Moving Object Detection via Background Model Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving-object recognition systems face challenges in accurately detecting and tracking moving objects, particularly in dynamic environments such as roads, due to changes in lighting conditions, object orientation, and occlusions, leading to low detection accuracy and frequent misidentification of recognition targets.
Innovation Solution
A processing apparatus and method that utilize a distance image acquirer, moving-object detector, and background recognizer to enhance moving-object recognition accuracy by generating a background model and adjusting detection methods based on the relative positional relation between moving objects and the background model, allowing for continuous tracking and prioritization of recognition targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a parallax image is used to detect moving objects, then the measurement accuracy for distance is improved, but the recognition accuracy of moving objects deteriorates
Solution Approach 1:
The patent divides the detection process into two independent stages: first detecting moving objects in the distance image to obtain distance information, then detecting the same objects in the original image to obtain recognition information. This segmentation allows each stage to optimize for its specific purpose without compromising the other, resolving the contradiction between distance measurement accuracy and object recognition accuracy.
Solution Approach 2:
The patent introduces a background model as an intermediary element that helps distinguish moving objects from stationary objects. The background model is generated by detecting stationary objects in the distance image and using this information to filter out false detections, thereby improving moving-object recognition accuracy while maintaining distance measurement precision.
2Measurement precision
If a background model is generated from distance image, then the detection accuracy of moving objects is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary actions by generating the background model and pre-processing the distance image before the actual moving-object detection. The background model is created once and reused for multiple detection tasks, which reduces the processing time required for each individual detection operation while maintaining high accuracy.
Solution Approach 2:
The patent applies partial action by only processing the regions of the distance image where moving objects are detected, rather than processing the entire image. This selective processing approach reduces the overall processing time while maintaining detection accuracy in the relevant regions.
Data Source
AI summary
A processing apparatus includes a distance image acquirer, a moving-object detector, a background recognizer. The distance image acquirer acquires a distance image containing distance information of each pixel. The moving-object detector detects a moving object from the distance image. The background recognizer generates a background model in which a stationary object recognized as background of the moving object is modeled, from the distance image acquired by the distance image acquirer. The moving-object detector changes a method for detecting the moving object, based on a relative positional relation between the moving object and the background model.