Human Intrusion Detection and Alarm System | HIDAS

Bispectral Sensing

Joy ride may not thrive

Prevent injuries to "free-riders" and protect your restricted zones from unauthorized human intrusion with HIDAS, a highly accurate detection system that has been field-proven to safeguard areas such as baggage belts.

Visible light sensitive

Thermography

Large-Scale Data Modeling

Deep Learning

De-risk your facilities with HIDAS

Viewer discretion is advised.

Incident footage in Atlanta's Hartsfield-Jackson International Airport in 2019.

Top-notched Performance

≥99.9% Recall, ≥99% Precision

HIDAS is a field-proven technology since 2016. It was rigorously tested and evaluated in the Hong Kong International Airport.

"Recall" is True Positive / (True Positive + False Negative);

"Precision" is True Positive / (True Positive + False Positive).

Tailored for human detection

including toddlers, kids and adults

No matter how cluttered the scene is, HIDAS is designed to detect any human intrusion and to safeguard the zone of interest.

Partially covered by baggages?

Not a problem

Even when key body features are blocked, the A.I. of HIDAS can also analyse the objects and deliver the same or better detection capability as human does.

Experienced false alarms with thermography?

Because you haven't met HIDAS

Our deep learning architecture projects heat signatures and chromatic features into N-dimensions, enabling us to leverage thermal characteristics to detect human intrusion while avoiding false alarms caused by heated baggage.

bag appearance in thermal and visibe light

Why thermography matters?

Using video analytics (VA) on visible light videos, aided by 3D vision alone doesn't work well. False alarms are frequently generated by this method due to human-like printing on baggage.
In contrast, HIDAS has been rigorously tested using various methods and imaging sources including depth+RGB cameras. Our research confirms that one of the best ways to minimize false alarms is not to ignore thermography.
If you evaluate other VA systems side-by-side with HIDAS, you will notice a difference - it's our promise.
The objects and scenes below easily demonstrate the weaknesses in systems that do not use thermography during processing.

Reject human pattern in printing

Robust against lighting and object-surface reflectance

HIDAS

(Bispectral Sensing + Deep Learning + Large-Scale Data Modeling)

VA by RGB Camera or

RGB+Depth Camera

Why choose HIDAS ?

Our detection system is best-in-class in terms of precision and recall, but we don't stop there. We work closely with our customers to continuously improve the system and tailor it to their unique needs.

SCADA

HIDAS supports integration with existing SCADA control systems, allowing operators to visually verify what is happening on site while monitoring data provided by other sensors. The live and processed footage from HIDAS can be integrated into the overall monitoring process within the SCADA system, enabling potential problems detected by video surveillance to be quickly addressed before they escalate further.

Customization

We understand that detection zones can be in close proximity to airline and terminal personnel. Our system can be configured to focus on specific areas of interest and avoid being triggered by just hands or legs.

Continuous Improvement

We work with our customers to provide tailored data characterization that improves their system. We understand that achieving 100% precision and recall is impossible, but we believe it is possible to continuously improve the customer's system using user data.

Drill Test

We regularly conduct drill tests on our system to reassure customers that it is always up to the task. These tests include actual pretend intruders whose looks are new to the system, as well as previous data samples. By doing so, we can demonstrate that a continuously trained system is not biased or saturated in detecting human intrusion.

Thermography

At our company, we utilize data collected from both visible light and thermal imaging sources. We carefully design detection algorithms to avoid false alarms triggered by non-human heated objects and believe that improving detection accuracy involves leveraging this key signal rather than abandoning it.