Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method, device and equipment for optimizing intelligent video analysis performance

An intelligent video analysis and performance technology, applied in the direction of digital video signal modification, TV, electrical components, etc., can solve the problems of difficult number of concurrent channels, low offline video analysis speed, low decoding performance, etc., to reduce performance delay loss Effect

Active Publication Date: 2019-05-17
武汉众智数字技术有限公司
View PDF6 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, when the intelligent analysis algorithm is faced with massive video processing scenarios, it faces huge performance pressure. Taking the most widely used 1080PH264 video stream as an example, the current mainstream intelXeon server based on x86 architecture usually only reaches about 200 based on CPU decoding. The performance of ~300fps, and the intelligent video analysis algorithm is usually a pipeline of video stream->decoding->YUV / RGB data->algorithm processing. After adding the algorithm link, since the image algorithm usually consumes a lot of CPU, the above decoding performance will be reduced. Lower, the specific performance is that for the two main application scenarios of video: offline video and real-time video streaming, the analysis speed of offline video will not be high, and the number of concurrent channels that can be supported by real-time video streaming is also difficult to increase. Through horizontal expansion analysis Nodes, if adding analysis servers to improve performance, the cost is too high and the cost performance is too low, it is difficult to support the application scenarios of large-scale video analysis

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
  • Method, device and equipment for optimizing intelligent video analysis performance
  • Method, device and equipment for optimizing intelligent video analysis performance
  • Method, device and equipment for optimizing intelligent video analysis performance

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0060] see figure 1 , the present embodiment provides a device for optimizing the performance of intelligent video analysis, including a GPU scheduling module, a decoding module and an algorithm analysis module, the GPU scheduling module is used to be responsible for the performance benchmark test of the GPU card and the allocation of the optimal number of parallel tasks and the The multi-card scheduling of the video analysis task, the multi-card scheduling of the video analysis task adopts the offline video analysis task resource scheduling step of Embodiment 2 or the real-time video stream analysis task resource scheduling step of Embodiment 3; Card scheduling, according to different video analysis tasks, is divided into two scheduling models: offline video and online video.

[0061] The GPU scheduling module can segment, schedule and manage GPU analysis tasks; accelerate optimization for offline video scenarios: call the GPU to perform hard decoding for each subtask, and di...

Embodiment 2

[0066] see figure 2 and image 3 , this embodiment provides a method for optimizing the performance of intelligent video analysis, using the following GPU resource scheduling processing steps for offline video:

[0067] (1) Detect and manage various GPU models, and automatically identify the card type and number;

[0068] (2) Use mainstream H264 or H2651080P video files as the benchmark source;

[0069] (3) Write a benchmark test analysis program, which can realize the decoding + algorithm analysis function of multi-channel video files, and can output the analysis frame rate fps of each channel;

[0070] (4) Slice the benchmark test file according to M=1, 2, 3, 4... etc. (or load M benchmark files at the same time) use the test analysis program in step 3 to perform M-channel video analysis, and record the fps value of each channel ;

[0071] (5) When the maximum fps*M appears for the first time, record the M value at this time as the optimal number of single-card GPU task...

Embodiment 3

[0082] see Figure 4 to Figure 6 , this embodiment provides a method for optimizing the performance of intelligent video analysis, using the following real-time video stream analysis task resource scheduling steps:

[0083] (1) Detect and manage various GPU models, and automatically identify the card type and number;

[0084] (2) For the specified GPU card type, use the mainstream H264 or H2651080P real-time video stream as the benchmark source;

[0085] (3) Write a benchmark test analysis program, which can realize the decoding + algorithm analysis function of multi-channel video files, and can output the analysis frame rate fps of each channel;

[0086] (4) For a single card connected to M channels of real-time streams, and at the same time print out the algorithm link to analyze the frame rate fps, start to increase M from M=1,2,3.., when the fps decreases to approach the Q value such as Q=25 (fps>=25, 25 is the most common real-time video stream frame rate in the field o...

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 relates to a method, a device and equipment for optimizing the analysis performance of an intelligent video, and the method comprises the steps: (1) carrying out a reference piperine test on a video file for the acceleration of an offline video file, and setting an optimal file slice number; slicing the video file, and issuing a slicing task to the GPU; calling a GPU to decode the slice file, and calling back a decoding result to an algorithm directly through a video memory address, and reducing the performance loss without the video memory-main memory copy, wherein the video analysis algorithm takes the decoded video memory address, calls a GPU for algorithm acceleration and outputs an analysis result; (2) optimizing and expanding the number of paths for real-time video stream algorithm analysis; and calling the GPU to decode each path of real-time video, calling back a decoding result to the algorithm directly through a video memory address, setting double caches by analgorithm end, storing decoded data in multiple paths, transmitting the decoded data to the algorithm for GPU batch processing, and switching the two cache functions after batch processing is completed to achieve the purpose of minimum system delay.

Description

technical field [0001] The invention relates to the technical field of video image processing, in particular to a method, device and equipment for optimizing the performance of intelligent video analysis. Background technique [0002] With the gradual advancement and implementation of large-scale security projects and projects such as "Safe City", "Smart City", and "Snow Bright Project", the construction of urban video surveillance has gradually entered an in-depth stage. While accumulating a large amount of video data, it has already Unsatisfied with the simple "watching" video stage: In the face of massive video scenes, the traditional manual inspection of video with the naked eye consumes a lot of manpower and material resources, but it often seems powerless and cannot adapt to the real case handling needs of the public security industry. Under this background, through intelligent video analysis algorithms, such as line detection, target tracking, face detection, etc., th...

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): H04N17/00H04N19/436H04N19/44H04N19/423
Inventor 谈鸿韬陆辉刘树惠杨波
Owner 武汉众智数字技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com