We never know who is right besides us on a road, who is walking with us or who is travelling around. Maybe a good person, maybe bad. What If the person has criminal record? What if the person is a most wanted criminal? Well the idea of "Eyes over street" could help. Eyes over street, is a Digital Mass Surveillance System over the streets of Kathmandu or any town. The main idea is the cameras installed over the street records every face coming and stacks them in a database and if there is criminal record of the person the authorities are notified instantly.
"How does the system work?”
First a computer (server) receives live feed from all the cameras around the city. Then it detects face in the feed and generates images and stacks them in a database and also searches in the criminal records if there are similar faces (Open source technologies like openCV and openBR can be used to analyze and processes the data). If a face similar to that of the criminal is found the authorities are notified by push notifications on the smart phone or their computers. In charge of the nearest police station gets the criminal details, location and also photos (depending on the technology they use) by SMS (for non-smartphone or offline devices) or by the dedicated application installed on the device (In case of smart phone and connected to internet). The notifications are sent to nearest possible location (Police Station), so they can mobilize the police or security officers faster and track down the person sooner.
The system is not just supposed to match the faces it is supposed to record the activities of every person on the road. So as a whole using the system I am proposing a better way for monitoring the roads , the parks , the buildings , malls, cinema halls with cameras and just few rooms in a central office ready to analyze data and implement the resources needed . Not just stop the criminal activities we can divide the same resources to the traffic control system and for slimmer traffic. And we could also share traffic information to people.
In a nutshell, few resources, cameras and computers can actually provide a better security system mobilizing lesser human power.
System is targeted to give security to every common person from theft, kidnapping or murder and also the monuments or the building and other things around street. This system can actually be helpful in maintaining security with Police or National Security Agencies. With lesser amount of human resources we can monitor the whole city and when necessary mobilize the ready to go force to help people or to take down criminals.
“Once implemented it can decrease the crime rate of the city and we can take some more time in surveying local areas to maximize security and implement the manpower at the need. But also you should maximize the security in the computers and network and create software accurate enough“
-Police officer at local police station
“Security in most crowded areas can be made better and VIPs can also get better security.”
As this system focuses on security and helping Police and National Security Agencies, we don’t have quite much competition in the current scenario, there are few surveillance systems being applied by traffic police in various areas the country as this system intends to give better services and fills up canvas with more features so this system must get over all the digital systems of the kind currently being used.
Shubham Lekhak, CDO:
Shubham Lekhak is Chief District Officer and he has to deal with administrative as well security based affairs and due to increasing crime rate his attention towards administration was decreasing but now with this system the crime rate is quite a lot lower. As security is better now the district officers can work more on administrative as well development works and don’t have to worry much about security.
Ankit Sharma, Commissioner of Police
Ankit Sharma is Commissioner of police of the district and he is very happy to see the implementation of the system in the city area and his reports says the criminals are being spotted at various areas and are captured within a few hours of seeing them and they now don’t have to implement many police officers for daily round instead they are preparing officers for other combat and quite few teams are working on the daily round trip of different areas.
Top User Stories:
They found our boy within 20 hours:
8 years old Subin Neupane was taken from a playground near his house. The kidnapper’s vehicle and kidnappers were identified and a search team was set up. Within few hours same vehicle was spotted in another area of the city and they followed the trail and further investigation led to the safe rescue of the boy with in around 20 hours.
Crime rates in city decreased by 42%
Police and Home Ministry published semiannual report on past few months of installation the system which explains crime rate has reduced by 42%. And curves explaining the reduction seem to go incredibly down towards axis. Ministry now can start new programs for public safety and now can take more time for border security issues and other national security issues.
Better Traffic System
Using same cameras and installing few more computers and adding custom software now Traffic Police can get the same feed from main junctions and busy areas, and now can implement their computer based traffic system which gives a lot of ease to maintain traffic on roads and same system is being used to notify common people about traffic through various medias.
As this technology is designed for digital surveillance by police or national security agency on roads and public places, so we might not have much business value but we can use the architectures of this system in small scale like in museums, schools, supermarkets, hotels or cafés maybe not to identify criminals but at least to store the footage and images of people in a smarter manner and maximize security.
As the system uses various well tested models and software architectures, I think this will work fine enough for longer time and we can be ready to implement new modifications with time and make system even stronger.
Major technologies required are huge numbers of cameras and very fast computing infrastructures and software technologies with fast enough algorithms to analyze hundreds and thousands of images in just seconds. Custom Built computers and software built using Open Source technologies like openCV and openBR and other surveillance technologies can help in analyzing faces and match with the database. Programming languages like c/c++ , java and web technologies can also be used in entire system. Also as we are focusing the notification systems for mobile phones and latest operating systems we are going to need various APIs and frameworks to build platform specific applications.
Open CV: http://opencv.org/
Open BR: http://openbiometrics.org/
Demo for facial recognition: http://skybiometrics.com/Demo
Youtube link: http://www.youtube.com/watch?v=tPPCfV7CI9A
<iframe width="853" height="480" src="//www.youtube.com/embed/tPPCfV7CI9A" frameborder="0" allowfullscreen></iframe>