Image Processing Apparatus Moving Object Detection Euclidean Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies fail to effectively detect moving objects that move in the same direction as the background, leading to difficulties in separating subjects from the background, especially when both move at different speeds.
Innovation Solution
An image processing apparatus and method that calculates a background vector based on motion vectors and detects moving objects by determining the Euclidean distance between these vectors and the background vector, allowing for accurate identification and separation of moving objects even when they move in the same direction as the background.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion vectors are used to detect moving objects based on angle difference from majority motion vectors, then moving objects moving in opposite direction can be detected, but moving objects moving in the same direction as background cannot be separated from background
Solution Approach 1:
The patent changes the detection parameter from angular difference to Euclidean distance in motion vector space. By representing motion vectors as 2D vectors with magnitude and direction, and calculating Euclidean distance between each motion vector and the background motion vector, the system can detect objects regardless of whether they move in the same or opposite direction, thus resolving the limitation of angle-based detection
Solution Approach 2:
The patent introduces a background motion vector as an intermediary reference. Instead of comparing motion vectors directly against each other, the system first establishes a background motion vector representing the overall background movement, then uses this intermediary to measure the deviation of individual motion vectors, enabling separation of foreground objects from background
2Measurement precision
If main subject is determined based on difference between overall image movement and local motion, then large subjects with large motion difference can be recognized, but not all user-intended main subjects are large subjects with large motion difference
Solution Approach 1:
The patent changes the measurement parameter from motion difference magnitude to Euclidean distance in vector space. This allows detection of subjects based on their motion vector characteristics rather than just the magnitude of motion difference, enabling detection of smaller subjects or subjects with subtle motion differences that were previously undetectable
Solution Approach 2:
The patent uses Euclidean distance thresholding to detect motion vectors that exceed a certain distance from the background vector. This partial detection approach focuses on vectors with significant deviation while ignoring minor variations, effectively identifying main subjects without requiring them to be large or have extremely large motion differences
Data Source
AI summary
A disclosed image processing apparatus calculates a background vector expressing a motion of a background based on a plurality of motion vectors detected between a plurality of images. Then the image processing apparatus detects a motion vector of a moving object from the plurality of motion vectors, based on a magnitude of Euclidean distance between each of the plurality of motion vectors and the background vector.