Bounding Box Calculation for Motion Objects Without Iteration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods require multiple iterations to calculate bounding boxes for motion objects, leading to high hardware costs and potential errors due to image noise, which separates complete objects into unconnected components.
Innovation Solution
An image processing method that converts raw input images into binary images, divides them into blocks, determines moving blocks, and calculates pixel distances to identify connected components without iteration, using a line-based algorithm to determine bounding boxes and reduce hardware costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing methods use multiple iterations to calculate bounding boxes, then measurement precision is improved, but device complexity and hardware costs increase significantly
Solution Approach 1:
The image is divided into multiple blocks, and each block is processed independently to determine moving blocks and connected components. This segmentation allows the bounding box calculation to be performed without multiple full-image iterations, reducing hardware complexity while maintaining precision through systematic block-level analysis
Solution Approach 2:
The method performs preliminary classification of blocks into moving and non-moving categories before conducting connected component analysis. This preliminary action organizes the data structure in advance, enabling the bounding box calculation to proceed directly without requiring multiple iterative passes through the entire image
2Device complexity
If conventional image processing methods use pure software programs to calculate bounding boxes, then hardware costs are reduced, but productivity and processing speed decrease
Solution Approach 1:
By segmenting the image into blocks and processing each block independently with simple hardware logic, the system achieves parallel processing capability. This segmentation enables the hardware implementation to process multiple blocks simultaneously, significantly improving productivity compared to sequential software processing while keeping hardware costs manageable
Solution Approach 2:
The method uses a simplified block-level analysis approach that processes only the necessary information (moving block detection and connected component identification) rather than performing complete image analysis. This partial action approach achieves sufficient precision for bounding box calculation while enabling faster hardware-based processing
3Reliability
If conventional image processing methods process images with noise, then complete objects may be marked as multiple unconnected components, but measurement precision deteriorates
Solution Approach 1:
The method performs preliminary classification of blocks into moving and non-moving categories before conducting connected component analysis. This preliminary organization of data allows the algorithm to systematically identify and connect fragmented components caused by noise, ensuring that complete objects are correctly identified as single connected components rather than multiple separate objects
Solution Approach 2:
The connected component analysis uses feedback from the block classification results to iteratively identify and merge fragmented components. By analyzing the spatial relationships between moving blocks and applying connection criteria, the system can reconstruct complete objects that were separated by noise, improving both reliability and measurement precision
Data Source
AI summary
An image processing method includes: receiving a raw input image; converting the raw input image into a binary input image; dividing the binary input image into N×M blocks; determining whether a specific block is a moving block; determining whether the specific block is a connected image component which is connected to at least one corresponding neighboring block when the specific block is determined as the moving block; and, determining a bounding box according to a result of whether the specific block is the connected image component connected to at least one corresponding neighboring block.