Before we begin. What is the purpose of people counting?
People counting is a fundamental activity for businesses that manage open spaces accessible to the public, such as retail chains, shopping malls, or passenger terminals.
It has numerous applications ranging from measuring sales conversion to operational space management based on foot traffic.
By implementing people counting technology, you can start making data-driven decisions, thereby increasing sales and optimizing space efficiency.
Important: If you're looking to make quick and easy money without lifting a finger, this is NOT for you. People counting yields results for companies with a professional, tech-savvy approach and a strong commitment to a data-driven decision culture. If you believe you or your company possess these characteristics, keep reading.
There are currently several people counting technologies available, and today you will see which is the most economical way to implement counting for your organization.
To determine the most cost-effective option for installing people counting, you must first familiarize yourself with the available technologies.
If you're new to the subject, we recommend you check out this post where the available technologies are explained. In summary, they can be categorized into two main groups:
Sensor-based counting technologies
AI-based counting technologies using security cameras
Both technologies have advantages and disadvantages, which you can find here.
People counting technologies based on sensors
People counting sensors are typically like this:
They should be installed above a door where you want to count people and must have an internet connection to upload the counting data to a Cloud platform. When installing sensors, you typically need to consider three things:
1. Sensor installation points: The doors where you want to install people counting in your physical spaces.
2. Sensor placement: Typically, one sensor every 2 meters of door.
3. Network connection: To enable data uploading to the cloud.
AI-based people counting technologies on security camera video analytics
Publicly accessible spaces typically already have cameras for security reasons. These cameras often provide coverage of the points of interest. This means that the raw data already exists, and AI can be applied to extract people counting from the existing infrastructure.
Security-installed cameras already have the raw data; you just need AI to extract people counting.
The architecture may vary, but we'll assume a common case in which the company has a camera system with the capability to remotely access real-time video streams. In this case, you should consider the following elements:
1. Camera system coverage: Verify if your camera coverage includes the points of interest.
2. AI installation: The AI should be installed on a PC or Workstation with access to the camera streams and then upload the counting data to the cloud. Typically, a Workstation costing between 500 and 1,000 USD can apply AI to 25 cameras in various locations.
How can you estimate a people counting project in your organization?
So you can understand the costs associated with a large-scale people counting project, you can consider the following three installation cases, which are the most common:
Traditional installation of people counting sensors
Installation of AI-based people counting on existing camera infrastructure
Installation of AI-based people counting on new camera infrastructure
| People counting sensors  | AI over existing camera infrastructure  | AI over new camera infrastructure  |
Hardware costs | 500-1.000 USD per sensor | 500-1.000 USD per 25 cameras (the PC or Workstation cost) | 50-200 USD per camera y 500-1.000 USD per 25 cameras (the PC or Workstation cost) |
Mounting costs | High (requires training and special locations) | 0 USD | Low (can be done with typical camera suppliers) |
In this table, the following are not taken into account:
Initial setup costs (for example, for data integration into your company),
Monthly maintenance costs (a monthly fee for platform maintenance, equipment monitoring, etc.). Both costs depend on the specific provider and may vary according to each business model.
Below, we provide some example cases in different industries for you to assess the required investment costs and compare the estimated investment for some example cases in various industries.
EXAMPLE. Cost of a people counting system with 25 counting points
You can consider the following numbers as a close estimate of the cost for the following example projects:
Implementation of people counting in a chain of 25 stores
Implementation of people counting in a shopping mall with 25 cameras
Implementation of people counting in a passenger terminal with 25 cameras
People counting sensors  | AI over existing camera infraestructure  | AI over new camera infraestructure  | |
Hardware costs (25 counting points) | 25.000 USD | 1.000 USD | 6.000 USD |
Mounting costs (25 counting points) | High | Very low | Low |
In conclusion, what is the most economical way to set up people counting?
Based on the analyzed information, we can conclude that the most economical way to set up people counting is: Reusing security cameras in a centralized manner using AI. Additionally:
Reusing existing cameras in a centralized AI-based people counting system is the most economical option in all aspects.
A centralized AI-based counting system with existing cameras is at least 25 times more cost-effective than implementing people counting sensors.
A centralized AI-based counting system with new cameras is at least 4 times more cost-effective than implementing people counting sensors.
A centralized AI-based counting system with existing cameras is at least 25 times more cost-effective than implementing people counting sensors.
We hope this information is helpful to you! If you have any questions, please don't hesitate to contact us at this link.