Scale and illumination adaptive structured multi-target tracking method and application thereof

A multi-target tracking and structured technology, applied in the field of multi-target tracking, can solve problems such as easy tracking failure and poor light adaptability, and achieve the effect of realizing scale adaptation, improving accuracy and robustness, and ensuring tracking accuracy

Pending Publication Date: 2021-04-23
NANJING INST OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, in the article "Structure Preserving Object Tracking" published on Computer Vision and Image Understanding (36(2014): 756-769), Lu Zhang et al. propose a structure suppression-based object tracking algorithm, by learning the graph structure model The elasticity parameter of determines the position of the target in the latest frame; the algorithm adopts the framework of Tracking by detection, and the detection part uses the HOG feature in the Histograms of Oriented Gradients for Human Detection of Computer Vision and Image Understanding (1(2005), 886-893), The SSVM classifier is used to realize target detection, and the structure suppression algorithm is further used to realize multi-target tracking, which effectively improves the tracking efficiency; however, the algorithm has poor adaptability to light and is prone to tracking failure when the target scale changes.

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Scale and illumination adaptive structured multi-target tracking method and application thereof
  • Scale and illumination adaptive structured multi-target tracking method and application thereof
  • Scale and illumination adaptive structured multi-target tracking method and application thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0072] Such as figure 1 As shown, a scale and illumination adaptive structured multi-target tracking method includes the following steps:

[0073] Step S1, using the multi-scale Retinex algorithm to preprocess the grayscaled sequence images (that is, to enhance the grayscaled image sequence using the multi-scale Retinex algorithm), the multi-scale Retinex algorithm, specifically:

[0074] Step S1.1, using the Gaussian function to estimate the illuminance component G(u,v), the calculation formula is as follows:

[0075]

[0076] Among them, (u, v) is the pixel point coordinates of the image, and δ is the scale parameter;

[0077] Step S1.2. Bring the illuminance component G(u, v) into the multi-scale Retinex expression to obtain the reflection image R of the essential characteristics of the object msr (u, v), the calculation formula is as follows,

[0078]

[0079] Among them, S(u, v) is the input sequence image, ρ q is the weighting factor for the qth scale, and St...

Embodiment 2

[0126] Such as figure 1 As shown, a scale and illumination adaptive structured multi-target tracking method includes the following steps:

[0127] Step S1.1. Use the multi-scale Retinex algorithm to preprocess the grayscaled sequence images. First, use the Gaussian function to estimate the illuminance component G(u,v), and the calculation formula is as follows:

[0128]

[0129] Among them, (u, v) is the pixel point coordinates of the image, and δ is the scale parameter;

[0130] Step S1.2. Bring the illuminance component G(u, v) into the multi-scale Retinex expression to obtain the reflection image R of the essential characteristics of the object msr (u, v), the calculation formula is as follows,

[0131]

[0132] Among them, S(u, v) is the input sequence image, ρ q is the weighting factor for the qth scale, and

[0133] Step S2.1. Manually select the multi-target rectangular frame area to be tracked from the first frame. The scale of target i is recorded as w(i, ...

Embodiment 3

[0176] Such as figure 1 As shown, a scale and illumination adaptive structured multi-target tracking method includes the following steps:

[0177] Step S1, using the multi-scale Retinex algorithm to preprocess the grayscaled sequence images, wherein the multi-scale Retinex algorithm is specifically:

[0178] Step S1.1, using the Gaussian function to estimate the illuminance component G(u,v), the calculation formula is as follows:

[0179]

[0180] Among them, (u, v) is the pixel point coordinates of the image, and δ is the scale parameter;

[0181] Step S1.2. Bring the illuminance component G(u, v) into the multi-scale Retinex expression to obtain the reflection image R of the essential characteristics of the object msr (uv), the calculation formula is as follows,

[0182]

[0183] Among them, S(u, v) is the input sequence image, S(u, v) is the weight factor of the qth scale, and

[0184] Step S2.1. Manually select the multi-target rectangular frame area to be trac...

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

No PUM Login to view more

Abstract

The invention discloses a scale and illumination adaptive structured multi-target tracking method, which comprises the following steps of: preprocessing a sequence image by using a multi-scale Retinex algorithm, extracting HOG (Histogram of Oriented Gradient) features of the appearance of a target region, classifying by using a structured support vector machine (SSVM), and taking a target optimal position obtained by the SSVM as a center; extracting a plurality of HOG features of different scales from the center of a target position to serve as samples, wherein a discriminant scale space tracking algorithm DSST is adopted to train a scale filter, and the scale of a target corresponding to a current frame is updated by utilizing the scale corresponding to the maximum response value of the scale filter; updating the spatial position relationship between the targets and the weight value of the SSVM through a stochastic gradient descent (SGD) method, and adjusting the scale weight of the SSVM according to the scale of the current frame target and by using bilinear interpolation. The method can be applied to the fields of intelligent video monitoring, enterprise production automation and intelligent robots.

Description

technical field [0001] The invention relates to a multi-target tracking method, in particular to a scale and illumination self-adaptive structured multi-target tracking method and its application. Background technique [0002] Object tracking is one of the important research directions in the field of computer vision, and it has a wide range of applications in public security monitoring and management, medical image analysis, behavior understanding, visual navigation, etc. At present, scholars at home and abroad are mainly concerned with the tracking robustness and accuracy improvement under the conditions of similar target interference, target scale change, blurred appearance, occlusion, and real-time performance of the target tracking system in practical applications. [0003] At present, in the article "Structure Preserving Object Tracking" published on Computer Vision and Image Understanding (36(2014): 756-769), Lu Zhang et al. propose a structure suppression-based objec...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/46G06K9/62
CPCG06V20/41G06V20/46G06V20/48G06V10/40G06F18/2411
Inventor 路红花湘陈桂彭俊胡云层
Owner NANJING INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap