Night Vision Image Resolution Reduction for Collision Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing night vision systems in motor vehicles face challenges in reliably detecting potential collision objects under varying driving and traffic conditions, especially at night when glare from headlights impairs vision, leading to increased accident risks.
Innovation Solution
A method and device that generate a second intermediate image by reducing the resolution of the original image, allowing a detector to evaluate it for specific object categories, such as pedestrians or cyclists, without the need for adjustments in object size, enabling early warning signals to be generated based on object distance and vehicle speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the original high-resolution image is evaluated directly by the detector, then detection precision for small objects is improved, but processing time and computational load increase
Solution Approach 1:
The image evaluation process is segmented into two stages: first evaluating a downsampled low-resolution image for quick processing, and only if needed, evaluating the original high-resolution image. This segmentation allows the system to achieve fast processing while maintaining detection precision through selective high-resolution analysis.
Solution Approach 2:
The system applies partial action by evaluating only the downsampled image for initial detection, and only performs full high-resolution evaluation when necessary. This partial approach reduces processing time while maintaining sufficient detection precision for most cases.
2Reliability
If the detector is trained for specific object size ranges, then detection reliability for target objects is improved, but adaptability to varying object distances decreases
Solution Approach 1:
The system dynamically adjusts the image resolution based on object distance estimation. For nearby objects, it uses high-resolution evaluation to maintain detection reliability. For distant objects, it uses downsampled evaluation, effectively adapting the detection parameters to the varying conditions while maintaining reliability across different distances.
Solution Approach 2:
The system changes the image resolution parameter dynamically based on the detected object's characteristics and distance. By adjusting whether to use original or downsampled images, the system maintains detection reliability across varying object distances while optimizing processing efficiency.
3Illumination intensity
If night vision systems use active infrared illumination, then visibility in dark conditions is improved, but glare from oncoming headlights impairs vision
Solution Approach 1:
The system creates a downsampled copy of the original image for initial evaluation, which allows processing of night vision images without the glare interference affecting detection. This copying approach enables the system to work with the night vision imagery while mitigating the harmful glare effect through selective processing of the replicated image data.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the reliability of detecting potential collision objects, reducing the likelihood of accidents by providing timely warnings to the vehicle driver, even in conditions where night vision is impaired, and allows for the recognition of heat-radiating objects like people and animals.
Implementation Method 1
Passive night vision systems capture ambient infrared light emitted by objects
Implementation Method 2
A first intermediate image with a predetermined first intermediate image size is generated by reducing a resolution of the original image in the sense of reducing pixels
Data Source
Figure 1~2
Figure 3~4
AI summary
A recording unit is arranged in a motor vehicle, said recording unit being designed make available, starting from an image that it has captured, a digital original image having a predefined image size. A first intermediate image having a predefined first intermediate image size is produced by reducing a resolution of the original image in the sense of reducing pixels. Furthermore, a second intermediate image having the predefined image size is produced such that it comprises the first intermediate image. The second intermediate image is evaluated by means of a predefined detector in order to check whether there is an object of a predefined object category in the second intermediate image, the detector being designed to evaluate a predefined image detail and to identify an object of a predefined object category having a predefined object size range.