Image Processing Apparatus Motion Detection Blending
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods, such as using mean absolute difference (MAD), often mistakenly identify rapidly moving objects as stationary, leading to image streaking due to incorrect blending of frames.
Innovation Solution
An image processing apparatus comprising a pixel buffer unit, a pixel difference calculating unit, a motion level determining unit, a blending value determining unit, and an output unit, which calculates pixel differences and determines whether pixels are static or moving, then applies weighted averages to generate an output image based on blending values, preventing mistaken identification of motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If mean absolute difference (MAD) is used to determine object movement, then the method is simple and widely applicable, but rapidly moving objects are mistakenly identified as stationary, causing image streaking
Solution Approach 1:
The patent divides the image into multiple blocks and processes each block independently. Within each block, it calculates motion vectors by comparing corresponding blocks between frames. This segmentation allows the system to detect motion at a finer granularity level, preventing rapid movements from being averaged out and misclassified as stationary, thereby resolving the contradiction between simplicity and precision.
Solution Approach 2:
The patent introduces a temporal dimension by calculating motion vectors that represent movement between frames. Instead of relying solely on spatial pixel differences (MAD), it adds the time component to motion detection. This dimensional expansion enables the system to accurately capture rapid movements while maintaining computational efficiency, thus improving measurement precision without sacrificing ease of operation.
2Reliability
If pixel differences are calculated between current and previous frames, then object movement can be detected, but rapidly moving objects are mistaken for stationary objects due to insufficient motion level analysis
Solution Approach 1:
The patent dynamically adjusts the motion detection process by calculating motion vectors for each block and comparing them against threshold values. The system adapts its analysis depth based on the detected motion level, using multiple threshold comparisons (first, second, and third threshold values) to reliably classify motion states. This dynamic approach improves reliability while keeping the complexity manageable through algorithmic optimization.
Solution Approach 2:
The patent introduces motion vectors as an intermediary representation between the raw pixel data and the final motion detection result. By calculating motion vectors for each block and using these as intermediaries for further analysis, the system achieves more reliable motion detection without directly comparing every pixel, thus reducing the effective complexity of the analysis process.
3Manufacturing precision
If blending is applied to correct motion detection errors, then image quality can be improved, but incorrect blending causes image streaking instead of correcting the error
Solution Approach 1:
Instead of applying blending to correct detected errors (which causes streaking), the patent inverts the approach by using the motion vector information to directly guide the blending process. It calculates motion vectors first, then uses these vectors to determine appropriate blending weights, ensuring that blending is applied only where and how motion actually occurred. This inversion of the correction logic eliminates the harmful streaking effect while maintaining processing precision.
Solution Approach 2:
The patent changes the blending parameter from a fixed or error-based approach to a motion-vector-based approach. By using motion vectors to determine blending weights, the system dynamically adjusts blending parameters according to actual motion characteristics. This parameter change ensures precise blending that respects the true motion model, improving image processing precision while avoiding the streaking caused by incorrect blending parameters.
Data Source
AI summary
This invention provides an image processing apparatus and an image processing method. By calculation of the pixel difference that is the difference of each corresponding pixels between the current image and the previous image with its neighbor pixel difference, this invention can determine the blending value. Furthermore, the image processing apparatus comprises a pixel buffer unit for storing the current image and the previous image; a pixel difference calculating unit for calculating a plurality of pixel differences; a moving judgment unit, coupled to the pixel difference calculating unit, for determining the pixel moving or static according to by the plurality of pixel differences; a blending value determining unit, coupled to the moving judgment unit, for determining a new blending value when the pixel is static; and an output unit, coupled to the blending value determining unit, for generating an output image by adding the pixels belonging to the same positions pixels on the current image and the previous image with a weighting according to the blending value.