Edge AI video anomaly detection application
x-RD developed an edge AI application to assist with physical security zone and health screening access control and visitor management.
Challenge
As part of a Defence innovation facility, the Defence client requested an application to enhance and streamline government physical security and health screening for building access (foyer) management.
Traditional, largely manual, access control and visitor management processes are time-consuming, prone to human error, and unable to scale to meet increased demand. Building access controls and visitor management require robust systems and processes to ensure only authorised individuals gain access, particularly during high-traffic periods.
The client wished to explore AI-enabled solutions to enhance and streamline building access control and visitor management, while meeting and exceeding physical security zone and health screening requirements. The solution was required to scale as the facility planned to grow in capacity and complexity, with increasing numbers and types of personnel and visitors across the ADF, APS, industry, and academia.
Solution
To enhance and streamline building access control and visitor management, x-RD utilised a high-resolution 360-degree camera capturing individuals entering the boundary of the physical security zone. As each person entered the facility, a fingerprint was assigned with a unique identifier, and securely and temporarily stored to protect privacy. Within the computer vision model, the foyer was defined as a region of interest. Whenever a person passed through the region, the system tracked their movement path and marked their access attempt as authorised or unauthorised. A number of non-PII biometrics were used to train a recognition model and, over repeated interactions, the system learned to recognise patterns, updating its recognition model to improve performance and accuracy over time.
By combining person detection, tracking, and machine learning, the solution provided a solution to minimise reliance on manual checks, reducing delays at entry points, and providing the facility with a scalable and intelligent system for monitoring and controlling building access and visitor management.
Impact
The AI computer vision application automated building access and visitor management controls for health screening and access control by leveraging computer vision and non-sensitive biometric signatures.