Method for tracking target based on graph theory cluster and color invariant space
A color-invariant, target tracking technology, applied in the field of target tracking based on graph theory clustering and color-invariant space, which can solve problems such as inability to process color images, mismatching of different moving objects, and limited SIFT.
Active Publication Date: 2013-06-05
NANJING HUICHUAN IND VISUAL TECH DEV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] The problem to be solved by the present invention is: in the existing local invariant feature extraction and matching research, SIFT is limited to grayscale images, cannot process color images, and its matching calculations are large, and there are sometimes false matches between different moving objects. occur
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0042] The present invention is based on the target tracking method of graph theory clustering and color invariant space, extracts feature points, then performs graph theory clustering on the movement trends of these points, classifies feature points into respective targets, and performs Track matches such as figure 1 , including the following processes:
[0043] 1) Carry out color-invariant space conversion to the image of the video stream, use color-invariant feature CSIFT feature extraction, detect and extract color-invariant and scale-invariant features, and calculate invariant feature vectors;
[0044] 2) Graph theory motion clustering of features: Based on graph theory, cluster the feature points of video frames according to the feature motion trend, take the current frame as the reference image, the next frame is the image to be matched, and use the CSIFT features of these two frames Points are used as the nodes of the graph to obtain the motion trend information of ea...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The target tracking method based on graph theory clustering and color-invariant space includes the following steps: 1) Convert the image of the video stream to color-invariant space, and use the color-invariant feature CSIFT feature to extract feature points; 2) Graph theory of features Motion clustering, feature points with the same motion trend belong to targets in the same motion state, and the feature points are tracked to the target in the video frame. The present invention adopts local invariant features and abandons the blindness of global features; realizes CSIFT extraction of color invariant features, increases color invariance while maintaining the advantage of geometric invariance of SIFT features, and upgrades from grayscale feature space to color space; for For moving objects on a static background, graph theory ideas are integrated into matching, and feature motion classification is used as preprocessing for matching, making the matching operation more accurate and faster.
Description
technical field [0001] The invention belongs to the technical field of video image processing, relates to target tracking in video, and is a target tracking method based on graph theory clustering and color invariant space. Background technique [0002] Object tracking has a very wide range of research and applications in the fields of visual navigation, behavior recognition, intelligent transportation and so on. Most of the current motion detection and target tracking algorithms are based on Harris and SIFT features, but it is difficult to solve problems such as matching efficiency and color changes caused by low light differences. Other tracking algorithms use luminance constancy to compute optical flow, however if the intensity distribution is relatively uniform, such as walls with less prominent features, or if it has a one-dimensional intensity distribution, such as edges, the image rate cannot be reliably estimated. [0003] The extraction and matching of image featur...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/20G06K9/62
Inventor 李勃陈抒瑢董蓉翟霄宇顾昊丁文杨娴陈启美郁建桥徐亮
Owner NANJING HUICHUAN IND VISUAL TECH DEV