Edge Trajectory Grouping for Tilted Vehicle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional environment recognition techniques struggle to accurately detect vehicles that are tilted or have curved shapes due to the limitations of histogram analysis in the horizontal and vertical directions, leading to reduced detection accuracy.
Innovation Solution
An environment recognition device and method that includes a luminance obtaining unit, edge deriving unit, trajectory generating unit, grouping unit, and specific object determining unit, which derive edge directions based on luminance differences and generate edge trajectories to accurately identify tilted or curved target objects by associating blocks and interpolating edge directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If histogram analysis is used to detect target objects, then the detection process is simple, but the detection accuracy deteriorates for tilted or curved objects
Solution Approach 1:
The patent divides the detection area into multiple blocks and derives edge directions for each block independently. This segmentation allows the system to handle tilted and curved objects by capturing local edge orientations rather than relying on global histogram analysis, thereby improving detection accuracy without significantly increasing overall system complexity.
Solution Approach 2:
The patent transitions from traditional 2D histogram analysis to a multi-dimensional approach by deriving edge directions (orientation information) for each block. This adds an orientation dimension to the detection process, enabling accurate detection of tilted and curved objects that conventional histogram methods cannot handle.
2Adaptability or versatility
If edge extraction is performed on tilted objects, then the edge detection can handle inclined surfaces, but the histogram peak becomes unclear and detection accuracy decreases
Solution Approach 1:
The patent applies local quality by deriving edge directions for each individual block based on its specific luminance characteristics. This allows each block to contribute its local edge orientation information to the overall detection, enabling accurate detection of tilted and curved objects while maintaining clear detection precision through localized analysis rather than global averaging.
3Ease of manufacture
If conventional histogram derivation is used, then the processing is straightforward, but vehicles with curved surfaces cannot be accurately specified
Solution Approach 1:
The patent segments the vehicle detection process into block-level edge direction derivation and subsequent trajectory generation. This segmentation enables the system to capture the curved surfaces of vehicles by accumulating local edge orientation information, achieving accurate vehicle specification while keeping each processing step relatively simple and modular.
Data Source
AI summary
There are provided an environment recognition device and an environment recognition method. The environment recognition device obtains a luminance of each of a plurality of blocks formed by dividing a detection area; derives an edge direction based on a direction in which an edge of the luminance of each block extends; associates the blocks with each other based on the edge direction so as to generate an edge trajectory; groups a region enclosed by the plurality of edge trajectories as a target object; and determines the target object as a specific object.