Skip to main content Skip to navigation

THE THIRD EYE -Smart Traffic Mirror-

Smart Traffic Mirror can be easily installed onto existing convex mirrors and uses three lights detecting oncoming vehicles, enhancing road visibility and reducing accidents caused by blind spot.

  • cover

  • Intro and prototype

    Intro and prototype

  • How to use

  • interview with Highway Bureau,MOTC(Taiwan)

  • prototype

What it does

Smart Traffic Mirror integrates an AI distance-sensing camera mounted onto existing convex mirrors. When a vehicle approaches, indicator lights will display green, yellow, then red light on how close the approaching car is, alert the driver in real time.


Your inspiration

According to statistics from the Taiwan National Police Agency, there were a total of 358,221 traffic accidents in 2022, and 276,013 of them occurred at intersection blind spots, accounting for 60% of all accidents. As a result, convex mirrors have been installed on roads to improve visibility. However, accidents caused by blind spots still continue to occur. A major reason is that drivers often have difficulty spotting the convex mirrors at night, and the mirrors themselves display small, unclear images of objects.


How it works

THE THIRD EYE -Smart Traffic Mirror uses an AI-based distance detection camera along with a Raspberry Pi motherboard, an HC-SR04 ultrasonic sensor, and the OpenCV object detection algorithm. This enables Smart Traffic Mirror to effectively use green, yellow, and red lights to alert drivers of oncoming vehicles. On the other hand, Smart Traffic Mirror is capable of recognizing vehicle license plates and colors, and uploading that data to the Highway Bureau,MOTC and the National Police Agency for data collection. It wil also detect if the equipment is broken. Using HTTP communication technology combined with RESTful API design, the Raspberry Pi can send error information in JSON format to a designated server via Python’s requests library when an error is detected.


Design process

Our design process ran in parallel with research, seeking to combine distance detection camera with artificial intelligence to solve real-world problems. We make questionnaires to interview some drivers and also use VR equipment to simulate the effects of the Smart Traffic Mirror when real vehicles are coming. During the test, the Smart Traffic Mirror display green, red, and red lights flashing according to the distance of the vehicle to display the distance of the vehicle. What’s interesting is that our original design only had green and red lights. After the test, these drivers told us that they cannot react in time when there is a red light. Just like a traffic light, a yellow light needs to be added to give them some time to react. After hearing what they said, I thought it made sense, so I add the concept of a yellow light. In assessing the feasibility of the product, we engaged with experts from the Taiwan Highway Bureau,MOTC. The Director expressed strong approval, noting its creativity and innovative nature. Furthermore, we learned that the government provides subsidies of NT$200,000 for projects implemented at non-signalized intersections, giving our team a clearer understanding of available public funding support.


How it is different

This is an entirely new product with no direct competitors currently in the market. The traditional convex only reflects the current view and do not actively detect vehicles or alert drivers. Our system uses AI recognition to actively detect approaching vehicles and provide real-time alerts, especially useful in poor visibility at night. This is a completely new product, and there are currently no similar competing solutions in the market. This presents a great opportunity for market entry.


Future plans

1Taiwan Highway Bureau MOTC and TMS Technologies explore cooperation with our team. We are currently in discussions. (Now in progress)We plan to work with government agencies and city departments that manage road safety.We aim to further optimize our system and then conduct real 1:1 on-road testing like mountain roads or sharp corners. We will enhance our product through rigorous testing and return with an improved version.For the next phase, we plan to grow through government contracts and public projects and work with city planners and road contractors to install our system during regular road upgrades, making it easy to use our solution.


Awards

2024 Data Innovation Challenge hosted by the Highway Bureau, MOTC (Taiwan)


End of main content. Return to top of main content.

Select your location