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

Long object identification method and identification system based on target detection and RNN

A target object and target detection technology, applied in the field of artificial intelligence, can solve the problems of long body, difficult to take complete and clear pictures of the whole vehicle, and difficult container loading methods, to meet real-time detection, without manual intervention, and easy to build. Effect

Pending Publication Date: 2020-01-21
聚时科技(上海)有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in reality, the body of the truck transporting the container is very long, and it is difficult to take a complete and clear picture of the whole vehicle. Therefore, we can only consider shooting the video of the vehicle passing at the crossing at a relatively short distance, so as to obtain a complete and clear picture. Full vehicle information
[0004] Since each frame of the video of the vehicle passing through the crossing is only an image of a part of the body, although the specific details are relatively clear, how to analyze and understand the video to obtain the loading method of the container is a difficult problem

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
  • Long object identification method and identification system based on target detection and RNN
  • Long object identification method and identification system based on target detection and RNN
  • Long object identification method and identification system based on target detection and RNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0038] like figure 1 As shown, this embodiment implements a long object recognition method based on target detection and RNN, including the following steps:

[0039] Obtain a video of the target object, which contains images of the target object moving from head to tail;

[0040] Traverse all the frames of the video, and use the deep learning-based target detection algorithm to detect the key position of the target object in each frame. The key position refers to the position that plays a decisive role in the final object recognition and classification, and saves the position of each frame in order Test results;

[0041] Generate a time series containing a set time series length of the target object according to the detection result;

[0042] Based on the time series, the RNN network is used to obtain the classification result of the target object.

[0043] During the video acquisition process of the target object, adjust the distance between the camera and the object to an...

Embodiment 2

[0060] This embodiment provides a long object recognition system based on target detection and RNN corresponding to Embodiment 1, including a video acquisition module, a target detection module, a time series generation module, and a classification module, wherein the video acquisition module is used to acquire the target The video of the object, which contains the moving image of the target object from beginning to end; the target detection module is used to traverse all frames of the video, and uses the target detection algorithm based on deep learning to detect the key position of the target object in each frame, in order Save the detection results of each frame; the time series generation module is used to generate a time series containing the set time series length of the target object according to the detection results; the classification module is used to obtain the classification results of the target object based on the time series using the RNN network .

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 long object recognition method and recognition system based on target detection and RNN, and the method comprises the following steps: obtaining a video of a target object,and the video comprises images of the target object moving from the head to the tail; traversing all frames of the video, performing key position detection of a target object on each frame by adoptinga target detection algorithm based on deep learning, and sequentially storing a detection result of each frame according to a sequence; generating a time sequence containing a set time sequence length of the target object according to the detection result; and based on the time sequence, adopting an RNN network to obtain a classification result of the target object. Compared with the prior art, the method has the advantages of being capable of accurately identifying long objects, low in cost and the like.

Description

technical field [0001] The invention belongs to the field of artificial intelligence, and mainly relates to video analysis and target detection methods, in particular to a long object recognition method and recognition system based on target detection and RNN. Background technique [0002] In recent years, with the continuous development of artificial intelligence, target detection and recognition technology has become more mature and widely used. The current mainstream object detection and recognition method is to detect and recognize a single image containing the target, but this design is not suitable for scenes with large object lengths (such as the detection of container trucks in ports). In the case of close-range shooting, it is impossible to capture a complete long target object with a single picture. In the case of long-distance shooting, a complete long target object can be captured with a single picture, but the local clarity of the obtained picture is not high e...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/049G06N3/08G06V20/40G06V2201/07G06N3/045G06F18/2415G06F18/241
Inventor 尹俊奇
Owner 聚时科技(上海)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