Monocular Camera Time to Collision Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating time to collision (TTC) in Advanced Driver Assistance Systems (ADAS) and Autonomous Driving rely heavily on costly Lidar and radar technologies, which are not feasible for consumer vehicles and have poor angular resolution.
Innovation Solution
A computer vision system using a monocular camera estimates TTC by determining the 3D position and velocity of detected objects relative to the camera, incorporating a ground plane estimation and Kalman filters to compute the TTC based on distance, projected velocity, and acceleration, while determining if objects are in the vehicle's trajectory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Lidar is used for TTC estimation, then measurement precision is improved, but cost increases significantly
Solution Approach 1:
The patent uses a monocular camera to capture 2D images and computationally reconstructs 3D information, effectively creating a virtual copy of Lidar's depth perception capability through software algorithms rather than expensive hardware
Solution Approach 2:
The patent replaces the mechanical/optical Lidar system with a computational vision system using a standard camera and image processing algorithms, substituting physical measurement mechanisms with computational methods
2Ease of manufacture
If radar is used for TTC estimation, then cost is reduced, but angular resolution deteriorates
Solution Approach 1:
The patent transitions from radar's limited angular measurement to computer vision's pixel-based 2D spatial resolution, adding dimensional precision through the camera's sensor array which provides much finer angular discrimination in the visual spectrum
3Ease of manufacture
If monocular camera is used for TTC estimation, then cost is reduced, but 3D position computation complexity increases
Solution Approach 1:
The patent extracts 3D position information from 2D images by isolating and processing specific features (object detection, ground plane intersection points) separately from the full image data, reducing computational burden through selective analysis
Solution Approach 2:
The patent introduces intermediate computational steps including object detection, ground plane estimation, and feature point extraction as mediators between the raw 2D image and the final 3D position calculation, breaking down the complex problem into manageable stages
Data Source
AI summary
A method for estimating time to collision (TTC) of a detected object in a computer vision system is provided that includes determining a three dimensional (3D) position of a camera in the computer vision system, determining a 3D position of the detected object based on a 2D position of the detected object in an image captured by the camera and an estimated ground plane corresponding to the image, computing a relative 3D position of the camera, a velocity of the relative 3D position, and an acceleration of the relative 3D position based on the 3D position of the camera and the 3D position of the detected object, wherein the relative 3D position of the camera is relative to the 3D position of the detected object, and computing the TTC of the detected object based on the relative 3D position, the velocity, and the acceleration.