Vehicle Object Recognition With Adaptive Headlamp Thresholding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object recognition systems in vehicles fail to accurately distinguish between irradiation areas and light control areas, leading to erroneous or undetected objects, especially when using adaptive high beam systems, which can result in inadequate correction processes.
Innovation Solution
An object recognition apparatus that includes an imaging unit, an object recognition unit, a light distribution control unit, and a reliability threshold change unit, which adjusts the reliability threshold based on the presence of objects in the irradiation area, and applies different image processing filters for irradiation and light control areas to enhance recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single correction threshold value is used for both high beam and low beam states, then the device complexity is reduced, but the object recognition accuracy deteriorates in adaptive high beam systems where irradiation area varies
Solution Approach 1:
The patent segments the headlight beam into two distinct areas: irradiation area (where light is emitted) and light control area (where light is shielded or dimmed). Correspondingly, it applies different correction threshold values to each area - a first correction threshold value for the irradiation area and a second correction threshold value for the light control area. This segmentation allows accurate object recognition in both areas while managing complexity through structured differentiation.
Solution Approach 2:
The patent applies different correction threshold values to different spatial regions of the image based on the headlight beam distribution. The first correction threshold value is applied locally to pixels in the irradiation area, while the second correction threshold value is applied locally to pixels in the light control area. This local quality approach ensures that each region is processed with the appropriate threshold for its specific lighting conditions.
2Reliability
If the reliability threshold is kept high to reduce false detections, then false detections are reduced, but object detection sensitivity deteriorates when no object is present in the irradiation area
Solution Approach 1:
The patent dynamically adjusts the reliability threshold value based on the detected object's position relative to the headlight beam areas. When an object is detected in the irradiation area, a first reliability threshold value is applied. When no object is detected in the irradiation area but an object is detected in the light control area, a second reliability threshold value (different from the first) is applied. This dynamic adjustment optimizes the balance between false detection rate and detection sensitivity based on real-time conditions.
Solution Approach 2:
The patent changes the reliability threshold parameter based on the spatial relationship between detected objects and the headlight beam distribution. The system switches between different reliability threshold values depending on whether objects are present in the irradiation area or light control area, allowing optimization of detection performance for different scenarios without requiring a single fixed threshold.
Data Source
AI summary
To effectively improve the recognition accuracy of the object, provided is an object recognition apparatus comprising an imaging unit for capturing front of a vehicle, an object recognition unit configured to calculate a reliability of an object recognition of an object in the image, and recognizes the object as an object actually present when the reliability is equal to or larger than a threshold value, a light control unit for controlling a light distribution of a headlamp in a light distribution pattern including an irradiation area and a light control area for shielding or dimming the irradiation light based on an existence information of the object, and a threshold change unit configured to changing, when the object in the image does not exist in the irradiation area, the threshold value to a smaller value than when the target is present in the irradiation area.