This is not only Gatekeeper’s position on undercarriage inspection but now a rapidly growing number of concerned and experienced security professionals across the globe as vehicles have rapidly become the weapon of choice and preferred means of transporting contraband of all descriptions, both inside and under motor vehicles.
These same concerned security professionals, in the main, were the first to embrace the significant benefits of video analytics. Video analytics or video content analysis is a subset of computer vision and is a true representation of Artificial Intelligence(AI). When AI is applied to the world of security technology a number of inherent challenges within the various technologies are mitigated.
The main challenge being that of the limitations of the human mind, an impediment to security personnel (guards) working at their optimal level, or put another way providing the standard of security we all seek. In order to meet the stated Security Function, security personnel must perform their tasks meticulously, efficiently and tirelessly. Sounds great as a stated aim for security personnel, however all three are individually, and or collectively almost impossible to achieve without the assistance of AI.
We have all seen the banks of video displays in control rooms and the impossible task required of guards to spot a person climbing a fence, entering a restricted area, or going through a door in the wrong direction etc.
These breaches of security rarely happen, however the guard has to be on his best game all of the time while their subconscious mind is telling them “relax nothing ever happens here”. Unlike the world before video analytics, security guards can now accurately and timely monitor such events through the deployment of AI contained within most modern security CCTV systems.
These systems automatically alert guards to the very issues they have been tasked to monitor, their Security Function, thus allowing the guards to act accordingly and provide the level of security required. The human mind is a wonderful thing, however numerous studies show that humans can only really concentrate on a single task for a maximum of twenty minutes before zoning out – drifting off to another place in our mind.
This is where the issues begin as you are asking a guard (a human) to resist human failings and be an eight hour fully focused security sensor – not possible. What AI has done both in the CCTV world and within the Gatekeeper technology is to change a guard from a sensor to a decision maker.A role that guards, with the correct amount of training, can perform almost to perfection.
Automatic v’s. Manual Inspection Systems
Gatekeeper’s development and application of AI to the world of under vehicle inspection has been a breakthrough for the security industry worldwide. AI has removed, to a greater part, the challenges associated with guards inspecting undercarriages of vehicles. These challenges are typically:
1. Provide a comprehensive search of a vehicle’s undercarriage in an efficient and timely manner.
2. Search to be conducted by personnel with little if any knowledge of the mechanical layout of a vehicle undercarriage.
3. Ability of guards to concentrate and conduct comprehensive searches for many hours at a time in varying work environments.
Past technologies or systems used to inspect the undercarriage of motor vehicles have all required the guard to deal with these challenges in combination with other issues related to the particular manual system deployed. For example:
1. Sticks and mirrors or cameras on sticks provide an extremely limited view of an undercarriage. You still see a number of sticks and mirrors leaning up against walls of guard rooms as a reminder of days gone by. That said there are still some security personnel out there where this is their only form inspection technology.
2. Running video systems (typically speed humps with little video cameras) produce moving, unrelated video streams that are virtually impossible for the human mind to decipher or understand let alone spot an anomaly. There are no reference images or video strips for the operator to try and compare just a series of videos moving at the speed of the vehicle in a blind hope that the operator knows what the vehicle undercarriage looks like and is looking when the vehicle, with the anomaly, runs over the video system.
There are several, mainly Chinese companies, that still offer this type of inspection device. However, most of the earlier manufactures of the running video style systems have gone out of business, as very few if any, end users purchase such equipment nowadays.
3. Line scan systems, up until the past several years, have been the main form of undercarriage imaging.However, they all, regardless of the manufacturer, suffer a significant technological impediment in that they are extremely speed sensitive so no two line scan images are alike (slightly slower vehicle speed than the scan rate of the camera produces an elongated image, slightly faster vehicle speed than the scan rate creates a short compressed image – illustrated below). This dissimilarity in images makes any form of automatic comparison virtually impossible, despite what the manufactures’ brochures say/show.
This unlikeness is an enormous challenge for a guard to manually spot the difference between two completely different images of the same vehicle all because they were scanned at slightly different speeds. This limitation/absolute impediment to the line scan technology was confirmed by Chemring (Ex BDL) recently in a document provided to a prospective Middle East customer saying “For accurate image reconstruction, the line rate of the scanning camera needs to be driven at a speed synchronised to the movement but often, constant speed is assumed, which in most cases gives acceptable results (but can give distortions in the direction of travel).” In realityvirtually all line scan images of vehicle undercarriages are distorted in one way or another.
Directly below is a color line scan image of a 30 seat bus when it crossed a line scanner slightly faster than the scan rate of the camera. Compare this with the Gatekeeper scanned image below of an almost identical 30 seat bus (compare features of air-conditioner, drive shaft, tires, exhaust system etc.)
The following image is an example of a line scan image of a car when it stops on top of a line scanner. A line scanner designed to scan vehicle undercarriages simply has no ability to stop recording lines or rows of pixels when a vehicle stops.It just keeps on recording and placing each consecutive row alongside the next to form an image until the recorder (line scanner) is switched off.
You can see from the above examples how it is virtually impossible for a line scanner to perform an automatic comparison of the same vehicle undercarriage image as they are only as accurate as their relative speed to the scan rate of the camera.
Over the past years much has been attempted to correct this limitation which is only made worse with high resolution, a claimed advantage of a line scan camera. In a machine vision environment where speed can be controlled very accurately, for example a conveyor belt, a line scanner can be deployed very successfully but not in situations where object speed (a vehicle) cannot be meticulously controlled.
