Struck target tracking method using GPU hardware acceleration

A target tracking and hardware acceleration technology, applied in the computer field, can solve problems such as limiting practical applications, slow processing speed, and lack of real-time performance, and achieve the effects of increasing speed, increasing speed, and overcoming the reduction in robustness

Active Publication Date: 2017-07-28
XIDIAN UNIV
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although this method has the advantages of high performance, to a certain extent, it overcomes the problem that traditional methods cannot solve the problem of reduced robustness caused by occlusion and illumination changes.
However, the shortcomings of this method are that the method uses serial calculations to extract the features of the training samples and test samples, obtain and update the weights and gradients of the support samples, and calculate the discriminant value of the test samples. The process involves a lot of calculations, so this implementation has the disadvantage of slow processing speed
Experiments show that for a long video sequence, the average processing speed of this method is about 5fps, obviously this method is not real-time, thus limiting its practical application

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
  • Struck target tracking method using GPU hardware acceleration
  • Struck target tracking method using GPU hardware acceleration
  • Struck target tracking method using GPU hardware acceleration

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] The present invention will be further described below in conjunction with the drawings.

[0043] The invention adopts the CUDA language and can be implemented on any GPU device of NVIDIA that supports the CUDA architecture. Before implementing the method of the present invention, the cudaMalloc function should be called to allocate eight memory areas on the GPU device. After using the method of the present invention, the cudaFree function should also be called to release these eight memory areas.

[0044] Reference figure 1 , The present invention can be realized through the following steps:

[0045] Step 1. Obtain a grayscale image.

[0046] Call the image loading function cvLoadImage in the open source computer vision library OpenCV to load the first frame image in the image sequence to be tracked into the memory of the computer host.

[0047] Call the color channel conversion function cvCvtColor in the open source computer vision library OpenCV to convert the image loaded i...

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 Struck target tracking method using GPU hardware acceleration, which solves the problems of low tracking performance and no real-time processing capability in the prior art. The steps that the present invention realizes: (1) obtain gray-scale image; (2) judge whether the image loaded is the first frame image; (3) initialize the position rectangle frame of tracking target; (4) extract all test samples of gray-scale image (5) Determine the position rectangle of the tracking target; (6) Extract the features of all training samples of the grayscale image; (7) Initialize the weights and gradients of the training samples; (8) Obtain and update the weights of the support samples value, gradient; (9) judge whether all images are loaded; (10) end target tracking. The invention can be used on a general computer to realize real-time tracking of the target in the video.

Description

Technical field [0001] The present invention belongs to the field of computer technology, and further relates to a structure output Struck target tracking method based on a kernel method using computer graphics processor GPU hardware acceleration in the technical field of computer video target tracking. The invention can realize the acceleration of the Struck target tracking method based on the structure output of the kernel method, and can be used on a general-purpose computer to realize real-time tracking of the target in the video. Background technique [0002] A high-speed, high-performance target tracking method is the core technology in the field of computer vision. The current target tracking methods are divided into two categories: one is the tracking method based on feature matching, which is mainly to construct features that can represent the target, and then determine the location of the target through the matching degree between the features; the other is based on the...

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 & AuthorityPatents(China)
IPC IPC(8): G06T7/223
Inventor李云松尹万春宋长贺
OwnerXIDIAN UNIV