Search

CN-121985109-A - Analysis method, analysis equipment and storage medium for multipath video

CN121985109ACN 121985109 ACN121985109 ACN 121985109ACN-121985109-A

Abstract

The embodiment of the application relates to the technical field of video analysis and discloses a method for analyzing multipath videos, which comprises the steps of obtaining first coded streams of a plurality of original video sequences; the method comprises the steps of obtaining a plurality of corresponding time stamp sequences based on a plurality of first coded streams, generating a video frame index matrix according to the plurality of time stamp sequences, and analyzing the video frame index matrix to generate an analysis report. The method obtains the time stamp sequence with higher precision based on the first code in the original video sequence, generates the video frame index matrix based on the obtained time stamp sequences, automatically generates the analysis report according to the generated video frame index matrix, and realizes the automation of multi-path video analysis.

Inventors

  • Request for anonymity
  • Request for anonymity
  • Request for anonymity

Assignees

  • 帕西尼感知科技(深圳)有限公司

Dates

Publication Date
20260505
Application Date
20251224

Claims (12)

  1. 1. A method for analyzing multiple video streams, comprising: Acquiring a first coded stream of a plurality of original video sequences; acquiring a plurality of corresponding time stamp sequences based on a plurality of the first coded streams; Generating a video frame index matrix according to the plurality of time stamp sequences; And analyzing the video frame index matrix to generate an analysis report.
  2. 2. The method of claim 1, wherein generating a video frame index matrix from the plurality of time stamp sequences comprises: determining one time stamp sequence from a plurality of time stamp sequences as a first time stamp sequence, and taking the rest time stamp sequences as a second time stamp sequence; And respectively carrying out time sequence alignment on each second time stamp sequence and the first time stamp sequence to generate the video frame index matrix.
  3. 3. The method of claim 2, wherein said time-sequentially aligning each of said second time-stamp sequences with said first time-stamp sequence, respectively, to generate said video frame index matrix, comprises: Determining a second target time stamp aligned with the first time stamp in the second time stamp sequence based on the first time stamp sequence and a preset rule; taking the video frame index corresponding to the second target timestamp as a matrix unit corresponding to the first timestamp to obtain an initial index matrix; And filling the empty matrix unit in the initial index matrix by adopting the lost frame identification to obtain the video frame index matrix.
  4. 4. The method of claim 3, wherein the predetermined rule comprises taking as the second target timestamp a timestamp in the second sequence of timestamps having an absolute value of a difference from the first timestamp of less than a predetermined threshold.
  5. 5. The method of claim 3, wherein parsing the video frame index matrix generates an analysis report, comprising: and counting the lost frame identifications in the video frame index matrix to generate the analysis report.
  6. 6. The method of claim 1, wherein prior to said obtaining a first encoded stream of a plurality of original video sequences, the method further comprises: performing function detection on a plurality of video acquisition devices; and if the function detection results of the video acquisition devices are normal, acquiring a plurality of original video sequences acquired by the video acquisition devices.
  7. 7. The method of claim 6, wherein the performing function detection on the plurality of video capture devices comprises: the network connection functions of a plurality of the video acquisition devices are detected by an inspection command and/or a detection tool.
  8. 8. The method of claim 6, wherein the performing function detection on the plurality of video capture devices comprises: And extracting key frames from the plurality of original video sequences according to a preset frequency, and detecting video acquisition functions of a plurality of video acquisition devices according to a preset function model and the key frames.
  9. 9. The method according to any one of claims 1-8, wherein said obtaining a corresponding plurality of time stamp sequences based on a plurality of said first encoded streams comprises: Decomposing the first encoded stream into a plurality of first units and taking the first units containing first data information as second units; And analyzing the second unit according to a preset protocol to obtain the time stamp sequence corresponding to the first coded stream.
  10. 10. A video analysis apparatus, comprising: At least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.
  11. 11. A video analysis system, comprising: at least two video acquisition devices; a network device, and a video analysis device as claimed in claim 10.
  12. 12. A computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, computer-executable instructions for causing a computer device to perform the method of any one of claims 1-9.

Description

Analysis method, analysis equipment and storage medium for multipath video Technical Field The embodiment of the application relates to the technical field of video analysis, in particular to a method, equipment and storage medium for analyzing multi-path videos. Background With the rapid development of microelectronic technology and computer vision, multi-camera (camera) array systems are widely used in the leading edge fields of automatic driving, three-dimensional reconstruction, robot navigation, industrial quality inspection, virtual reality and the like. In these applications, the system needs to collect video data synchronously from multiple spatial perspectives to obtain complete environmental information, and therefore, there is a high requirement for the integrity of each video data stream and the precise synchronization of the time stamps between the video data streams. At present, the prior art has the problems of single fault detection dimension, low time synchronization precision and the like in the aspect of ensuring the reliability and synchronization precision of data acquisition of multiple cameras, and meanwhile, the conventional fault detection also has the problems of difficult positioning of synchronization abnormality, lack of automatic quantification and the like. Content of the application The embodiment of the application mainly provides a multi-channel video analysis method, which adopts more accurate time stamps to automatically generate an analysis report, thereby improving the analysis accuracy and realizing automation. In order to solve the technical problems, the embodiment of the application adopts the following technical scheme: in a first aspect, an embodiment of the present application provides a method for analyzing multiple video, including: Acquiring a first coded stream of a plurality of original video sequences; acquiring a plurality of corresponding time stamp sequences based on the plurality of first coded streams; Generating a video frame index matrix according to the plurality of time stamp sequences; and analyzing the video frame index matrix to generate an analysis report. In some embodiments, generating a video frame index matrix from a plurality of time stamp sequences includes: Determining one time stamp sequence from the plurality of time stamp sequences as a first time stamp sequence, and taking the rest time stamp sequences as a second time stamp sequence; and respectively aligning each second time stamp sequence with the first time stamp sequence in time sequence to generate a video frame index matrix. In some embodiments, time-sequentially aligning each second time-stamp sequence with the first time-stamp sequence, respectively, generates a video frame index matrix, comprising: Determining a second target timestamp aligned with the first timestamp in a second timestamp sequence based on the first timestamp sequence and a preset rule; Taking the video frame index corresponding to the second target timestamp as a matrix unit corresponding to the first timestamp to obtain an initial index matrix; and filling the empty matrix unit in the initial index matrix by adopting the lost frame identification to obtain a video frame index matrix. In some embodiments, the preset rule includes taking as the second target timestamp a timestamp in the second sequence of timestamps having an absolute value of a difference from the first timestamp that is less than a preset threshold. In some embodiments, parsing the video frame index matrix generates an analysis report, including: and counting the lost frame identification in the video frame index matrix to generate an analysis report. In some embodiments, prior to obtaining the first encoded stream of the plurality of original video sequences, the method further comprises: performing function detection on a plurality of video acquisition devices; And if the function detection results of the video acquisition devices are normal, acquiring a plurality of original video sequences acquired by the video acquisition devices. In some embodiments, functional detection of a plurality of video capture devices includes: the network connection function of the plurality of video capture devices is detected by an inspection command and/or a detection tool. In some embodiments, functional detection of a plurality of video capture devices includes: And extracting key frames from the plurality of original video sequences according to the preset frequency, and detecting video acquisition functions of the plurality of video acquisition devices according to the preset function model and the key frames. In some embodiments, obtaining a corresponding plurality of time stamp sequences based on the plurality of first encoded streams includes: Decomposing the first coded stream into a plurality of first units and taking the first units containing first data information as second units; And analyzing the second unit according to a preset protocol