Today, smart search of video is critical for law enforcement and city surveillance officials. In this regard, AI can make smart search even smarter. This article discusses how.
Today, smart search of video is critical for law enforcement and city surveillance officials. In this regard, AI can make smart search even smarter. This article discusses how.
We’ve established that video surveillance plays a critical role in
smart city projects. Yet as the camera count gets higher, so does video footage. This, then, presents a difficulty for law enforcement officials as they look for relevant video should something happen.
“A common challenge for security operators is managing the size of some video surveillance systems and the difficulty in monitoring a growing number of cameras in real-time. Surveillance operators simply cannot focus on thousands of camera feeds simultaneously,” said Andreas Göransson, Global Marketing Manager at Axis Communications.
“Investigating crimes is often tedious due to the large volume of stored video that needs to be searched,” said AJ Frazer, Chief Revenue Officer at Irisity. “Most police departments need to pay officers overtime to manually search through video during investigations. Even with a tool that accelerates the replay 10 to 1, it will take 10 hours to search 100 hours of video. This is both tedious and it takes officers off other more important duties.”
Smart search and how AI can help
This is where smart search can be helpful, a function that more and more law enforcement officers rely on.
“Smart search allows looking for crimes with a similar modus operandi, which will help LEOs identify a potential perpetrator faster. Smart search allows your software to search through the recorded video and evidence data looking only for activity that matches your smart search settings, saving you the time of sifting through video. Smart searches can help filter down from hours of video footage to a few seconds/minutes, so it makes the entire investigation process much more efficient,” said Augusto Chiaravalloti, Industry Marketing Manager at Genetec.
Now smart search has been made smarter, thanks to AI and deep learning-based analytics, which have greatly improved the scope and accuracy of smart search.
“Video analytics and deep learning technologies can be used to detect objects within a frame of video, along with information about the type, attributes and behavior of those objects. Analysts are finding these technologies invaluable when trying to track specific individuals, vehicles and even animals. They can quickly search through hours of video to locate and isolate something or someone of interest by size or color, its speed or direction,” Göransson said. “For instance, if a call comes in about a pickpocket wearing a blue shirt, surveillance operators can locate all individuals wearing blue shirts in a live video streaming from the scene, isolate the perpetrator and convey their location to officers in the vicinity, making it easier to apprehend the criminal.”
“AI has the capability of searching for almost any specific object, person, vehicles and other critical identifiers (e.g., color, type, quantity) within seconds where traditionally may have taken hours or even days,” said David Ly, CEO of Iveda. “AI is here now and it will be a part of all our lives from hereon. We must learn how to best leverage what’s available today to help us fill in the missing gaps within current capabilities.”
“AI software can be used to rapidly scan through large volumes of video to find clips of interest – like cars matching a description in a police report,” Frazer said. “A smart search tool like innoVi that creates metadata as video is stored can perform the same 100 hour video search in mere seconds by only presenting the investigator with clips that are relevant to the investigation.”
Integration of systems critical in smart cities
The need for smart search arises due to the sheer amount of video generated in a city. The video comes from various sources including typical cameras deployed on streets, traffic cameras and body-worn cameras, among others.
Indeed video surveillance is critical in smart city initiatives. Yet video alone isn’t enough for law enforcement and city officials to do their jobs effectively. Only when video is
integrated with other security systems can law enforcement become smarter and more efficient.
“When teams are relying on fragmented, under-utilized investigative technologies that cannot communicate with each other, there is a risk that crucial details will fail to reach beat officers and detectives until it is too late. To speed up the flow of information to first responders on the scene of a crime, all these systems, including surveillance, need to be integrated,” Göransson said. “Most cities use a comprehensive network of cameras to both reduce the overall crime rate, respond to incidents, and bring criminals to justice. As technology evolves, microphones and sensors can be integrated with network cameras to detect threats such as gunshots, raised voices, people entering prohibited areas or suspicious vehicles loitering near critical infrastructure.”
“Gunshot detection is sometimes used for public spaces in large cities. Integrating other systems like access control and alarm panels with the video system can help with detection and better response to incidents. Most of these types of integrations have been offered for years to cities that want a more advanced system. Using AI to assist in detection and response is the major new development in the last 3 years,” Frazer said.
Finally, integrating video with other systems has benefits beyond security. “In the age of digital transformation, it is all about data. Security systems focus so much on the negative events, something bad happening, but if we can add sensors that will provide us more environmental data, as to how crowded a certain area is or what the temperature is – is it dry/wet, etc.? Environmental sensors will help contribute to safer and more secure environments. So, it is not just more cameras, not just more sound detection, it’s beyond that. And giving all that data AI plays the key role into determining and interpreting what exactly is going on,” Ly said.