Vehicle Object Detection Switching Recognition Dictionaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection apparatuses in vehicles face reduced detection performance due to varying environmental conditions caused by changes in external environments and vehicle operating states, such as different lighting conditions, which affect the visibility of target objects.
Innovation Solution
An object detection apparatus that stores multiple image recognition dictionaries and techniques, allowing it to switch between them based on the vehicle's lighting device states, such as headlight on/off or wiper operation, to adjust image recognition processing parameters like brightness, contrast, and sharpness for optimal detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single image recognition dictionary and technique are used, then the device complexity is reduced, but the detection reliability deteriorates under varying environmental conditions
Solution Approach 1:
The system dynamically switches between multiple image recognition dictionaries and techniques based on detected environmental conditions (headlight state, wiper state). This allows the detection system to adapt its parameters in real-time to match current lighting and weather conditions, thereby maintaining high detection reliability without requiring a single overly complex universal system
Solution Approach 2:
The system changes recognition parameters (dictionary selection, technique selection) based on environmental parameter changes such as headlight on/off state and wiper operation state. By adjusting the recognition parameters to match environmental conditions, the system maintains reliable detection across varying conditions without increasing overall device complexity
2Reliability
If multiple image recognition dictionaries and techniques are stored for different conditions, then the detection reliability improves, but the device complexity increases
Solution Approach 1:
The image recognition system is segmented into multiple specialized dictionaries and techniques, each optimized for specific environmental conditions (e.g., daytime, nighttime, rainy conditions). This segmentation allows the system to store multiple condition-specific models rather than one large universal model, improving reliability for each condition while organizing storage in a manageable, structured way
Solution Approach 2:
Multiple image recognition dictionaries and techniques are pre-prepared and stored in advance for different environmental conditions. The system performs preliminary classification of current conditions (headlight state, wiper state) and selects the appropriate pre-prepared dictionary and technique, avoiding the need for real-time generation or complex processing of multiple models
3Measurement precision
If image recognition is performed without adjusting for lighting conditions, then the processing speed is maintained, but the detection precision deteriorates
Solution Approach 1:
The system performs preliminary detection of environmental conditions (headlight state, wiper state) before executing image recognition. Based on this preliminary information, it selects the most appropriate pre-prepared dictionary and technique that matches current lighting conditions. This preliminary selection avoids the need for time-consuming real-time image adjustment or trial-and-error recognition attempts, maintaining processing speed while improving detection precision
Solution Approach 2:
The system changes recognition parameters (dictionary and technique selection) based on lighting condition parameters (headlight state, wiper state). By matching recognition parameters to lighting conditions, the system achieves high detection precision without requiring complex real-time image processing adjustments that would slow down processing
Data Source
AI summary
An object detection apparatus mounted in a vehicle for detecting a target object in various changing environmental conditions. In the apparatus, a storage prestores plural image recognition dictionaries each describing reference data for the target object, and plural image recognition techniques each used to detect the target object from an input image with use of one of the plural image recognition dictionaries. A first acquirer acquires an operating state of a lighting device of the vehicle. A selector selects, according to the acquired operating state of the lighting device, one of the plural of image recognition dictionaries and one of the plural of image recognition techniques. A detector detects the target object in the input image by applying image recognition processing thereto with use of the selected image recognition dictionary and technique.