Slant Correction via Motion Vector Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image correction processing techniques, such as slant correction in video cameras, face challenges in accurately detecting correction parameters without pre-obtained information about the object, especially in varying environments, leading to high operational loads and power consumption, making them unsuitable for mobile devices.
Innovation Solution
An image processing device that generates frame feature data, calculates motion vectors between consecutive frames, specifies a region to be masked based on these vectors, and detects correction parameters from the unmasked region, reducing processing load and power consumption by focusing on background features like architectural structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Hough transform is used to detect the slant angle from frame picture data, then the slant correction accuracy is improved, but the operation load increases significantly
Solution Approach 1:
The patent extracts only the necessary edge information from the frame picture data that is relevant for slant angle detection. By identifying and extracting edge components that indicate the slant angle while excluding other unnecessary image data, the system achieves accurate slant detection without processing the entire high-resolution image, thus reducing the operation load significantly
Solution Approach 2:
The patent applies different processing quality levels to different regions of the image. Instead of uniformly processing all pixels with high computational algorithms, the system focuses computational resources only on edge regions that contain slant angle information, using simpler and more efficient detection methods localized to these critical areas
2Loss of information
If the Hough transform is applied to high-resolution images (1920×1080), then comprehensive image analysis is achieved, but the processing time and power consumption increase
Solution Approach 1:
The patent segments the image processing task into distinct stages: edge detection, edge component extraction, and slant angle detection. By dividing the high-resolution image into edge components first, the system processes only the essential features rather than all image data, significantly reducing processing time while maintaining analysis completeness
3Measurement precision
If conventional slant correction is applied to videos taken while walking, then horizontal orientation correction is achieved, but the correction accuracy decreases due to camera shakes
Solution Approach 1:
The patent implements a dynamic correction approach that adapts to varying shooting conditions. By detecting the slant angle from edge components in real-time and applying corrective rotation dynamically, the system can handle different camera shake patterns and shooting environments, maintaining high correction accuracy whether the camera is held steady or moving
Data Source
AI summary
A motion vector calculation unit calculates inter-frame movement amounts. A masked region specification unit separates the entire edge image of frame feature data into (i) an edge image showing relatively large movement amounts and (ii) an edge image showing relatively small movement amounts. The masked region specification unit then specifies the edge image showing relatively large movement amounts as a region to be masked. This way, a correction parameter is detected from a region other than the masked region. When the correction parameter is a slang angle, a slant correction unit performs slant correction on a frame picture obtained by an image sensor.