Image Processing Apparatus Noise Evaluation for Low-Light Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In low-illuminance environments, existing imaging technologies that combine visible and invisible light images suffer from significant noise, leading to reduced recognizability of objects in the output images due to the influence of noise from the visible light image.
Innovation Solution
An image processing apparatus that evaluates the brightness and noise levels in an imaging region, switching between outputting a visible light image, a combined image, or an invisible light image based on threshold evaluations to minimize noise and improve image recognizability, by using a control unit to generate and output either a combined image or an invisible light image when the brightness falls below a certain threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If visible light image and invisible light image are combined to capture color image in low-illuminance environment, then color reproduction capability is improved, but noise level increases significantly
Solution Approach 1:
The patent uses an evaluation unit as an intermediary to assess noise levels in the combined image, and based on this evaluation, selectively outputs either the combined color image or the invisible light image. This intermediary evaluation mechanism resolves the contradiction by introducing a decision layer that prevents harmful noise from being propagated to the final output while preserving the beneficial color reproduction capability when conditions are favorable.
Solution Approach 2:
The patent changes the output parameter selection based on evaluated noise levels. When noise in the combined image exceeds a threshold, the system switches from outputting the combined color image to outputting the invisible light image instead. This parameter change strategy resolves the contradiction by dynamically adjusting which image type is output based on real-time noise evaluation, thereby maintaining color reproduction capability when possible while avoiding noise propagation when necessary.
2Loss of information
If visible light image is output in low-illuminance environment, then color information is preserved, but recognizability deteriorates due to high noise
Solution Approach 1:
The evaluation unit acts as an intermediary between image generation and output, assessing whether the combined image contains excessive noise. This intermediary evaluation prevents the system from outputting low-quality images with poor recognizability while preserving color information capability, thereby resolving the contradiction between information preservation and reliability.
Solution Approach 2:
The system implements feedback by evaluating noise levels in the combined image and using this evaluation to control the output selection. The feedback mechanism compares the evaluated noise level against a threshold and adjusts the output accordingly, resolving the contradiction by ensuring that color information is only output when it does not compromise object recognizability.
3Illumination intensity
If combined image is always output in low-illuminance environment, then color reproduction is enhanced, but detection accuracy decreases due to noise influence
Solution Approach 1:
The evaluation unit serves as an intermediary that assesses the quality of the combined image before output. By introducing this evaluation layer, the system can enhance color reproduction through combined imaging when noise levels are acceptable, while preventing detection accuracy degradation when noise levels are high, thereby resolving the contradiction between color enhancement and measurement precision.
Solution Approach 2:
The system changes the output parameter from always outputting combined images to selectively outputting based on noise evaluation. When noise levels exceed the threshold, the system switches to outputting invisible light images instead, thereby maintaining detection accuracy while preserving color reproduction capability when conditions permit, resolving the contradiction between these two objectives.
Data Source
AI summary
There is provided with an image processing apparatus. An obtaining unit obtains a visible light image and an invisible light image of an imaging region. A first evaluation unit evaluates a brightness of the imaging region. A second evaluation unit evaluates noise in an object in the imaging region. A combining unit generates a combined image by combining the visible light image and the invisible light image. An output unit outputs either the combined image or the invisible light image in accordance with an evaluation result on the noise without outputting the visible light image in response to the brightness becoming lower than a first threshold during an operation of outputting the visible light image.