Vehicle Hazard Detection and Alert System
Zone: Field House
Location: Gordon Field House and Activities Center (GOR/024) -
Time: All Day
The goal of the Imaging Science Graduate Project is to develop and build a system that will detect road hazards from warm-blooded animals the size of a house cat and larger and provide a driver with an alert that is sufficient for avoiding a collision. This system is referred to as the Hazard Detection and Alert System, or HDAS for short. The Imagine RIT exhibit will represent the culmination of this two semester, multidisciplinary effort by a team of 18 PhD students from the Chester F. Carlson Center for Imaging Science. The HDAS is comprised of many different components. For detecting hazards, the HDAS utilizes two forward looking thermal-IR cameras mounted to the hood of a truck with overlapping fields of view. Video from these cameras are fed to an image processing algorithm which filter out the road hazards from the rest of the scene. The algorithms used were developed by the HDAS team in-house and are based on the latest in image processing technology. Information from the algorithms is then passed to an alert system which provides the driver with both visual and audio cues which provide the location and severity of the road hazard. The audio cues are provided by stereo speakers which indicate whether or not the hazard is on the left or right hand side and the threat severity by repetition rate of the beeping cue. The visual cue consists of an array of twelve eye-safe lasers which shine onto reflective surfaces on the windshield. the location of the cue, which the driver sees as a bright dot, corresponds to the location of the hazard from the drivers point of view. The Imagine RIT exhibit will consist of a scaled demonstration of the alert system as well as a video presentation showing the image processing algorithm in action. For the alert demonstration, a visitor will have the opportunity to sit inside an HDAS simulator. The simulator will have a screen showing video from the perspective of a person driving down a street and will end with a deer suddenly popping out. Between the visitor and screen will be a glass pane simulating the HDAS windshield. The visitor will be alerted to the oncoming deer just as they would in the full system, with visual cues provided by laser spots reflected off the reflective coating on the windshield and audio cues from the speakers. In addition to the HDAS simulator, a visitor will also have the opportunity to see how the image processing algorithm performs. This will be done by a presentation which shows the algorithm identifying road hazards in thermal video. The thermal video will be actual HDAS test data taken while the system was being driven down a road.
Douglas Macdonald, Brittany Ambeau, Jie Yang, Kamran Binaee, Kelly Anderson, Lauren Taylor, Osborn De Lima, Viraj Adduru