Object Identification via Angle-Normalized Vector Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object identification technologies in surveillance systems often misjudge whether objects in different monitoring images are the same due to variations in viewing angle, as the front and rear sides of a person or object may exhibit different colors or patterns.
Innovation Solution
An object identification method that captures and compares monitoring images from the same angle of view at different times, using similarity estimation and weighting adjustments based on object dimensions and visible area ratios to determine if the objects are the same, employing image receivers and operation processors to analyze and process the images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional object identification technology compares characteristic vectors from monitoring images captured at different angles, then object identification can be performed, but misjudgment occurs due to angle differences causing diverse colors and patterns
Solution Approach 1:
The patent transforms the comparison parameters from direct characteristic vector comparison to angle-normalized comparison. It calculates the viewing angle of each monitoring image relative to the object's moving direction, then uses this angle information to select or adjust characteristic vectors for comparison, ensuring that objects viewed from different angles are compared using appropriate parameters that account for the angle difference.
Solution Approach 2:
The patent introduces the viewing angle as an intermediary parameter between the monitoring images and the comparison process. By calculating and utilizing the angle information, the system mediates the comparison between characteristic vectors from different angles, selecting or adjusting the vectors based on angle similarity rather than directly comparing vectors from potentially very different viewing angles.
2Measurement precision
If monitoring images are captured from the same angle of view at different times, then identification accuracy improves, but the system complexity increases due to angle calculation and image selection requirements
Solution Approach 1:
The patent performs preliminary calculation of the viewing angle for each monitoring image relative to the object's moving direction before the actual object identification process. This pre-computation of angle information allows the system to efficiently select or adjust characteristic vectors based on angle similarity, avoiding more complex real-time calculations during the identification phase.
Solution Approach 2:
The patent adds the viewing angle as an additional dimension to the object identification process. Instead of only comparing characteristic vectors in the original feature space, the system incorporates angle information as an extra dimension for filtering and selecting images for comparison, enabling more accurate identification while managing complexity through dimensional organization.
Data Source
AI summary
An object identification method determines whether a first monitoring image and a second monitoring image captured by a monitoring camera apparatus have the same object. The object identification method includes acquiring the first monitoring image at a first point of time to analyze a first object inside a first angle of view of the first monitoring image, acquiring the second monitoring image at a second point of the time different from the first point of time to analyze a second object inside the first angle of view of the second monitoring image, estimating a first similarity between the first object inside the first angle of view of the first monitoring image and the second object inside the first angle of view of the second monitoring image; and determining whether the first object and the second object are the same according to comparison result of the first similarity with a threshold.