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

Real-time intelligent target detection method and system based on embedded platform

A target detection and embedded technology, applied in the field of target detection, can solve the problems of low robustness, scale change, background interference, etc., and achieve the effect of speeding up the recognition rate, reducing the size and improving the accuracy

Pending Publication Date: 2021-12-07
EAST CHINA INST OF COMPUTING TECH
View PDF1 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when applied to actual scenes, there are still many problems to be solved, such as occlusion, scale change, background interference, etc.
Although some algorithms can achieve fast tracking of the target to a certain extent, such as CSK, MOSSE, TLD, etc., when the target is affected by illumination, deformation, occlusion, etc., the tracking process is greatly disturbed, resulting in low accuracy of the tracking algorithm, and The robustness is extremely low, problems such as tracker drift and tracking target loss may occur, and accurate target tracking cannot be completed
With the further research and exploration of detection and tracking algorithms, some algorithms with better precision and stronger robustness have appeared, but these algorithms cannot locate the target object in real time, and have high requirements for hardware resources

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
  • Real-time intelligent target detection method and system based on embedded platform
  • Real-time intelligent target detection method and system based on embedded platform
  • Real-time intelligent target detection method and system based on embedded platform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0108] A kind of real-time intelligent target detection method based on embedded platform provided by the present invention, such as figure 1 shown, including:

[0109] Step S1: Obtain real-time video data, and use the improved three-frame frame difference method to perform dynamic target detection and extract key areas;

[0110] Step S2: selecting key frames and non-key frames by means of frame extraction;

[0111] Step S3: Based on the extracted key area, use the improved target recognition model after training to identify the target position and category through key frames, and use the trained target tracking model to perform target positioning and detection using the multi-model target tracking strategy for non-key frames, and repeatedly trigger Step S1 to step S2, until the video data cannot be obtained;

[0112] The improved three-frame frame difference method is to reprocess the difference results of adjacent frames, screen out possible moving target sets to be detect...

Embodiment 2

[0197] Embodiment 2 is a preferred example of embodiment 1

[0198] The research and application of embedded system target detection and tracking algorithms is of great significance. With the overlapping and progress of the times, the application environment of related algorithms has become more and more complex, and the challenges encountered have also increased. In order to ensure that the task can be completed quickly and accurately, the real-time and accuracy of the tracking algorithm and the target recognition algorithm must be considered.

[0199] Difficulties in target detection:

[0200] (1) The size of the target is different: the size of the objects that need to be detected in the actual scene is different, especially the detection of small targets has always been one of the difficult problems in the target detection technology.

[0201] (2) The variability of the target: In the application of computer vision, the shape of the target is varied, and there will also b...

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 real-time intelligent target detection method and system based on an embedded platform, and the method comprises the steps: S1, obtaining real-time video data, and carrying out the dynamic target detection through an improved three-frame difference method, and extracting a key region; S2, selecting a key frame and a non-key frame in a frame extraction mode; and S3, based on the extracted key area, identifying the position and category of a target by using a trained improved target identification model through a key frame, carrying out target positioning and detection on a non-key frame by using a trained target tracking model, and repeatedly triggering the S1 to S2 until video data are not obtained. The invention provides a real-time tracking algorithm which combines an improved model and target tracking and fuses a moving target detection technology. According to the improved frame difference method, an original three-frame frame difference method is improved, and the accuracy of boundary detection can be improved.

Description

technical field [0001] The present invention relates to the technical field of target detection, in particular to a real-time intelligent target detection method and system based on an embedded platform. Background technique [0002] In recent years, artificial intelligence technology has developed rapidly. At the same time, machine vision technology, as an important branch of artificial intelligence, is also developing rapidly. With the help of vision, humans can obtain important information to recognize the world. Compared with the limitations of human vision, machines can help humans obtain rich information and greatly broaden human vision. With the help of advanced machine tools, human beings can look up to the bright moon and down to the nine oceans, ranging from the vast universe to a tiny example. As an important part of the field of computer vision, target detection and tracking technology is a difficult and challenging task. The main task of target detection is to...

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): G06T7/246G06K9/00G06K9/32G06K9/62G06N3/04G06N3/08
CPCG06T7/251G06N3/082G06N3/045G06F18/253Y02T10/40
Inventor 林敏郭威张浩博
Owner EAST CHINA INST OF COMPUTING TECH