Vehicle Image Feature Selection by Distance for Real-Time Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing systems for vehicle external situation detection face challenges in reducing processing load while maintaining accuracy, particularly in real-time applications, as they do not adequately suppress the number of feature points in image analysis.
Innovation Solution
An image processing apparatus that extracts feature points from a camera image, divides the image into regions, and adjusts the number of feature points based on their distances from the vehicle, prioritizing closer objects for processing to optimize the number of feature points used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of feature points is reduced to decrease processing load, then processing speed improves, but detection accuracy deteriorates
Solution Approach 1:
The patent applies local quality by differentiating feature point selection based on spatial location. Feature points are categorized into nearby objects (within threshold distance) and distant objects, with different selection criteria applied to each region. This allows optimized processing where nearby objects receive more feature points for accurate detection, while distant objects use fewer feature points, resolving the contradiction between processing speed and detection accuracy.
Solution Approach 2:
The patent segments the image processing task by dividing feature points into two distinct groups: those from nearby objects and those from distant objects. This segmentation enables independent optimization of processing parameters for each group, allowing the system to maintain high detection accuracy for critical nearby objects while reducing overall processing load through selective feature point reduction in distant regions.
2Measurement precision
If all extracted feature points are used for processing, then detection accuracy is maintained, but processing load increases
Solution Approach 1:
The patent extracts and separates nearby feature points from the complete set of feature points using distance threshold comparison. By extracting only the necessary nearby feature points for detailed processing while excluding distant feature points from intensive processing, the system maintains detection accuracy for critical objects while significantly reducing overall processing load and energy consumption.
Solution Approach 2:
The patent applies partial action by processing only a subset of feature points (those from nearby objects) with high priority and detailed analysis, while applying reduced or no processing to distant feature points. This selective partial processing maintains sufficient detection accuracy for safety-critical nearby objects while avoiding the excessive processing load that would result from analyzing all feature points equally.
Data Source
AI summary
An image processing apparatus is configured to detect an external situation of a vehicle based on image information acquired by a detection unit mounted on the vehicle. The image processing apparatus includes a microprocessor configured to perform: extracting feature points of an object included in the image information, selecting feature points to be used for processing from a plurality of feature points, dividing the image information into a plurality of regions, and adjusting the number of feature points to be selected in each of the plurality of regions based on distances to objects included in the plurality of regions.