Comparing Camera Analytics to External Detectors


Since 1983 GJD has developed numerous advanced external motion detectors, we have noticed an increase in public demand for reliable human detection products, partially due to the rise of camera analytics.

Whilst researching data for reliability, we discovered that no extensive testing had taken place for comparing camera analytics against external motion detectors. We wanted to know which product delivers the most reliable human detection, as well as the fewest false alarms so we decided we would create our own tests and present the results to the security industry.

Purchasing cameras from two of the leading brands in the surveillance industry, we installed them using the manufacturers’ recommended settings, we deleted obvious areas of potential false triggers like trees from the detection zones and spent some additional time adjusting sensitivity, field of view etc to produce the best performance. We then installed a D-TECT 2 IP motion detector configured to cover the same field of view as the cameras. The D-TECT was linked to a 3rd camera to verify recorded events. These cameras were recorded onto the latest Milestone XProtect VMS software.

Month 1 Results Table:

GJD is currently conducting the second month of tests and will be presenting them once they have been verified and ready for publication.

If you would prefer to conduct your own tests, GJD will be happy to supply evaluation detectors for comparison, we will also support you in setup and configuration and explain how to ensure the optimum performance.

Whilst the results so far appear favorable to the motion detector we recognise that the camera performance could potentially be improved by further optimisation of the motion detection, we will also look at using either line crossing or deep learning video analytics for comparisons in the future.