Pedestrian Distance Estimation Using Static Point Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for autonomous driving struggle to accurately estimate the distance between a pedestrian and a camera, especially when the camera is tilted or the ground is sloped, due to inaccuracies in positioning the pedestrian.
Innovation Solution
A method and apparatus that detect a pedestrian region in images from a camera, determine a static point within this region, and calculate the distance to the camera based on the static point and the camera's position, using techniques such as structure from motion (SFM) and deep learning-based object recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If camera projection geometry is used to calculate distance from floor line, then distance estimation is simplified, but accuracy deteriorates when camera is tilted or ground is sloped
Solution Approach 1:
The pedestrian detection region is segmented into multiple candidate regions, and a static point selection module selects the most appropriate static point from these candidates. This segmentation approach allows the system to handle different scenarios (tilted camera, sloped ground) by choosing the best candidate point, thereby maintaining accuracy without significantly increasing overall system complexity.
2Productivity
If simple floor line projection is used, then processing speed is improved, but reliability deteriorates in complex scenarios
Solution Approach 1:
The system performs preliminary actions by pre-identifying multiple candidate static points and their corresponding floor lines before final distance calculation. This preliminary segmentation and candidate selection enables the system to quickly evaluate multiple possibilities and choose the most reliable one, maintaining real-time processing speed while improving reliability in complex scenarios like tilted cameras or sloped grounds.
Data Source
AI summary
A method with distance estimation includes detecting a pedestrian region of a pedestrian comprised in a plurality of images received from a camera; determining a static point in the detected pedestrian region; and determining a distance between the pedestrian and the camera based on the static point in each of the images and a position of the camera corresponding to each of the images.