Vehicle External Environment Recognition Apparatus Sidewall Occlusion Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle external environment recognition systems face challenges in accurately tracking preceding vehicles, especially on curved roads, due to potential hiding by sidewalls or other vehicles, leading to mistaken identification of objects.
Innovation Solution
A vehicle external environment recognition apparatus that uses processors and memory to calculate three-dimensional positions of blocks in images, group them to determine objects, identify preceding vehicles, and estimate their future positions, while determining if they are hidden by sidewalls or other vehicles, using pattern matching and stereo methods to improve tracking accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system tracks preceding vehicles using standard object recognition methods, then tracking can be performed, but tracking accuracy deteriorates on curved roads where vehicles may be hidden by sidewalls or other vehicles
Solution Approach 1:
The system performs preliminary actions by predicting the future position of the preceding vehicle before it becomes completely hidden. The prediction unit calculates where the vehicle will be in the next frame based on current motion trends, allowing the system to maintain tracking continuity even when the vehicle temporarily enters a blind region obscured by sidewalls or other vehicles.
Solution Approach 2:
The prediction unit acts as an intermediary between the object recognition unit and the tracking unit. When the recognition unit fails to detect the vehicle due to occlusion, the prediction unit provides intermediate position estimates that bridge the detection gap, allowing the tracking unit to maintain continuous tracking without interruption.
2Measurement precision
If the system expands the detection range to cover blind regions, then more vehicles can be detected, but false identification of sidewalls as vehicles increases
Solution Approach 1:
The system dynamically adjusts the detection range based on the predicted position and motion of the preceding vehicle. Rather than maintaining a static expanded detection range that increases false positives, the detection range is adaptively focused on regions where vehicles are likely to be located, reducing unnecessary detection of sidewalls and other static objects.
Solution Approach 2:
The system applies different detection strategies to different spatial regions. In areas where vehicle presence is highly probable (based on prediction), the system performs detailed vehicle-specific detection. In other regions, particularly areas likely to contain sidewalls or static structures, the system applies more restrictive detection criteria to avoid false positives.
Data Source
AI summary
A vehicle external environment recognition apparatus to be applied to a vehicle includes one or more processors and one or more memories configured to be coupled to the one or more processors. The one or more processors are configured to: calculate three-dimensional positions of respective blocks in a captured image; group the blocks to put any two or more of the blocks that have the three-dimensional positions differing from each other within a predetermined range in a group and thereby determine three-dimensional objects; identify each of a preceding vehicle of the vehicle and a sidewall on the basis of the determined three-dimensional objects; and track the preceding vehicle. The one or more processors are configured to determine, upon tracking the preceding vehicle, whether the preceding vehicle to track is to be hidden by the sidewall on the basis of a border line between a blind region and a viewable region.