Combined Image Background Subtraction for Motion-Stable Distance Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional background noise removal methods in auto clean apparatuses, which rely on image processing, often result in incorrect target images due to movement, leading to inaccurate distance calculations, especially at high speeds or frame rates.
Innovation Solution
An image processing method involving a light source and image sensor that captures multiple images under different light modes, generates a combined image, and subtracts this from a second image to acquire a target image after background noise removal, ensuring accurate distance computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If background noise removing is performed by subtracting image B from image A, then background noise can be removed, but wrong target images are acquired when the apparatus moves during capture
Solution Approach 1:
The patent applies preliminary action by capturing a reference image (first image) before the measurement image (second image) when the apparatus is at a reference position. This reference image is used to generate a background image that accounts for the apparatus position, ensuring that background subtraction is performed on images taken from the same position, thereby preventing wrong target images during apparatus movement.
Solution Approach 2:
The patent changes the parameter of image selection criteria by introducing position information as a key parameter. Instead of simply subtracting any background image, the system selects and matches images based on apparatus position, ensuring that only images captured at the same position are used for background subtraction, thus maintaining target image accuracy during movement.
2Productivity
If image capture frequency is increased to improve responsiveness, then distance measurement can be performed more frequently, but wrong target images occur more often due to movement
Solution Approach 1:
The patent changes the parameter of image matching from simple temporal sequence to position-based matching. By using position information as the matching criterion, the system can process images at high frequency while ensuring that background subtraction is only performed on images captured at the same position, thus maintaining reliability even with increased measurement frequency.
Solution Approach 2:
The patent applies feedback by using position information to control the background subtraction process. The position information serves as feedback to determine whether the apparatus has returned to the reference position, and only when the position matches does the system perform background subtraction, ensuring accurate target images even at high capture frequencies.
3Adaptability or versatility
If the apparatus moves to perform cleaning operations, then cleaning coverage is improved, but background noise removing becomes inaccurate due to position changes
Solution Approach 1:
The patent applies preliminary action by capturing a reference image at the reference position before the apparatus moves for cleaning operations. This reference image establishes the background at the starting position, and the system uses position matching to ensure that distance measurements are only performed when the apparatus returns to this reference position, maintaining measurement precision while allowing versatile cleaning operations.
Solution Approach 2:
The patent applies local quality by treating the reference position as a special location with specific processing rules. At this local position, the system performs background subtraction using the pre-captured reference image, while at other positions during cleaning operations, the system does not perform background subtraction, thus maintaining measurement accuracy at the reference position while allowing flexible cleaning elsewhere.
Data Source
AI summary
An image processing method, applied to an image processing apparatus comprising a light source and an image sensor. The image sensing method comprises: acquiring a first image via the image sensor if the light source operates in a first mode; acquiring a second image via the image sensor if the light source operates in a second mode; acquiring a third image via the image sensor if the light source operates in the first mode; generating a combined image based on the first image and the third image; and acquiring a target image after background noise removing based on the second image and the combined image.