What is people counting for?
People counting in public spaces is essential for the efficient management of places such as stores, malls, airports and passenger stations. Beyond counting visitors, it offers advantages such as:
Operations and Resources Management: Helps to plan and allocate resources effectively according to the influx at different times, improving customer service and adapting to demand.
Space Optimization: Provides information on space occupation, allowing areas with high influx to be reorganized and maximizing the use of underused areas.
Strategic Decision Making: Food data over time informs decisions about hours, promotions, and preparation for high-demand events.
Evaluation of Campaigns and Events: Quantifies the impact of events and campaigns, allowing future adjustments based on public response.
Customer acquisition and sales conversion: Measure these key efficiency indicators in your commercial spaces to make decisions based on data.
Implementing people counting systems is essential to create attractive and well-managed environments in your spaces open to the public.
What is the traditional approach to counting people?
Traditionally, specialized people counting sensors have been used, which have evolved over the last 30 years as follows:
Years 1990-2000. Infrared people counters. (Up to 30% error) : These sensors count people by emitting beams of infrared light and detect interruptions to estimate the number.
Years 2000-2010. Stereo people counters. (5-10% error): Stereo sensors improved counting accuracy by using cameras to calculate 3D distances and estimate the number of people based on size and shape.
Years 2010-2020. ToF (Time-of-Flight) and LiDAR people counters. (3% error): Technologies like ToF and LiDAR have made people counting more accurate by measuring distances with reflected light and creating 3D maps with lasers.
Currently these types of sensors coexist and are installed in a large number of retailers and spaces open to the public. However, they have reached the stabilization of their development because they are mature technologies with little capacity for innovation.
The new method for large-scale people counting: AI on security cameras
Recently, a new actor has emerged in the ecosystem, a counting technology that has several advantages: AI applied to image processing with security cameras.
It is a developing technology that extends the counting coverage with high precision (greater than 99%). In addition, it allows the generation of large-scale projects by reusing the infrastructure of security cameras, which reduces the necessary investment costs by 10 times.
The risk of innovating a new counting technology with AI
AI has many advantages; but it has a risk; the current maturation of technology. Being a relatively new product, there are providers that do not have reliable counting algorithms and can present bad results.
An AI-based counting system should be evaluated by a professional and well-trained team within the company, since it is a maturing technology and you can find unreliable providers
When selecting an AI counting system, we recommend taking the following into account:
The supplier must be reliable, have proven experience and well-established customers in your industry.
It's important to request proof of data accuracy. These can be live demos, and count validation audits.
Evaluates the scalability of the system. Is it simple to install? is it plug and play? Does it have the capacity to support many count points? Does it integrate easily with your corporate systems?
Keep in mind also that AI can extract much more data than just customer counts, so it is recommended that you investigate the extra features each system offers.
Pros and cons of AI for people counting
The pros and cons of AI for people counting are:
Pros:
Very low infrastructure costs: with the cost of 1 traditional meter, you can install 10 AI-based counting points. In addition, it may be that you have zero cost if your cameras already exist for security reasons.
More range and accuracy: A good quality AI algorithm outperforms all types of traditional people counting sensors (overhead) in counting range and accuracy.
Scalability: having very low investment barriers, it is possible to implement it on a large scale with ease.
More intelligence: By using image analytics, the AI is able to detect more information and analytics found in images.
Cons:
You could be hit with a flop: Because it's a new technology, you'll need to thoroughly and professionally evaluate your vendors.
In conclusion, if you want a safe and traditional approach, it is recommended that you use traditional people counting technologies. In the case of wanting to take advantage of current AI technologies to obtain their benefits, it is better that you have a well-trained team trained to evaluate different providers to ensure that you will be successful