Dutch startup brings crowd analytics through machine learning

One of the characteristics of smart cities is that they seek to be more efficient than conventional cities in terms of reducing energy usage, reusing technology and information, etc. A Dutch analytics solutions provider believes this can be achieved by understanding the behavior of the citizens and making smart decisions based on it.
 
ViNotion has developed video analytics systems for analyzing crowded situations of pedestrians, bicycles and vehicles in wide streets. Unlike several other companies that depend on motion sensing for analytics, the company’s uses statistical learning.
 
ViNotion’s solution works by the machine learning how an object looks like. This would involve taking several pictures of objects like bicycles in different situations and the machine learning about what is common in all these pictures. Once the machine has learned the commonalities in all the images, it can scan the footages from the camera and look for patterns.
 
“We also do that on different scales of the image, so that even if the object is far or near, small or big, we can always detect it,” said Egbert Jaspers, Director of ViNotion at a seminar recently. “So even if the camera is moving, even if there are shadows or if the lighting conditions changes, we will be able to detect the object reliably. That is the advantage of using this kind of technology instead of motion sensing where you will have problems such as moving trees, shadows, reflections from water etc.”
 
Understanding citizens key to smart cities    
 
Jaspers explained everything that you do in smart city applications is about citizens. Hence, you first need to understand where the citizens are, how many there are and how they move through the city. As cities grow problems like traffic become a major issue. ViNotion’s solution provides the right answer for this.  
 
“We have a product to, in real-time, look at the traffic, see where the citizens are, how many there are,” said Jaspers. “Once you have this information you can start controlling it. Our product looks through the camera to the traffic, we can observe pedestrians, we can observe bicycles, we can observe trucks, cars, etc. and we can do this in very wide streets of up to 80 meters with an accuracy of over 95 percent.”
 
If you are able to analyze the people, you are able to use it for traffic management and crowd management. For example, if there is an event in a city, you will be able to analyze and understand the number of people that present in the area and use this information for better management. This could include controlling signage informing people where there is congestion and guiding them to routes that are more convenient for them.
 
Traffic management is not the only market that ViNotion’s solution can be applied in. Other verticals include retail, spatial planning city marketing, etc.