Foot Position Estimation in Distorted Wide-Angle Human Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for estimating foot positions in images captured by cameras with wide-angle or fisheye lenses fail to accurately determine foot positions due to image distortion, leading to incorrect estimations.
Innovation Solution
A foot position estimating device that detects a human region in an image and estimates the foot position as the intersection of a line from a reference point to the vanishing point, correcting the position based on the distance between the reference point and vanishing point to account for distortion, and tracks the foot position over time to determine movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a wide-angle lens or fisheye lens is used to capture a large region, then the monitoring coverage area is improved, but the image distortion increases causing inaccurate foot position estimation
Solution Approach 1:
The patent changes the parameter of foot position estimation by using vanishing point geometry instead of direct pixel coordinate measurement. The estimation unit calculates the foot position as the intersection of a line from reference point to vanishing point with the edge of the human region, which accounts for the distortion characteristics of wide-angle lenses without requiring complex calibration data.
Solution Approach 2:
The patent introduces the vanishing point as an intermediary element to resolve the contradiction. The vanishing point serves as a geometric reference that naturally accounts for the distortion field of wide-angle lenses. By using this intermediary, the system can accurately estimate foot positions in distorted images while maintaining wide monitoring coverage.
2Device complexity
If the lowest point in the human region is used to estimate foot position, then the estimation process is simple, but the accuracy fails in distorted images
Solution Approach 1:
The patent changes the estimation parameter from simple lowest-point detection to vanishing-point-based geometric intersection. This parameter change maintains computational simplicity while achieving accurate foot position estimation in distorted images by leveraging the geometric properties of perspective projection.
Solution Approach 2:
Instead of directly measuring foot position from the human region boundary, the patent inverts the approach by using the vanishing point (a point at infinity in the scene) as the reference. The foot position is derived as the intersection of a line connecting the reference point to the vanishing point with the human region edge, which correctly accounts for distortion.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A foot position estimating device (4) includes a detection unit (21) that detects a human region representing a human in a predetermined region from an image generated by an image capturing unit (2) configured to capture the predetermined region, and an estimation unit (22) that estimates a point of intersection of a line from a reference point in the human region to a vanishing point of the image with an edge of the human region as a foot position of the human, thereby correctly estimating a foot position of the human represented in the image.