Counting inputs and outputs (customer counter) - types of solutions and technologies used

Measuring the number of people who have entered and exited a particular room or building serves us in very many types of applications. A customer counting system will help to adjust the display in our store in an advanced way, prepare dedicated marketing campaigns and determine their effectiveness, and check at which times of the day the traffic is the highest, etc. Read our post and see which customer counter will work best for your application. 

Customer counting system

Sensors dedicated to checking the movement of people come in various forms and use different methods of power supply, data transmission and, of course, different methods of measuring the movement of people in a space. In this text, we want to highlight the types of customer determination devices present and help you choose the best-fitting solution for your needs.

Measurement methods used by people counters

The many methods of measurement available in human traffic counters make specific devices suitable for different applications. We are able to tailor the solution to the budget required by the customer, along with the price, however, the accuracy of the device also goes down. The measurement methods cited below have different effects on device prices and measurement accuracy.

The second important feature of a good meter is the detection range. Different methods of measurement allow for completely different ranges, combining devices operating side by side, and mounting in different places such as walls, ceilings, etc.

Additional functions that a customer counter can offer, such as distinguishing between employees and visitors, are also only available using certain measurement methods. When considering the choice of a counter, therefore, we should also pay attention to the additional applications for which we want to use the device. Other examples of additional functions include: checking the filling of shopping carts, checking whether the person being surveyed has turned back in the aisle, distinguishing between children and adults, and much more. We write about the specific additional functions of the devices in the description of the specific equipment.

Compliance with the General Data Protection Regulation (GDPR).

RODO - the Data Protection Regulation is the most restrictive law governing the privacy of collected data in the world. The law was passed by the European Union in 2016 and applies to all entities interested in collecting any data on European Union residents.

All of the solutions we describe below meet the requirements of RODO, the most important features include:

- No local data storage: No additional space on the device for data collection. Information is directly processed by the AI algorithm without local storage.

- No access to public networks: The option without PoE power, instead using USB allows the device not to be plugged into the public network, providing additional data security. 

- AI algorithm: It is used to recognize and analyze human movement through data collected by the device. 

- LoRaWAN: All of the following are in LoRaWAN, which means the technology allows data packets from 0.3 to 50 kbps, so it can transmit the result of AI analysis on the device, but cannot forward the videos and images collected by the device. In addition, the collected data is only sent to the gateway and client server.  

PIR method - cost-effective solution

A people counter of this kind uses dual PIR (Passive Infrared Sensor) sensors to identify the presence of people by detecting changes in infrared radiation. To distinguish bidirectional movements and realize accurate bidirectional people counting, two PIR sensors work simultaneously to detect the directions of movement determined by the signal waveform.

This is the cheapest and also the least accurate method of measurement, with an accuracy of 80%. However, in some implementations this value is sufficient, and the low cost of the device makes the solution attractive for some applications.

Not very satisfactory in many deployments can also be the range of this method. This is because it plateaus at about 2 meters when mounting the device on a wall in an aisle.

The lack of additional functionality, such as detecting changes in direction of movement and distinguishing one person from groups of people, can further affect problems when using the device for deployments that require high accuracy measurements.

Low battery consumption, meanwhile, allows devices of this type to operate for many years on replaceable batteries, without the need for mains power. This greatly simplifies deployment and reduces operating costs

Examples of the PIR method:

Milesight VS350

VS350 is an indoor aisle people counter that detects and analyzes the flow of people, enabling optimal space management and utilization. Equipped with dual PIR sensors, it offers considerable accuracy for bidirectional people counting. The device additionally features a temperature sensor, providing additional measurements of the environment in which it is installed. As a Milesight D2D controller, the VS350 communicates seamlessly with other Milesight D2D devices and lorawan gateways.

The device can be mounted on the wall or on the ceiling, ensuring measurement accuracy of 80-95%. More information about the device:

Radar method

Radar, originally an acronym for the term: radio detection, is a dedicated system that uses radio waves to determine the distance, angle of incidence and speed of object movement. It is an "active" method, that is, one that uses its own radiation source (transmitter) to locate objects. It usually operates in the microwave regions of the electromagnetic spectrum, at frequencies from 400 Mhz to 40 Ghz. 

When using radar devices , there is no need to worry about privacy. They also provide great distance and width for taking measurements. On the downside, however, it should be noted that this method is prone to problems related to the detection of static objects and very close objects. 

Thermoelectric method

This method involves the direct conversion of a temperature difference into an electrical voltage and vice versa. A thermoelectric device produces a voltage when there is a different temperature on either side of it. Conversely, when a voltage is applied to it, heat is transferred from one side to the other, creating a temperature difference.

This ensures reliable and accurate data collection with up to 95% accuracy. When counting only single-person walks, by contrast, the wireless counter achieves exceptional accuracy of up to 99%. This level of precision allows you to rely on the collected data to make critical decisions.

This solution combines cost-effectiveness, high accuracy and low energy consumption. Using thermoelectric and radar technologies simultaneously, people passing through designated areas can be accurately counted in both directions.

