Target tracking algorithm based on fusion 2D detection

A target tracking and algorithm technology, applied in computing, computer components, instruments, etc., can solve problems such as degradation and tracking algorithm needs to be initialized

Inactive Publication Date: 2015-01-14
SHANGHAI YINGLIAN SOMATOSENSORY INTELLIGENT TECH CO LTD
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above problems brought about by supervised learning
The invention discloses a method for offline semi-automatic training of cascaded classifiers, which solves the problem of initialization and "degeneration" of tracking algorithms

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
  • Target tracking algorithm based on fusion 2D detection
  • Target tracking algorithm based on fusion 2D detection
  • Target tracking algorithm based on fusion 2D detection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] The technical scheme of the present invention is described in further detail below:

[0024] S1: Establish a statistical model of the appearance of the target object.

[0025] S1a: Mark a rectangular frame containing the target object in the first frame.

[0026] S1b: Using the sliding window, select the 10 patches closest to the marked target object as positive examples: randomly select 100 patches in the sliding window set whose overlapping area with the target object patch is less than 20% of the target object patch area, as a negative example.

[0027] S1c: Build a statistical model of the target object. The establishment of the statistical model of the target object includes two parts, the establishment of relevant parameters for the random forest classifier, and the establishment of the sample set for the nearest neighbor classifier.

[0028] The features used in the Random Forest classifier are defined as follows:

[0029] The random forest classifier we use c...

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 relates to a target tracking algorithm based on fusion 2D detection. The method comprises the steps that an appearance model of a target object at the current background is established, continuous tracking is carried out on the target object through a method based on LK sparse light stream, the position of a palm is detected in a picture through the appearance model of the target object, the result of the tracking step and the result of the detecting step are fused, and therefore a more reliable and more accurate target object position can be obtained. The cascading-classifier-based detecting method in which a semi-automatic offline training method is adopted and the tracking result and the detecting result can be fused at the same time solves the problems that according to a traditional detecting method only based on a classifier, deformation of the target object is hard to deal with and a large number of manual marks are needed in the training process, and the degeneration problem of a traditional tracking algorithm.

Description

technical field [0001] The invention belongs to the technical field related to machine vision, and more specifically relates to a real-time target tracking algorithm integrated with a 2D detection method. Background technique [0002] In the field of computer vision, the tracking of real-time moving objects is a core research direction. Widely used in video surveillance, natural human-machine interface, augmented reality, military guidance and other fields. Therefore, how to achieve stable, accurate and real-time tracking of motion is a problem that needs to be studied intensively. [0003] The traditional moving target tracking algorithm has the following problems: 1) The tracker needs to be initialized in advance, that is, the position of the target in the picture is marked, and then the tracking algorithm can start to work; 2) When the target object disappears in the picture, the tracker cannot perceive it in time ; 3) The target object disappears in the picture, and wh...

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): G06K9/62
CPCG06V20/52G06V2201/07G06F18/2415
Inventor 严嘉祺王亚捷
Owner SHANGHAI YINGLIAN SOMATOSENSORY INTELLIGENT TECH CO LTD
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