I recall my brief report of the Smart City Council’s program in Bangalore on 28th of July. The central theme of the program was “Surveillance”. However the focus was more on the technology implications rather than ethical or legal issues.
While discussing the issues of “Surveillance” through CCTV cameras the phenomenal increase in the need for storage capacities as well as the band width issues in moving the images collected from the CCTV cameras to the central processing station on a real time basis so as to be of use in real time decision making was also briefly discussed.
At the same time, the futility of having a low resolution image which later turns out to be of no use in recognizing the persons whose images are captured particularly when the person has committed some crime was also highlighted. This is a problem which Police normally experience. (Other than in the cases where the Police report that the CCTV cameras intelligently stopped functioning exactly when the crime was about to be committed!).
The Smart City Council’s program stopped at creating a case for buying more high resolution cameras and enhancing bandwidth infrastructure and data storage facilities which are all good for the commercial development of the “Surveillance industry”. However, there was no attempt to discuss the possible use of “Appropriate Technology” that can improve the efficiency of the image collection system along with reducing the burden on the data transmission and storage infrastructure. Since there was no representation of a Citizen centric panel member in the discussion, this point did not come up for discussion.
I hope Smart City Council incorporates this discussion in their subsequent so called “Round Tables”, one of which is scheduled in Mumbai for August 4th.
In the meantime, I would like to discuss two specific use cases and solutions that need further discussion in the appropriate decision making fora.
They are
- Use of Image enhancement Technologies to upgrade image quality
- Use of smart strategies to get proper images useful for prevention of crimes in the ATM security scenario
This is not meant to be a technology paper and hence it may leave certain technical details uncovered.
Image Enhancement
The idea of using image enhancement technologies is to manage with a low resolution image at the time of capturing and transmission back into the control room but enhance its quality when required with the use of Video image enhancement software running on the back end systems.
The attempt is to use Video enhancement on real time basis so that where necessary quick decisons can be arrived at.
This may be even treated as a “Video Forensic” strategy to create a better video from which vehicle number plates are easily identified, face recognition is reliably achieved so that instructions can be transmitted to operational police or medical or other disaster management units to either rush to the trouble spots or try to intercept a criminal trying to get away.
Image enhancement technologies work on two levels. One is “Optical level” and the other is “Geometric level”.
In “Optical level Super Resolution Imaging Technique”, the limitations of the optical device arising due to the “Diffraction Properties” of light are sought to be corrected for getting an output which is more useful.
In the “Geometric Super Resolution Imaging Technique”, the pixalations are corrected by removing noise so that more details of the picture could get revealed.
Here is an example of a real time video enhancement that can be achieved through software which may use the combination of both the above techniques and apply it on multiple video frames to generate a more intelligible video than what the camera first generates.
The above is just an example of what can be done back in the Police Control room to make surveillance through CCTV cameras more effective when we are constrained to use low resolution cameras because the budgets donot permit. It will also not require huge enhancement of storage and data transmission capabilities.
Those interested can research more on the possibilities. If we can direct some of our research capabilities in IISc type of organizations, we can perhaps develop some of these software indigenously so that the cost of Smart City surveillance comes down significantly.
Smart ATM Security
The image enhancement techniques will work when the problem is really of the quality of image received. It cannot however address the other situations where the CCTV owner does not bother to check if the installed cameras are working or not.
Let’s leave out deliberate deletion of CCTV footage which I have already discussed in earlier articles as a Section 65 Offence and look at an instance where we can force the use of a device only if the CCTV footage is working.
I have proposed this strategy for the ATM security system where
a) Entry to the ATM is through a biometric lock which captures the finger print.
We May or may not authenticate the person in real time with the customer data base but store the data for some time for use when required.
b) Face recognition camera is fixed directly to the ATM so that any body who does not expose his face will not be able to carry out the transaction.
For this purpose, the ATM is operated with a lock which gets opened only when the face recognition camera transmits a proper image back to the server and a “Go Ahead” signal comes from the Bank’s server.
The above strategy (IPR with Naavi) is simple and inexpensive but no Bank has introduced it so far because they are either not “Smart” or because they donot want to spend that “Extra Rupee” for security.
When we are prepared to develop the so called “Smart Cities”, such simple “Smart Solutions” need to be tried out as a part of our “Smart City Policing” requirement.
At present I am not sure if Police are part of the initial planning of any Smart City. Most such projects are driven by the vendors and the politicians who look at how fat is the project cost. If the cost is lower, politician is often not interested. ( I can vouch for this from my personal experience in the past). Police are brought in only at a much later stage when the technical infrastructure is already firmed up and there is no scope for structural changes.
I wish a demand is made by the Police that they be made part of any planning to introduce high end technology into our Governance system because they are the people who have to carry the baton when things go wrong.
At the same time, in order to avoid some bad elements in the Police who are subservient to the political masters corrupting the system, there needs to be some checks and balances including taking the CCTV footage archive out of control of the operational people so that it cannot be manipulated on a selective basis.
These are matters of detailing which can be handled in Smart City Policing strategies when required.
In summary, I would like to state that Smart solutions using appropriate technologies are essential to ensure that available technologies are used efficiently while we do also strive to increase the technology boundaries itself with better cameras, more storage space etc.
I urge IISc to dedicate some research in this direction if they have not already done.
Naavi
