UVU Computer Engineering Capstone Project

Adaptive Video Projector Headlights (2019-2020)

Project Sponsor: Students
Team Members: Christian Garrett and Jason Wright

Faulty Advisor: Dr. Afsaneh Minaie

  • Hardware layout overview diagram
  • Feature breakdown of raspberry pi
  • Thermal image of raspberry pi
  • LCD projector operation
  • Light detection and regional dimming process flow
  • Luminous channel image
  • Luminous channel with applied threshold filter/detected points
  • X (bottom left) and Y (top right) axis luminance summation graphs
  • Filtered Luminance Peaks with False Positives
  • Filtered Luminance Peaks

In the United States, personal automotive vehicles make up the majority of transportation. Automotive vehicle accidents are also a majority contributor to national deaths. With the invention of modern-day computers and advances in computing, cars have become more technologically advanced than ever before in history. Many of these advancements have been to enhance the comfort and safety of this automotive transportation. This includes the potentially life-changing development of the adaptive headlight.

The adaptive headlight is a device that augments the abilities of the modern automotive headlight. Currently, most single headlight assemblies only allow for a few modes of main beam outputs. Two of the current headlight setups use reflector or projector-style bulbs. Although these assemblies and setups allow for simple and sufficient visibility for the driver, they do little beyond this in terms of being adaptive. Some systems seek to enhance these headlight capabilities by allowing auto-switching between low and high beams. Other adaptive headlight systems exist that use electronic sensors on the car to adjust the direction of the vehicle's headlights. In the United States, this adaptive headlight system is one of the most advanced DOT (Department of Transportation) approved headlight systems available. In Europe, however, Audi has developed the Matrix LED headlights, which can modify the light output of a matrix of LEDs based on current driving conditions through the use of a mounted camera. The headlight adjusts to oncoming vehicles, turns and more.

Using this inspiration, the proposed adaptive headlight system uses a single-board computer (Raspberry Pi 4), a camera, and a video projector to create a simpler and more versatile headlight that allows for object detection and locality dimming of light, among other capabilities. The current scenario programmed into the adaptive headlight is for the system to detect an oncoming vehicle. The vehicle is tracked, and light shined out of the video projector is locally dimmed in a rectangle around the oncoming vehicle. This allows the safety-targeted headlight to continuously be in a high-beam state while also creating a non-blinding scenario for oncoming vehicles on the road. The device utilizes OpenCV to detect bright objects, which are then analyzed to determine whether the object is an oncoming vehicle. After detecting an object, the video output to the projector is adjusted in a calibrated manner to avoid shining on the vehicle. A video projector and web camera are used for versatility purposes, and a Raspberry Pi 4–based design allows for many more future capabilities to be added to the adaptive headlight system through simple software updates. 

Students' Presentation