The new generation under vehicle inspection system is fully automatic
As referenced above “manual” systems all essentially suffer from the same challenges i.e. reliance on the guard to spot the difference. In most cases with no useful reference image to assist, require the guard to have specific knowledge of the mechanical layout of almost any vehicle, be able to look on top of cross beams and axels and for the guard to perform the inspection task for hours on end. Gatekeeper’s experience in countries like Iraq has seen our technology find mechanical changes to vehicle undercarriages that to the human eye look normal.
For example, the removal of a vehicle’s spare tire and replacing it with a metal plate, creating a storage pocket accessible through the vehicle’s trunk. The Gatekeeper automatic under vehicle inspection system (AUVIS), is able to find such anomalies purely because the system/computer knows what the vehicle undercarriage should look like and when it is different (depending on the sensitivity setting down to small wires) automatically highlights the difference on screen and alerts the operator/guard to the anomaly.
Patented and Unique Technology
Gatekeeper has numerous US and international patents either awarded or pending surrounding its technologies. One of these patents relate to the “fingerprinting” or vehicle printing of a vehicle’s undercarriage. What this means is that the AUVIS can recall the same or similar vehicle from a system database purely by its undercarriage image (pattern).
In this way the Gatekeeper AUVIS does not rely nor require a vehicle identifier (license plate, RFID tag etc.) to recall an image for automatic comparison. The AUVIS can run in either a one to one mode through the use of a specific vehicle identifier or on a one to many mode by searching the system database for the same or similar vehicle based on the vehiclefingerprint.
This dual ability is extremely valuable as the system does not have to have seen every vehicle before it can perform its automatic searches. Instead, if the AUVIS has seen a similar vehicle in the past that undercarriage image will be used as the reference image for automatic comparison. The AUVIS search algorithms are so accurate that they can, with >96% tested accuracy, inform an operator of a vehicles make, model, year of manufacture, engine size, transmission type and style of vehicle. All of this information gathered by just optically analyzing the undercarriage.
Environmental and Driver Challenges
In almost all cases, vehicle inspections are required to be conducted out in the open and with tolerance for normal driver behavior. These requirements are another area where the Gatekeeper AUVIS excels as the technology has been designed and has an IP68 (Ingress Protection) certification and has proven capability of handling these challenges, namely;
1. Withstanding harsh environmental conditions proven by systems installed from the North of Russia to Asia and the Middle East for over nine years.
2. Ambient lighting conditions are dealt with by matching the wavelength of the illumination with the optimal wavelength of the scanner’s sensor – reducing dramatically the effects of ambient light. Apart from back ground conditions day and night images are the same due to the illumination technology used.
The use of monochrome array is to provide greater contrast within the grey scale as it is the computer that is providing the image analysis. This capacity is enhanced by the strict control over illumination wavelength verses the scanner’s optimal sensitivity at 650 nm. Optical filters within the scanner controls the ambient light outside of the 650nm wavelength and blocks these from the sensor array.
3. Challenges of normal driver behavior i.e. stopping and starting during the scanning process are not a problem to the Gatekeeper AUVIS. The intelligence within the algorithms recognizes when a vehicle has stopped and only restarts the stitching process when new/different pixels are detected, ensuring the same high quality images regardless of a vehicle’s behavior during the scan process.
4. Variation in speed of vehicles being scanned is also a strength of the AUVIS as a driver can drive very slowly or accelerate up to speeds of 35kph without affecting image quality or the system’s ability to automatically search and compare vehicle images. The same cannot be said of other technologies/systems referred to above.
5. High vehicle through put at pressure times of the day is achieved by the systems Hands Off function which allows for the system to work fully automatically without the guards involvement by automatically inspecting one vehicle every 4 seconds.
6. Watch lists and alerts are features of the modern under vehicle inspection system.However, the Gatekeeper system takes this to another level by deploying its fingerprint technology. For example, if a license plate is removed from one vehicle and placed on another, the system will alarm as the AUVIS will compare the fingerprint of the vehicle with the registered license plate in the database.
Networking and Integration
The world of security has moved from a series of disparate, non-integrated systems to a series of integrated technologies, each providing a layer of security designed to provide a current overall threat condition viewed in a central control room. Systems are sharing data and comparing results, looking for an indication of a potential threat.
Gatekeeper is involved in a number of high profile installations worldwide where the company’s participation in large networked and complex security systems is required. In fact, as much as Gatekeeper has developed the automatic undercarriage inspection technologies the company has also lead the way in large system integration of undercarriage scanning via the use of a central database configuration.
All scanners retrieve reference images and report all vehicle activity to a central database. They also integrate with other access control technologies separately or through various PSIMs.
Misrepresentations of Manufacturers Capabilities – Gatekeeper Offers USD100,000plus unlimited video rights of the test to any party who can prove they can beat Gatekeeper’s AUVIS.
As much as under vehicle inspection technology has move forward to provide and meet a higher Security Function many under vehicle inspection technology manufactures have chosen to misrepresent their technology’s capabilities to the detriment of the end user.
In response to this ongoing issue Gatekeeper has for the past nine years offered to manufacturers, end users and consultants the ability to video tape a side by side test of the various claims made by the associated manufacturers against those of Gatekeeperwith the video tape being available on an unrestricted basis for marketing purposes by either party.
To date not one manufacturer has agreed to demonstrate their claimed capabilities. As a result Gatekeeper is now offering USD100,000 and unrestricted video rights to a party judged by an independent party to be superior to the Gatekeeper technology. Let’s see who will step forward.