Season 2 Challenges

OPENED ON: 07 JUL 2017  |  CLOSED ON: 31 OCT 2017  |  REWARD: INR 1,00,000
Reward money is paid in exchange of legally acquiring the solution, implementing it to solve the problem and meeting the success criteria. Milestones for paying the reward money would depend upon the complexity of challenge and maturity of the proposed solution, which would be discussed with the solver as soon as the proposed solution is selected by us.

Short Description:

ACTi CCTV cameras are installed in our premises. Solutions are invited for automatically assessing feed from these cameras to identify violations and trigger an alert.

Challenge Details

Host of ACTi CCTV cameras are installed in Tata Steel campus. There are around 22 cameras across the sprawling campus measuring around 50 acres. The CCTV cameras stream videos from various sections of the campus, mainly covering all the roads and open areas. Currently, the video streaming is manually verified for any violations which may result in missing of any event that may occur on the campus. Typical violations may include, speeding (limit is 35KMPH), not wearing helmet/seat belt, pedestrians not crossing at zebra crossing locations, parking in No Parking zones and driving in No-drive zones and during No-drive time limits as well. 

The violations should be automatically identified and an alert should be triggered on the video dashboards. 

Current State: 
Currently, video analytics is not implemented for the video streams. The security personnel keep a watch on the video streams and manually report the violations. The manual screening for identification of violations is risky as the operator may miss out on the same. 

The video streams should be subject to analytics and any violation should be visually marked and an alert should be sent to the specific area security personnel. The defaulter vehicle number should be recorded and should be traced back to the vehicle owner/driver. The driver should be removed from the campus for repeated violations (say 3 consecutive speeding beyond 35KMPH, or not wearing seat belt after 3 consecutive warnings etc.,). In addition, the analytics should also indicate potential near misses. For example, jumping the signal in the last minute, driving very close to or slightly above 35KMPH. 

Success Criteria & Measures of success:

  • No false alarms - Example, identifying seat belt violation when the user is wearing the seat belt or mis reading within speed limit as over speed limit
  • Timely alerts to the specific security personnel
  •  No misreading vehicle numbers

Have any query or need more clarification about this challenge?

Other Open Challenges