Map Generation Apparatus for Vehicle Position Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing map generation methods for vehicle position estimation face challenges in managing the increasing data size of maps due to the large number of feature points extracted from in-vehicle camera images, which can overwhelm memory capacity and reduce estimation accuracy.
Innovation Solution
A map generation apparatus that detects the vehicle's surroundings, extracts feature points, divides the map area based on road division lines, and corrects the number of feature points within each divided area by calculating centroids and thinning out non-central points, thereby reducing the data amount while maintaining estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a map is created using all feature points extracted from captured images, then the accuracy of vehicle position estimation is improved, but the data size of the map increases greatly, reducing memory capacity
Solution Approach 1:
The map area is divided into multiple regions based on road division lines (lane markings, curbs, etc.), and feature point correction is performed separately for each region. This segmentation allows selective retention of feature points in different areas, reducing overall data size while maintaining local estimation accuracy.
Solution Approach 2:
Different correction strategies are applied to different regions based on their characteristics. Regions with many feature points undergo more aggressive thinning, while regions with fewer points retain more features. This local differentiation optimizes the balance between data reduction and accuracy preservation.
2Quantity of substance
If the number of feature points in the map is reduced to decrease data size, then memory capacity is improved, but the accuracy of vehicle position estimation deteriorates
Solution Approach 1:
Feature point correction is performed in advance during map generation, dividing the area by road division lines and selectively thinning points before the map is used for position estimation. This preliminary processing ensures that the reduced feature set is optimized for accuracy from the start.
Solution Approach 2:
Road division lines serve as intermediary structures that guide the feature point correction process. These lines (lane markings, curbs) act as references to determine which feature points to retain or remove, ensuring that the reduced set still provides accurate position estimation.
3Area of stationary object
If feature points are uniformly distributed across the map area, then coverage is improved, but the number of points in high-density areas increases unnecessarily, increasing data size
Solution Approach 1:
The map area is segmented into multiple regions using road division lines as boundaries. This segmentation allows the system to handle different areas independently, applying appropriate thinning ratios to each region based on its specific characteristics rather than treating the entire map uniformly.
Solution Approach 2:
The correction ratio for feature point thinning is adjusted based on the number of points in each divided area. Areas with more feature points receive higher thinning ratios, while areas with fewer points receive lower ratios, dynamically optimizing the balance between coverage and data size.
Data Source
AI summary
A map generation apparatus includes: an in-vehicle detection unit configured to detect a situation around a subject vehicle in traveling; and a microprocessor. The microprocessor is configured to perform: extracting one or more feature points from detection data acquired by the in-vehicle detection unit; generating a map using the extracted feature points; dividing an area on the map generated in the generating in a traveling direction of a subject vehicle and dividing the area in the vehicle width direction based on a division line information regarding a division line of a road to form a plurality of divided areas; and correcting a number of feature points for each of the divided areas formed in the dividing. The microprocessor is configured to perform the correcting including reducing the number of feature points based on a distribution of feature points included in each of the divided areas.