Examples of thermoelectric and radar methods:

Milesight VS351

The VS351 is a compact AI thermoelectric people counter designed for indoor and exit applications. It offers high accuracy in bidirectional people counting, enabling efficient analysis of pedestrian movement and effective space management. Combined with a radar sensor for presence detection, it intelligently schedules the activation time of the second sensor to optimize energy consumption.

The device at low cost and energy consumption provides 95-99% efficiency depending on the measurement conditions. 

More information about the device:

ToF method

ToF (time-of-flight) technology uses emitting light in the direction of the object under study and checking how quickly the light will reflect and return to the device. Such a method allows the device to determine the surveyed object's shape, position and surroundings. This makes it possible not to worry about privacy issues, since only the silhouette of the person is examined, without video.  

Sensors of this type are most often also equipped with a range of AI solutions to distinguish human silhouettes in order to take measurements with greater accuracy and, for example, not confuse the movement of animals under the sensor with people going in and out.

The accuracy of the device can be as high as 99.9% in the most advanced equipment, and this is an excellent method for making measurements even in the most demanding implementations. ToF sensors are also an active method of measurement, as they use the devices' emission of light instead of collecting light from the environment. This allows devices using this technique to work even in total darkness. 

Examples of the ToF method:

Milesight VS135

VS135 is a high-end people counting sensor based on second-generation AI and ToF deep learning technology. It is capable of adapting to a variety of complex scenarios while providing 100% privacy protection for customers. The sensor has an impressive 99.8% accuracy in people counting, fully satisfying the most demanding needs and provides exceptional performance in both indoor and outdoor applications.

With a ceiling mounting height of up to 6.5 meters and an IP65 waterproof rating, the sensor adapts seamlessly to any environment.

More information about the device:

Ultrasonic method

Ultrasound is any vibration or sound wave that has a frequency higher than the upper limit of sound audible to humans. That is, all sounds greater than 20kHZ. Ultrasonic technology, unlike passive methods such as PIR, emits waves on its own.

The device using this technology will emit ultrasonic waves. When an object appears in the space to be surveyed, the emitted ultrasonic waves will be reflected and travel back to the device. The device can thus determine the distance to the object under study by examining the wavelength and the time it took for the wave to return. 

This technology is suitable for use when we have a small field of view for the sensor. The technology has no privacy issues. However, it may have a problem detecting more people at once.

Stereovision method

Stereo vision systems rely on two cameras placed at a fixed distance from each other to capture two images of the same scene from slightly different perspectives, allowing them to calculate the distance between objects based on their disparity.

By using this method in conjunction with the AI system sewn into the device, it is possible to accurately determine the number of people who have passed in the vicinity of the device. With the ability to connect up to 16 devices, it provides comprehensive coverage of most detection areas, enhancing the capabilities of industries such as retail and building management.

It is an excellent solution when we need to study the movement of people in a larger space. Besides, it allows checking such parameters as the gender of visitors, whether they are children or adults, etc.

Examples of the stereovision method:

Milesight VS125

The VS125 is a professional people counting sensor that is based on deep learning AI and Binocular Stereo Vision technology. The sensor has an impressive accuracy of up to 99.8% in person counting and provides exceptional performance even in low-light and total darkness conditions. In addition, it can achieve rich attribute recognition, including gender and facial expressions. It is designed with privacy in mind, which is GDPR-compliant.

More information about the device:

Vision method

As a key component of CCTV systems, network cameras typically collect images through the use of lenses, an imaging sensor (CCD/CMOS), an image processor, a digital processor and a circuit. Camera image collection can be useful in many applications. However, it is not an option designed to count people.

However, based on the images it collects, it can realize people counting through analysis using AI algorithms sewn into the device. Implementing an entire camera system for the purpose of counting people also most often proves to be cost ineffective and complicated. It will also generate privacy issues by capturing and storing camera images. 

Examples of the vision method:

Milesight VS121

VS121 is an advanced sensor that uses AI algorithms to check occupancy in the space it is observing. The sensor has the ability to program up to 16 workspaces in which it will detect with up to 98% accuracy if a person is present or if an area is empty, it will also count the current number of people in the surveyed space.

With an installation height of up to 7 meters, the sensor can check an area of as much as 78m2. The use of LoRaWAN technology, on the other hand, allows for low power consumption and a very long transmission distance of collected data, i.e. up to 15 km in a straight line in undeveloped space or up to 2 km in a city. The collected data can be easily checked in the application. VS121 can be widely used in building smart office systems, building a system for checking the number of visitors to specific exhibitions in museums, etc.

A table with a summary of each solution:

Mode of communication

A very important factor in choosing the right device is also to select a matching communication method. Many devices will be able to be installed by providing cable communication and network power, but this is not always possible.

The devices in our portfolio use a variety of communication methods such as WiFi, Ethernet, LoRaWAN, LTE/4G, etc. All to ensure versatility in adapting the device installation method to the specific space and customer requirements. 

Stovaris - a value-added distributor

At Stovaris, we offer assistance at every stage of implementation: from selecting a solution for the customer, to learning how to manage the devices and software layer, to installing the sensors.

We encourage you to contact us and work with us on specific implementations. We have extensive experience from working on the above-mentioned equipment, many completed projects and satisfied customers.