Vehicle Night Vision Template Matching for Camera Angle Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle vicinity monitoring systems face challenges in accurately determining the position and distance of objects relative to a vehicle, especially at varying distances, due to errors caused by slight differences in imaging unit angles and unstable imaging environments, which affects the accuracy of obstacle detection and mounting angle calibration.
Innovation Solution
A vehicle vicinity monitoring apparatus with an imaging unit, an object distance detecting unit, and an object information calculating unit that uses template matching and perspective transformation models to accurately determine object positions and angles, selecting appropriate templates based on distance and using either short-distance or long-distance pin-hole models depending on the object's distance from the vehicle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a simplified perspective transformation model is used for long-distance objects, then calculation speed is improved, but measurement precision deteriorates
Solution Approach 1:
The patent changes the parameter of perspective transformation model selection based on object distance. For distant objects (beyond predetermined distance), a simplified model is used to achieve fast calculation. For nearby objects (within predetermined distance), a general model is used to ensure high precision. This dynamic parameter adjustment resolves the contradiction between calculation speed and measurement precision.
2Measurement precision
If template matching is performed for aiming process, then mounted angle determination is improved, but device complexity increases
Solution Approach 1:
The patent uses template images (copies of target images at various distances) stored in memory to perform matching with actual captured images. This copying approach allows automated determination of mounted angles and object positions without complex manual aiming procedures, improving measurement precision while managing device complexity through pre-prepared templates.
3Measurement precision
If imaging units are mounted with high precision, then position measurement is improved, but manufacturing complexity increases
Solution Approach 1:
The patent implements self-service through automated aiming processes using template matching and parallax-based distance measurement. The system automatically determines mounted angles and corrects position measurements without requiring manual adjustment during manufacturing, thereby improving measurement precision while simplifying the manufacturing process.
4Measurement precision
If multiple templates are stored for different distances, then measurement precision is improved, but memory requirements increase
Solution Approach 1:
The patent changes the template selection parameter based on detected object distance. The system stores multiple templates for different distances but dynamically selects only the appropriate template for the current measurement scenario. This approach improves measurement precision across varying distances while optimizing memory usage by not loading all templates simultaneously.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This solution enables highly accurate calculation of object positions and angles regardless of distance, reducing errors and simplifying the process, even in unstable imaging conditions, by selecting the appropriate template and model based on the object's distance, thus enhancing obstacle detection and mounting angle determination.
Implementation Method 1
an imaging unit for obtaining an image of a vicinity of a vehicle
Implementation Method 2
measuring the distance up to the object based on the parallax between the obtained images
Data Source
AI summary
An ECU of a night vision system stores small obtained images of provisional targets as six templates depending on the distance from infrared cameras, and selects one of the templates depending on the distance up to an actual object. Using the selected template, the ECU performs template matching on images obtained by the infrared cameras, and calculates coordinates of the inspection target. The ECU compares the calculated coordinates of the inspection target and stored reference coordinates with each other, and determines mounted angles of the infrared cameras.