Vision System Height Correction for Object Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vision-based object positioning systems inaccurately determine the position of objects located above the ground, such as corners of parking spots in mechanical parking systems, due to the assumption that objects are on the ground.
Innovation Solution
A vision system for vehicles that includes a camera system, a dead reckoning system to calculate vehicle trajectories, and a vision-based positioning system to calculate object trajectories and height values from multiple images. The system identifies and deletes outliers in height values, calculates the object's height, and corrects object coordinates using the height-to-camera-height ratio.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If vision-based positioning assumes objects are on the ground, then the positioning system is simple, but positioning accuracy deteriorates for objects above ground
Solution Approach 1:
The patent introduces the height dimension to the positioning system by calculating the vertical position of objects relative to the camera. The system computes object height values from image data and uses this third dimension (height) to correct the projected 2D position, thereby accurately locating objects above the ground without significantly increasing system complexity.
Solution Approach 2:
The system changes the parameter used for position calculation from ground-level 2D coordinates to 3D coordinates that include height information. By calculating height values from image data and using these to adjust the projected position, the system adapts to objects at different vertical levels while maintaining computational simplicity.
2Measurement precision
If multiple height values are calculated from images, then positioning accuracy improves, but data processing complexity increases
Solution Approach 1:
The system calculates multiple height values from different image frames and uses feedback to identify and remove outliers. By comparing height values across multiple measurements and using statistical criteria (mean and standard deviation), the system refines the final height estimate, improving accuracy while keeping the processing approach systematic and manageable.
Solution Approach 2:
The patent extracts and removes outlier height values from the set of calculated values. By identifying and deleting erroneous measurements (those deviating significantly from the mean), the system isolates the valid height data, thereby improving measurement accuracy without requiring complex processing of all raw data points.
Data Source
AI summary
A vision system for a vehicle includes a camera system including one or more cameras. A dead reckoning system is configured to calculate a first trajectory of the vehicle from a first location to a second location. A vision-based positioning system is configured to calculate a second trajectory of an object located at a height above ground as the vehicle travels from the first location to the second location using a plurality of images generated by the camera system; calculate a plurality of height values for the object from the plurality of images; and identify and delete outliers in the plurality of height values.