Target tracking state model training method and device

A state model, target state technology, applied in the field of target tracking, can solve problems such as low processing accuracy

Inactive Publication Date: 2017-06-20
BOCOM SMART INFORMATION TECH CO LTD
View PDF3 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to overcome the defect that the existing target tracking state is processed according to manual rules with low accuracy

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 state model training method and device
  • Target tracking state model training method and device
  • Target tracking state model training method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0056] This embodiment provides a method for training a target tracking state model, the flow chart of the method is as follows figure 1 shown. Including the following steps:

[0057] S1: Obtain the known trajectory information of the target in different states in the video sequence and the tracking state model of the target. The known trajectory information may include the trajectory position of the target and characteristics of the target. The tracking state model of the target includes a similarity equation and a reward function.

[0058]Specifically, the different states may include a hypothetical state, a normal state, a lost state and a deleted state. The hypothetical state means that when the target appears for the first time, it is taken as the target to be tracked. Preferably, in order to prevent interference caused by false detections caused by target extraction or detection, a target that only appears in one frame is not immediately used as a target to be tracke...

Embodiment 2

[0097] This embodiment provides a target tracking state model training device, the schematic diagram of the device is as follows figure 2 shown. include:

[0098]The acquiring unit 10 is configured to acquire the known track information of the target in different states in the video sequence and the tracking state model of the target. The known trajectory information may include the trajectory position of the target and characteristics of the target. The tracking state model of the target includes a similarity equation and a reward function.

[0099] The output unit 20 is configured to use the tracking state model to track the target in the video sequence, and obtain trajectory information output by the tracking state model in different states.

[0100] The selection unit 30 is configured to select the trajectory information in one state from the trajectory information output by the tracking state model in different states. Specifically, when the trajectory information ou...

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 provides a target tracking state model training method and device. The method comprises the steps that the known trajectory information of a target in a video sequence in different states and the tracking state model of the target are acquired; the tracking state model is used to track the target in the video sequence, so as to acquire trajectory information outputted by the tracking state model in different states; trajectory information in a state is selected from trajectory information outputted by the tracking state model in different states; according to the trajectory information in the state and the known trajectory information of the corresponding state, whether the tracking state of the target is correct is determined; and if the tracking state of the target is incorrect, the tracking state model is updated. According to the invention, the tracking state model is trained by constantly updating the tracking state model through the known trajectory information of the target in different states, so as to realize automatic conversion of the tracking state of the target in different cases.

Description

technical field [0001] The invention relates to the technical field of target tracking, in particular to a method and device for training a target tracking state model. Background technique [0002] Target tracking technology has a wide range of applications in the field of intelligent security, vehicle auxiliary systems or military fields. Target tracking technology usually detects the target first, and then tracks the detected target. For example, in the process of multi-target tracking, multiple targets need to be managed. Since each target has its own life cycle, that is, from the appearance of the target to the disappearance of the target, what is the normal tracking state and lost state of each target in the life cycle? For correct conversion, the target is occluded or not detected in a certain frame, and how the new target detected in each frame is related to the existing tracked target. In the process of target tracking, when targets cross or stick together, it is ...

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): G06K9/00G06K9/62
CPCG06V20/48G06F18/22G06F18/214
Inventor 谯帅彭莉张如高
Owner BOCOM SMART INFORMATION 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