Neural Labs Ecosystem of Access Control Solutions
Access control is one of the most important sectors for Neural Labs, together with ITS, security, traffic, mobility and logistics.
There are several special cases and scenarios within access control that I intend to address in this article presenting the Neural Labs proposal.
The provision of increasingly more top-down solutions is one of our company objectives, without forgetting customers who only need to buy the "raw" technology.
In general, the use of video analytics for recognition of license plates and additional features such as make, color and type of vehicle offers mau benefits to the access control sector.
- Increased control of parking lots.
- Increased parking lot safety.
- Automated access to parking lots.
- Reduction in internal and external fraud.
- Segmentatioin of parking lot users for marketing purposes.
Here's a reminder of the applications that a video analytics system can offer in terms of access to a parking lot:
1. In a rotation parking lot, the license plate recognition system can simply pass the data to the management system for it to deal with the parking logics. This system will typically associate the license number to the ticket collected by the user to, among other things, be able to check that the ticket and license plate match on exit.
2. In this type of parking lot, Neural Labs can also classify the vehicle (car, motorcycle, etc.) and pass the data for application of the corresponding fee. In some countries, the law requires different pricing depending on the type of vehicle.
Detection of this on entry helps to select the right fee from the start. Historically, other technologies have been used to perform this function, such as ribbon sensors, weight sensors, etc. Nowadays, the breakthrough simpilicity of Deep Learning means we have a non-invasive option for determining the type of vehicle.
3. Automate access to a parking lot (building, company, etc.) based on lists of authorized users. The system opens and closes teh barriers if the license number is on the list of authorized users. These lists, of which there can be several, can contain residents, employees, workers, suppliers, etc. each with their own authorized hours, etc.
4. Manage lists of suspicious vehicles or vehicle types of specific interest. The system supports "surveillance" lists and issues warnings if it detects that any of the vehicles on such a list (or lists) are attempting to enter. In shopping malls, vehicles can be added to these lists if you know they are coming to commit crimes or if they have previously left the parking lot without paying.
5. Auditar de forma independiente los accesos a su aparcamiento. Instalado como sistema independiente al sistema de gesión de parquing. Algunos clientes desean una gestión independiente de accesos para auditar su sistema de gestión del aparcamiento, ya sea manual o semi automatizado. De esta auditoría suelen aparecer descuadres muy a menudo, debidos a fraude interno y externo.
6. Check that the license is not "cloned". Our system not only detects license plates but also color, make and type of vehicle. This information is detected as the vehicle approaches and is compared with the information about the vehicle stored in the system. If it does not match, access is denied and the corresponding warning is sent to the operator. In some countries, it is common for license plates to be stolen or duplicated, so allowing access to a residential complex, for instance, using only license plates could be very risky.
7. Check the physical condition of vehicles on entry to the parking lot. More general view cameras (evidence cameras) can be added to our system to support the license plate recognition cameras. Then the system can store filmed evidence of a vehicle's condition on entry. This information is extremely useful for avoiding false claims of vehicle damage.
8. Control parking lot capacity. By installing cameras at all the entrances and exits to and from the parking lot, the system can keep track of the number of vehicles on the premises.
In these cases, statistics on how long vehicles have been on the premises can be very useful, and warnings can even be issued when vehicles have been on the premises for longer than a given period of time.
9. Automate access and increase traceability in ports and logistic zones. Access to ports can be automated through recognition of the IDs of intermodal containers combined with those of truck license plates, always integrated in the logistics management system.
10. Wiegand card reader emulation. Users sometimes like to integrate access to a parking lot with management of the presence/access of people based on the proximity cards that are already being used in the building/company.
The idea is to translate the license plates of the vehicles entering the car park into de IDs on the cards of employees, residents, etc.
Using a license plate to Wiegand converter, the access control system treats license plate recognition cameras as proximity readers and licese plates as personal IDs. The access logistics are stored in the presence management software.
