a solution that did not require specialised equipment and worked similarly to the current site monitoring systems.
Our team delivered a qualitative step on the value that monitoring systems provide. For instance, estimating activity efficiencies, safe distances, and identifying potential contagions.
Working with Costain, we deployed CCTV cameras that were powered by green energy solutions and utilised a variety of wireless and radio communications to test and prove the new models across two separate Costain construction sites (Preston & A30).
The Zest Consult collaborative team’s synthetic data approach was proven to be highly effective when applied to machine learning. (Patent pending) Their Object Recognition AI models performed with a high level of accuracy, as evidenced by the significantly high rate of ~80% reduction in false positives yielded in the field.
This Synthetic Data approach to AI/Machine Learning provided a qualitative improvement over the traditional manual site monitoring services and enabled the team to achieve their goal of delivering a system that performs robust object recognition, tracking, and photogrammetry that would result in boosting environmental, safety, and productivity capabilities.
Overall, our team found when we were not bound by the restrictions of collecting and using solely real-world data for machine learning, we were able to generate the vast amounts of synthetic data required for machine learning in significantly less time and at a much lower cost that resulted in the development of a more accurate AI model.
Automated and accurate labelling
Enabling diverse datasets including edge cases
Large cost savings when compared to real-world data collection
Faster model development
Improvement in detection accuracy
Following the successful completion of this project, Zest is now testing and trialling this technology with multiple customers across various sectors (Construction, Finance, Health & Safety, and many more).