Image Registration via Iterative Perspective Projection Parameter Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for associating captured images with spatial positions in an imaging target range suffer from inaccuracies due to deviations between actual and calculated imaging ranges, leading to poor registration between maps and images.
Innovation Solution
An image processing apparatus that acquires captured images, extracts three-dimensional position information, applies perspective projection transformation based on imaging conditions, evaluates the match rate between transformed and captured image segments, and iteratively adjusts parameters for optimal registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera position and posture are specified from detection unit output signals to calculate imaging range, then registration with map is performed, but deviation between actual imaging range and calculation result is larger, leading to poor registration accuracy
Solution Approach 1:
The patent applies feedback by comparing the calculated imaging range with the actual captured image, detecting deviations between them, and using this comparison information to correct the camera position and posture parameters. This closed-loop feedback mechanism continuously refines the accuracy of the imaging range calculation, reducing deviation and improving registration precision with the map.
Solution Approach 2:
The patent replaces direct reliance on detection unit signals with an image-based verification and correction system. Instead of trusting the mechanical/sensor-based position and posture data alone, the system uses visual feedback from the captured image to substitute and refine the position calculation, achieving higher accuracy through optical measurement rather than purely mechanical sensing.
2Measurement precision
If parameter values of perspective projection transformation are changed to improve registration accuracy, then match rate evaluation is performed multiple times, but processing time and computational complexity increase
Solution Approach 1:
The patent applies partial action by performing match rate evaluation only for specific critical parameters of the perspective projection transformation rather than all parameters simultaneously. The system identifies and adjusts only the most influential parameters that have the greatest impact on registration accuracy, reducing the number of iterations needed while maintaining high precision.
Solution Approach 2:
The patent systematically changes parameter values of the perspective projection transformation based on the detected deviation between calculated and actual imaging ranges. By adjusting parameters in proportion to the measured deviation and evaluating match rates iteratively, the system efficiently converges to optimal parameter values that maximize registration accuracy without requiring excessive computational iterations.
Data Source
AI summary
One or more processors acquire a captured image, acquire three-dimensional position information indicating positions of specific points in a space of an imaging target range, set a value of a parameter of perspective projection transformation of transforming the three-dimensional position information into two-dimensional image coordinates based on an imaging condition of the captured image, transform the position information of the specific points into data of the image coordinates by using the perspective projection transformation, evaluate a rate of match between a first line segment extracted based on the data of the image coordinates obtained by the transformation and a second line segment extracted from the captured image, evaluate the rate of match a plurality of times while changing the value of the parameter of the perspective projection transformation, and associate the captured image with the positions of the specific points based on results of the evaluation.