Unlock instant, AI-driven research and patent intelligence for your innovation.

Front-end real-time target tracking method based on Jetson NX platform

A target tracking and platform technology, applied in the field of target tracking, can solve the problems of JetsonNX's slow processing speed and lack of real-time requirements, and achieve the effect of improving real-time performance and reducing power consumption

Pending Publication Date: 2021-11-16
西安中科西光航天科技有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

During the tracking process, the processing speed of Jetson NX is slow and cannot meet the real-time requirements

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
  • Front-end real-time target tracking method based on Jetson NX platform
  • Front-end real-time target tracking method based on Jetson NX platform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The present invention will be described in further detail below in conjunction with the accompanying drawings.

[0022] combined with Figure 1~2 , the present application discloses a front-end real-time target tracking method based on the Jetson NX platform, wherein the real-time tracking system adopts a camera terminal on the premise that there is only one GPU and is completely offline, and the camera terminal is embedded in the Jetson NX platform.

[0023] Transplant a specific network model (tracking model) to the C compiler of the Jetson NX platform development board on the camera side to accelerate tracking reasoning.

[0024] Such as figure 2 Shown, the step of a kind of front-end real-time target tracking method based on Jetson NX platform of the present invention is as follows:

[0025] 1) Save the end-to-end offline model that has been trained on the host or server. Specifically, the saved tracking algorithm should be transplanted to the Jetson NX platform,...

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 front-end real-time target tracking method based on a JetsonNX platform. The method comprises the following steps: 1) storing a trained end-to-end offline model on a host or a server; 2) converting a pth file of a tracking model into a pt file special for C compiling on Jetson NX; (3) compiling a pt file on the Jetson NX by using the C; (4) reading a real-time camera data flow on the Jetson NX; 5) specifying a target to be tracked in a first frame, and then running the model by a C program; 6) continuing to jump to the fourth step to continue execution; and 7) manually exiting and / or automatically terminating the tracking process. The method has the advantages that front-end (or local video) real-time tracking is carried out on camera under the condition that only one GPU is provided and the video memory is low, a mainstream tracking model is deployed on mobile terminal equipment, the real-time performance of the tracking model based on deep learning applied to edge equipment (embedded type) is greatly improved, and power consumption is remarkably reduced.

Description

technical field [0001] The invention relates to the field of target tracking, in particular to a front-end real-time target tracking method based on the Jetson NX platform. Background technique [0002] Since our mainstream deep learning models are directly stored on the host or server, the calculation is often performed on the host or server when testing the speed. When the deep learning model is deployed on the edge device (embedded), because the CPU and GPU resources of our mobile edge device (embedded) are limited, and the computing resource consumption is very large, the preprocessing time before tracking and tracking The processing time in the process will be relatively long, and the real-time performance cannot reach the speed of the host or server test. Therefore, it is not advisable to directly deploy the deep learning model to the mobile edge device (embedded). [0003] In the prior art, the NVIDIA Jetson NX kit is configured with 8GB128-bit LPDDR4 and 64GB eMMC ...

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
IPC IPC(8): G06T7/20G06F8/41G06F9/448
CPCG06T7/20G06F8/41G06F9/4482G06T2207/10016
Inventor 秦静祝青
Owner 西安中科西光航天科技有限公司