Image Processing Apparatus Adaptive Threshold for Double Image Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic camera shake correction methods fail to accurately combine images when displacement detection is erroneous, leading to noticeable double images, especially in bright scenes with high edge contrast or during macro shooting, where the displacement amount of object images is large or difficult to detect.
Innovation Solution
An image processing apparatus that sets a reference image and detects motion vectors for other images, comparing the scalar amount of these vectors to a threshold value to determine if they can be combined, and adjusts this threshold based on factors like sensitivity, number of images, shooting mode, distance to objects, presence of people, and image height to prevent double images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If images with high edge contrast are used for electronic camera shake correction, then image capturing quality is improved, but double images become highly noticeable when displacement detection is erroneous
Solution Approach 1:
The patent applies preliminary action by detecting the presence of people in the image before performing pattern matching and displacement detection. This preliminary detection allows the system to adjust the threshold value accordingly, preventing erroneous displacement detection in advance. The people detection step is performed prior to the main image combination process, enabling proactive adjustment of processing parameters to avoid double image artifacts.
2Productivity
If a small number of images are combined for electronic camera shake correction, then processing time is reduced, but double images become highly noticeable when displacement detection is erroneous
Solution Approach 1:
The patent performs people detection as a preliminary step before image combination processing. This allows the system to identify situations where displacement detection is likely to fail and adjust the threshold value in advance, ensuring reliable results even when combining only a small number of images. The preliminary detection enables quality control without significantly increasing processing time.
3Quantity of substance
If the threshold value for displacement limit is set high, then more images can be combined, but erroneous displacement detection leads to noticeable double images
Solution Approach 1:
The patent implements dynamics by making the threshold value adaptive rather than fixed. The threshold is dynamically adjusted based on the detected presence of people in the image. When people are detected, the threshold is set to a value that prevents erroneous displacement detection, even if this reduces the number of combinable images. This dynamic adjustment ensures measurement precision is maintained while allowing maximum possible images to be combined under each condition.
Data Source
AI summary
An image processing apparatus sets one of a plurality of obtained images as a reference image, and detects a motion vector relative to the reference image for the other images. A scalar amount of the detected motion vector is compared with a threshold value representing an image displacement limit for combining an image with the reference image, and whether a condition for combining with the reference image is satisfied is determined. If the condition is satisfied, a corresponding image is moved to cancel the detected motion vector and combined with the reference image, and an image in which displacement between images has been corrected is output. The threshold value is changed by a control unit depending on an image-capturing setting set when the plurality of images are captured, the condition of an object image in the plurality of captured images, or the like.