Seeking Technologies for Automated Detection and Identification of Vehicles with Smoke Emissions

Background / Description

The Open Innovation Platform (OIP), an initiative by the Info-Communications Media Development Authority of Singapore (IMDA), aims to connect real business challenges to problem solvers. 



The National Environment Agency (NEA) is the leading public organisation responsible for ensuring a clean and green environment, and the sustainable development of Singapore. Its key roles are to protect Singapore from pollution, maintain a high level of public health and provide timely meteorological information. Vehicular exhaust emission is a key contributor to air pollution. Under the Environmental Protection and Management (Vehicular Emissions) Regulations, it is an offence for any vehicle to emit smoke while on the road (“smoky vehicle”).

Currently, NEA conducts surveillance operations at designated hotspots on public roads and at the land entry checkpoints to detect smoky vehicles. For roadside surveillance, officers are equipped with video-cameras to capture video-graphic evidence of smoky vehicles spotted. At the land entry checkpoints, officers visually screen for smoky vehicles and send these smoky vehicles for further testing. The current operation is manpower-intensive and is not scalable to cover all the smoky vehicle hotspots simultaneously.


Problem Statement

How might we automate remote detection of smoky vehicles on the roads and at the land-entry checkpoints accurately and cost-effectively?

Technical Specifications

What Are We Looking For

  • A Proof-of-Concept (POC) solution that could remotely detect smoky vehicles on the roads and at land-entry checkpoints so as to improve operational effectiveness and reduce reliance on manpower
  • Proposed solution could support scenarios related to roadside surveillance or checkpoint (as described below)
  • Proposed solution should give details on the scope of work relating to the collection of video footages at sites selected by NEA (by mounting new cameras on infrastructure or manually by hand) and the development of relevant video-analytics
  • Proposed solution should demonstrate the technical feasibility of automated remote smoky vehicle detection

Proposed solution should:

  1. record the video-footage of all the vehicles travelling at the designated hotspots within the lanes monitored on the public road and designated locations at the checkpoint, including their licence plate numbers for the entire monitoring duration with the date and time stamp
  2. use video-analytics solution to identify and output spliced video footage of the smoky vehicles detected including their licence plate number, location, date and time
  3. take into consideration the accuracy and false alarm rate of detecting smoky vehicles and correctness of enforcement data (i.e. licence plate number, location, date and time)

Desired effectiveness of solution is defined by detection accuracy of more than 80% (i.e. out of 100 smoky vehicles on the road, more than 80 of them are detected by the system), a false alarm rate of less than 10 % (i.e. out of 100 smoky vehicles detected, less than 10 vehicles are not smoky) under reasonably visible conditions

Ability to distinguish and exclude false alarm cases due to environmental conditions such as rain, shadows from clouds, passing vehicles and roadside objects, water splashes from vehicles, and tyre mist from wet roads, etc.

In addition, the proposed solution shall meet the following location-specific requirements:

  1. Roadside Surveillance: Should be able to detect fast-moving smoky vehicles.
  2. Checkpoint: Detection shall take into account the different exhaust pipe configurations of vehicles (e.g. left or right undercarriage exhaust pipe, left or right vertical exhaust pipe).
  3. Processing of videos for detection must be real-time to enable real-time alerts (e.g. SMS) to enforcement officers to intercept and subject smoky vehicles.


Possible Use Case

  • Automated remote detection of smoky vehicles 24/7 without the need for officers onsite
  • Solution is mounted on existing public road or checkpoint infrastructure using e.g. off-the-shelf high-resolution video cameras supported by video analytics and licence plate recognition system
  • For public road operations, the videographic evidence and enforcement information (location, date, time, and vehicle licence plate number) are transmitted to a centralised server for verification and follow up action
  • For checkpoint operations, the photographic evidence and smoky vehicle licence plate number are transmitted (e.g. SMS) in real-time to the enforcement officer stationed there to intercept and subject smoky vehicles to smoke-test 


What's In It For You

  • SGD $50,000 of prize money for each winner of this challenge (see Award Model).
  • NEA can provide low-resolution archived video footages as samples, if required.
  • Potential to work with NEA to operationalise the solution
  • Opportunity to co-develop an innovative Proof-of-Concept which has high commercial value for applications such as automated detection of smoking in prohibited areas, gas leakages, fire detection and smoke emissions from industrial and commercial facilities



We are open to global and Singapore-based individuals, research providers and companies that have digital technology expertise and innovation capabilities. Some of these include Info-Comm and Media (ICM) start-ups, SMEs and larger companies. Problem Solvers should have the ability and desire to co-create and innovate with Problem Owners or/and other partners to translate their ideas into solutions that can address real business needs. 

Problem solvers who choose to apply to this challenge must be registered and operating in Singapore. The POC needs to be demonstrated in Singapore.


Further Details

Technical details of the problem statements as well as information related to the innovation calls can be found on the IMDA Open Innovation Platform (OIP) which you will be redirected to after clicking on the 'Proceed To Details' button below and filling up a short form with your basic information.

Preferred Business Model

  • R&D Collaboration
  • Business Collaboration (Joint Venture, Distribution)

Make an EnquiryProceed to Details