11. Marketing. Marketing departments always like to segment customers. It means they can focus highly targeted advertising on certain types of people, thus greatly increasing the likelihood of gaining a new client.
By detecting the color, make, type and age (from the license number) of their vehicles, marjeting departments can segments potential customers.
At the entrances to a shopping mall, for example, they can present highly targeted advertising to users on arrival simply by analyzing the kind of vehicle they drive.
A simple example could be a shopping mall advertising luxury when detecting arrival os a high-end vehicle, but advertising more accessible stores to low-middle range vehicles.
IMPLEMENTATION OF THE TECHNOLOGY
In some sountries it is already hard to find parking lots of any kind that don't have license plate recognition. While the installarion of these systems has already been made compulsory by law in some countries like Spain, the technology is becoming essential in this kind of business.
But in other countries these systems are still hardly found at all. At Neural Labs we are convinced that the benefits that these systems bring to mature markets will soon be recognized in emerging markets, and will be installed en masse.
Historically, the cost per lane of these systems has been a barrier in developing conuntries. Lately, with the emergence of very low cost smart cameras with embedded license plate recognition and more and more other functions, this technology has become much easier to implant, though not always successfully.
At Neural Labs, in March 2019, we launched our Neural Edge smart camera for situations where edge processing is required with very high recognition rates.
At Neural Labs, we always highlight two key factors for the successful implementation of this type of technology:
1. The first is that there should always be a preliminary analysis of the customer's needs. LIcens plate recognition is not enough if it does not serve a purpose and offer a definitive solution to a customer's problem.
This problem may be the need to automate revenue, detect internal theft, etc. For technology to serve a purpose it often needs to be integrated into other systems, such as parking lot administration systems, facial recognition systems, RFID etc. So, the preliminary analysis must make any integration needs, and even bespoke developments, clear in order to 100% fulfill the customer's specific praking requirements.
2. The second factor for success is good performance, i.e. the accuracy of the technology. Imagine a bar code reader taht only reads 70% of codes. This would make any store management application totally useless..
It's the same in this sector. Unless the license plate recognition success rate is extremely high, the system would be so totally unusable and unreliable that it would almost certainly end up being a failure.
In the vast majority of cases, non-recognition leads to manual correction, and low recognition rates lead customers to give up on the system.
In short, the question we always need to ask is:
- Why do I want to recognize license plates?
- How would it help with my operation?
- What problem would it solve?
- What is the retirn of the investment?
- With what systems should the license plate communicate?
- And we should always demand the very highest recognition rates.
There are infinite number of scenarios in the world of access control and at Neural Labs we treat each one as different and as having different levels os verticality. Some of these area:
- Shopping malls - with or without payment.
- Access to companies.
- Access to residential areas.
- Pay-as-you-go parking lots.
- Access to logistic zones and ports.
Neural networks and now Deep Learning are changing the world we live in by making systems much smarter. Although there is a lot of work to do, the current applications of this technology are endless. Neural Labs has been using neural networks since the 1990s and in recent years we have turned to Deep Learning to:
- Recognize vehicle license plates
- Recognize the make of vehicles
- Recognize types of vehicles (car, motorcycle, van, etc.)
- Recognize the color of vehicles.
THE NEURAL LABS PROPOSAL
Neural Labs offers products o address all the aforesaid cases.
Our first job is always to listen to the customer to try to work out how our technology and solutions can help them with their operations. Then comes the design of a solutions, which usually involves some integration/customization to make sure that the system really meets the targets set with the customer.
Hardware agnosticism is in our DNA, so our solutions are almost always designes with hardware that can be easily sourced in any market.
Support and training both in testing and implementation are also vital for the success of a project of a project and we always take special care over these phases.
In short, there is everything to be done in the access control sector. Neural Labs gets the job done with quality, with commiment and with the intention of forgoing long-term relationships with our customers.
Elías Valcárcel Torres
CEO & Co-Founder Neural Labs, S.L.