US 10,949,984 B2Granted

Object tracking system and method thereof

AssigneeNational Central University

InventorsTsung-Han Tsai, Ching-Chin Yang, Rui-Zhi Wang

Life of this patent

Estimating the term…

Loading patent…

Abstract

An object tracking system includes a foreground identifying module, an object grouping module, and an object tracking module. The foreground identifying module determines an attribute information of each pixel position of a current processing frame according to a difference between a pixel value of each pixel position of the current processing frame and that of a background model frame, so as to generate a current foreground frame. The object grouping module sets a label to each pixel position according to the attribute information of surrounding pixel positions of each pixel position, and connects adjacent pixel positions with the same label to form an object. The object tracking module determines the object being a tracked object according to whether the pixel positions corresponding to the object are at least partially overlapped with the pixel positions corresponding to the tracked object.

Claims20

Dependent claims

An object tracking system, comprising:

a foreground identifying module, determining an attribute information of each pixel position of a current processing frame of a frame series according to a difference between a pixel value of each pixel position of the current processing frame and that of a background model frame, so as to generate a current foreground frame, wherein the attribute information is of a foreground attribute or a background attribute;
an object grouping module, setting a label to each said pixel position according to the attribute information of a plurality of surrounding pixel positions of each said pixel position, and connecting a plurality of adjacent said pixel positions with the same label to form at least one object;
and
an object tracking module, determining the at least one object being a specific object according to whether the pixel positions corresponding to the at least one object in the current foreground frame are at least partially overlapped with the pixel positions corresponding to the specific object in a previous foreground frame of the frame series, and converting the label corresponding to the at least one object into the label of the specific object.

1 more element

An object tracking method, which is performed by an object tracking system, comprising the steps of:

using a foreground identifying module to determine an attribute information of each pixel position of a current processing frame of a frame series according to a difference between a pixel value of each pixel position of the current processing frame and that of a background model frame, so as to generate a current foreground frame, wherein the attribute information is of a foreground attribute or a background attribute;
using an object grouping module to set a label to each said pixel position according to the attribute information of a plurality of surrounding pixel positions of each said pixel position, and connect a plurality of adjacent said pixel positions with the same label to form at least one object;
and
using an object tracking module to determine the at least one object being a specific object according to whether the pixel positions corresponding to the at least one object in the current foreground frame are at least partially overlapped with the pixel positions corresponding to the specific object in a previous foreground frame, and convert the label corresponding to the at least one object into the label of the specific object.

1 more element

Dependencies are inferred from the claim text.

Facts

Application: 16/241,613Filed 7 Jan 2019
Priority: TW 10711867831 May 2018
Published: US 2019/0370984 A15 Dec 2019
Granted: 16 Mar 2021
Term ends: Not estimatedThe source records don’t support an estimate.

Classification17 codes

G06T 7/246
Analysis of motion: using feature-based methods, e.g. the tracking of corners or segments
G06T 5/50
Image enhancement or restoration: using two or more images, e.g. averaging or subtraction
G06T 5/70
Image enhancement or restoration: denoising; Smoothing
G06T 7/13
Segmentation; Edge detection: edge detection