Vehicle Camera Motion Detection Using Distance-Based Image Grids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving object detection technologies, such as those using deep neural networks, require significant data preparation and processing load, leading to delays in detecting moving objects near vehicles.
Innovation Solution
A moving object detection device that calculates difference images between frames, binarizes pixel differences, and extracts grids with varying sizes based on distance from the camera, using a processor to detect moving objects efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep neural network is used for moving object detection, then detection accuracy is improved, but processing load increases and detection speed decreases
Solution Approach 1:
The patent divides the detection process into multiple processing stages: frame acquisition, difference image calculation, grid extraction, and moving object detection. This segmentation allows each stage to be optimized independently, reducing overall processing load while maintaining detection accuracy.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image by dividing the difference image into multiple grids. Each grid can be processed with appropriate threshold values and detection criteria, allowing efficient processing of distant regions while maintaining high accuracy for nearby regions.
2Measurement precision
If deep neural network is used for moving object detection, then detection accuracy is improved, but data preparation requirements increase
Solution Approach 1:
The patent uses the vehicle's own movement information (speed, yaw rate) to automatically correct and align frames, eliminating the need for extensive manual data preparation and labeling. The system leverages readily available sensor data to improve detection accuracy without requiring large annotated datasets.
Solution Approach 2:
The patent performs preliminary frame correction and alignment using vehicle movement data before moving object detection. This preliminary processing prepares the data in advance, reducing the need for complex data preparation and making the detection process more efficient.
3Ease of operation
If uniform grid size is used in difference image, then processing simplicity is maintained, but detection accuracy for distant objects decreases
Solution Approach 1:
The patent applies different grid sizes to different regions based on distance from the camera. Near-region grids use larger sizes for detailed detection, while far-region grids use smaller sizes to capture more detail at distance. This local adaptation improves overall detection accuracy while maintaining processing efficiency.
Solution Approach 2:
The patent dynamically adjusts grid sizes based on the spatial position in the difference image. Rather than using a static uniform grid, the system adapts grid dimensions to the detection requirements of each region, improving accuracy for both near and distant objects.
Data Source
AI summary
Provided is a moving object detection device including a storage medium storing computer-readable commands and a processor connected to the storage medium, the processor executing the computer-readable commands to: acquire image data including a plurality of frames representing a surrounding condition of a mobile object, which are photographed by a camera mounted in the mobile object in time series; calculate a difference image between the plurality of frames by calculating differences between the plurality of frames and binarizing the differences using a first value and a second value; extract a grid for which the density of pixels with the first value is equal to or larger than a first threshold value from among a plurality of grids set in the difference image; and detect the extracted grid as a moving object.