Unsupervised multi-object detection tracking method and storage device and camera shooting device thereof

A technology of multi-target tracking and detection method, applied in the field of unsupervised multi-target detection and tracking method and storage device and camera device, can solve the problems of difficult classification accuracy, comparison, high labeling error rate, and achieves the improvement of classifier performance. Effect

Inactive Publication Date: 2018-05-15
CHINA UNIV OF GEOSCIENCES (WUHAN)
View PDF8 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is that for the above-mentioned existing methods based on collaborative training, methods based on background modeling, methods based on production models and methods based on tracking, the samples near the classification surface (that is, difficult samples) The labeling error rate is high, and its classification accuracy rate is difficult to compare with the manual labeling offline learning classifier. It provides an unsupervised multi-target detection and tracking method and its storage device and camera device.

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
  • Unsupervised multi-object detection tracking method and storage device and camera shooting device thereof
  • Unsupervised multi-object detection tracking method and storage device and camera shooting device thereof
  • Unsupervised multi-object detection tracking method and storage device and camera shooting device thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, the specific implementation manners of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] Such as figure 1 As shown, it is a flow chart of the unsupervised multi-target detection and tracking method of the present invention. A kind of unsupervised multi-target tracking detection method in the present embodiment, comprises the following steps:

[0025] S1. Perform affine transformation on each frame of the acquired image data to construct a training sample set. The affine transformation is performed according to the artificially selected frame selection target of the first frame, and the frame selection target also includes the artificially selected frame selection target. Positive samples formed by targets to be tracked and negative samples formed by artificially selected targets not to be tracked; ...

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 an unsupervised multi-object detection tracking method and a storage device and a camera shooting device thereof. The method comprises the following steps: performing affine transformation on each frame of the acquired image data so as to build a training sample set, classifying the image data after the first image by adopting an OSF classifier according to a preset confidence threshold beta and an adaptive threshold theta, classifying OSF positive samples, OSF negative samples and OSF difficult samples, classifying the OSF difficult samples by virtue of an ISVM (Interleaved Space Vector Modulation) classifier, updating the adaptive threshold theta by virtue of new theta formed after classification of the ISVM classifier, repeating the previous steps until the adaptive threshold theta converges to a preset degree, and performing tracking detection on an object acquired in a video object by utilizing the completely trained OSF classifier and ISVM classifier. According to the method disclosed by the invention, under the condition that human intervention is not needed at all, only objects needing to be subjected to detection tracking need to be manually selected in the first frame of the video, continuous autonomic learning can be realized, the performance of the classifier is gradually improved, and finally, the multi-target detection and tracking can be realized.

Description

technical field [0001] The present invention relates to the field of target tracking, in particular to a color-based target tracking method, and more specifically, to an unsupervised multi-target detection and tracking method and its storage device and camera device. Background technique [0002] Video object detection and tracking is a research hotspot in the field of computer vision, and it has important theoretical research significance and practical value in applications such as video surveillance, virtual reality, human-computer interaction, and autonomous navigation. At present, most video target detection and tracking systems include three modules: target positioning, target data association tracking and trajectory generation. Among them, the data association tracking algorithm has made a great breakthrough with the in-depth research of the detection-based tracking method. However, the object localization module still adopts a large number of off-line learning algori...

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): G06K9/62G06K9/00
CPCG06V20/48G06F18/241G06F18/214
Inventor 罗大鹏杜国庆曾志鹏牟泉政魏龙生高常鑫马丽王勇
Owner CHINA UNIV OF GEOSCIENCES (WUHAN)
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