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

Labeling method and device for road information in video

A technology in road information and video, applied in the field of road information labeling in video, which can solve the problems of low video labeling efficiency and labor-intensive

Inactive Publication Date: 2020-10-30
MOMENTA SUZHOU TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the existing video labeling process requires manual labeling of each frame of video data, which is extremely labor-intensive, leading to the technical problem of low efficiency of the existing video labeling

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
  • Labeling method and device for road information in video
  • Labeling method and device for road information in video
  • Labeling method and device for road information in video

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0071] see figure 1 , figure 1 It is a schematic flowchart of a method for labeling road information in a video provided by an embodiment of the present invention. This method is applied in automatic driving, and is typically applied in the application scenario of labeling road information that does not change much between consecutive video frames. The method can be performed by a device for labeling road information in the video, which can be realized by means of software and / or hardware, and can generally be integrated in vehicle-mounted terminals such as a vehicle-mounted computer and a vehicle-mounted industrial personal computer (Industrialpersonal Computer, IPC). Examples are not limited. Such as figure 1 As shown, the method provided in this embodiment specifically includes:

[0072] 110. Extract current video frame data in the video stream.

[0073] The object to be marked in this embodiment is the road information in the video stream, and the road information may...

Embodiment 2

[0096] see figure 2 , figure 2 It is a schematic structural diagram of a device for labeling road information in videos provided by an embodiment of the present invention. Such as figure 2 As shown, the device includes: a video frame data extraction module 210 , an initial labeling result determination module 220 and a first labeling module 230 . in,

[0097] The video frame data extraction module 210 is configured to extract the current video frame data in the video stream;

[0098] The initial labeling result determination module 220 is configured to label the road information in the current video frame data based on a preset road information detection model to obtain an initial labeling result of the current video frame data, wherein the preset The road information detection model makes each road information in each frame of video data associated with its corresponding position;

[0099] The first labeling module 230 is configured to, if the reference labeling resul...

Embodiment 3

[0121] see image 3 , image 3 It is a schematic structural diagram of a vehicle-mounted terminal provided by an embodiment of the present invention. Such as image 3 As shown, the vehicle terminal may include:

[0122] A memory 701 storing executable program codes;

[0123] a processor 702 coupled to the memory 701;

[0124] Wherein, the processor 702 invokes the executable program code stored in the memory 701 to execute the method for marking road information in the video provided by any embodiment of the present invention.

[0125] The embodiment of the present invention also provides another vehicle-mounted terminal, including a memory storing executable program codes; a processor coupled to the memory; wherein, the processor invokes the executable program codes stored in the memory to execute any embodiment of the present invention Annotation method for road information in the provided video.

[0126] The embodiment of the present invention discloses a computer-rea...

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 embodiment of the invention discloses a labeling method and device for road information in a video. The method comprises the steps: extracting the current video frame data in a video stream; marking the road information in the current video frame data based on a preset road information detection model to obtain an initial marking result, the preset road information detection model enabling theroad information in each frame of video data to be associated with the position of the road information in the video data; and if a reference labeling result obtained after the initial labeling result is corrected is obtained, labeling the road information of the extracted next frame of video data by utilizing the reference labeling result of the current video frame of data based on the preset road information detection model. By adopting the technical scheme, the manual time for road information video annotation is greatly shortened, and the cost of road information annotation in the video stream is effectively reduced.

Description

technical field [0001] The invention relates to the technical field of automatic driving, in particular to a method and device for marking road information in videos. Background technique [0002] Human-annotated data is the fuel for deep learning. In order to obtain a more accurate and robust deep learning model, it is often necessary to label a large amount of expensive data. Video data annotation, since it can be understood as a very dense collection of pictures, requires even greater financial resources. [0003] Taking a 5-minute driving recorder as an example, the general video is roughly 25fps (frames per second), that is, a total of 5*60*25=7500 pictures need to be marked. On the one hand, the road conditions that can be covered by a 5-minute video are far less abundant than 7,500 randomly selected pictures. On the other hand, the multi-frame information contained in continuous video is what we need. Therefore, in the field of data labeling, the labeling of contin...

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): G06K9/00G06K9/62
CPCG06V20/588G06V20/56G06F18/214
Inventor 陈佳辉张驰蒋竺希
Owner MOMENTA SUZHOU TECH CO LTD