Video-to-Map Positioning Using Multilayer Homography
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for tracking customer movement in retail environments suffer from decreased accuracy when the feet of individuals are hidden, leading to errors in coordinate conversion from camera images to maps.
Innovation Solution
Employ a multilayer homography approach using multiple homography matrices to convert coordinates of visible body parts, such as the waist and head, to accurately determine the position on a map even when feet are obscured.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If coordinates of feet are used for position determination, then position accuracy is improved when feet are visible, but position accuracy deteriorates when feet are hidden
Solution Approach 1:
The patent segments the person into multiple parts (head, body, feet) and tracks each part separately. When feet are visible, they provide accurate position data; when feet are hidden, other segments (head or body) are used instead. This segmentation allows the system to maintain position determination reliability by switching between different body segments based on visibility conditions.
Solution Approach 2:
The patent changes the parameter used for position determination based on the visibility condition. When feet are visible, the system uses foot coordinates; when feet are hidden, it switches to using head or body coordinates. This parameter change ensures that position accuracy is maintained regardless of whether feet are visible or hidden.
2Ease of operation
If homography matrix is applied to convert coordinates, then coordinate conversion is simplified, but conversion accuracy deteriorates when feet are hidden
Solution Approach 1:
The patent makes the coordinate conversion process dynamic by selecting different body segments (feet, head, or body) as the basis for conversion depending on visibility conditions. The system dynamically adjusts which segment's coordinates are converted using the homography matrix, ensuring conversion accuracy is maintained even when feet are hidden, while still utilizing the simplicity of homography matrix transformation.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
An analysis program causess a computer to execute a process including specifying (S103) a plurality of portions of a person included in an image configuring a video obtained by capturing in a facility by analyzing the video, converting (S203, S204) each coordinate position on the image of the plurality of specified portions of the person into each coordinate position on a map in the facility, extracting (S104) any target portion among the plurality of portions based on each positional relationship of the converted coordinate positions on the map, and setting (S106, S107) the coordinate position of the extracted target portion on the map as a position where the person exists.