Haar detection method based on GPU acceleration

A detection method and scanning window technology, which is applied in the field of Haar detection, can solve the problems such as the slow implementation speed of the Haar algorithm and the inability to meet actual needs, and achieve the effects of improving the adaptability and practical value, shortening the data processing time, and increasing the calculation density

Active Publication Date: 2015-12-16
深圳市哈工交通电子有限公司
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a Haar detection method based on GPU acceleration, which solves the problem that the implementation speed of the haar algorithm in the prior art is slow and cannot meet the actual needs

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
  • Haar detection method based on GPU acceleration

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0030] Such as figure 1 As shown, the Haar detection method based on GPU acceleration disclosed by the present invention mainly aims at the problem of slow implementation speed of the haar algorithm in the prior art, and has carried out various technical improvements. Generally speaking, it includes the following steps:

[0031] (1) System initialization, the CPU organizes the scan window, feature frame information and classifier parameter data under all magnification factors into a matrix in the GPU global memory in a one-by-one arrangement, and saves it to the GPU device;

[0032] (2) Set the number of threads in sequence according to the number of columns of the image matrix and call the kernel function twice to perform matrix transposition twice to obtain the integral map and square integral map of the image;

[0033] (3) Arrange the scan window and feature rectangles in the GPU global memory for haar algorithm detection;

[0034] (4) Combine the scanning windows of all s...

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 Haar detection method based on GPU acceleration. The method is characterized by comprising the following steps that: (1) system initialization is carried out, and a CPU transmits all scanning windows under an amplification coefficient, characteristic frame information and classifier parameter data and stores all the scanning windows under the amplification coefficient, the characteristic frame information and the classifier parameter data to GPU equipment; (2) thread numbers are set in sequence according to the column number of a image matrix, a kernel function is called twice, matrix transposition is carried out twice, and integration graphs and square integration graphs of images are are obtained; (3) the scanning windows and characteristic rectangle frames are arranged in a global memory of a GPU, and haar algorithm detection is carried out; and (4) rectangle frame merging is carried out on all the scanning windows passing strong classifiers, and a detection result is obtained. According to the invention, an existing haar algorithm is improved in three aspects of data preparation modes, rapid calculation of the integration graph, and staged and pointed Haar detection, so that the realization speed of the haar algorithm is substantially increased, the practical value is fully reflected, and the practical value and the popularization value are very high.

Description

technical field [0001] The invention relates to a Haar detection method, in particular to a Haar detection method based on GPU acceleration. Background technique [0002] The Haar method is a method for detecting objects of interest on images by using Haar features and cascaded classifiers through training. Haar detection algorithm is widely used due to its excellent performance, but its complex data structure, large amount of calculation, and slow implementation speed cannot meet the actual requirements of fast detection. Contents of the invention [0003] The purpose of the present invention is to provide a Haar detection method based on GPU acceleration, which solves the problem that the implementation speed of the Haar algorithm in the prior art is slow and cannot meet actual needs. [0004] In order to achieve the above object, the technical scheme adopted in the present invention is as follows: [0005] A Haar detection method based on GPU acceleration, comprising ...

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
CPCG06V2201/07G06F18/2411
Inventor 曹泉余坚毅
Owner 深圳市哈工交通电子有限公司
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