CN118839031A - Multi-level file index management method for video data - Google Patents
Multi-level file index management method for video data Download PDFInfo
- Publication number
- CN118839031A CN118839031A CN202411045767.6A CN202411045767A CN118839031A CN 118839031 A CN118839031 A CN 118839031A CN 202411045767 A CN202411045767 A CN 202411045767A CN 118839031 A CN118839031 A CN 118839031A
- Authority
- CN
- China
- Prior art keywords
- video
- video data
- index
- information
- file
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000007726 management method Methods 0.000 title claims abstract description 20
- 238000013507 mapping Methods 0.000 claims abstract description 17
- 239000013598 vector Substances 0.000 claims description 110
- 238000000034 method Methods 0.000 claims description 37
- 230000002159 abnormal effect Effects 0.000 claims description 18
- 238000010276 construction Methods 0.000 claims description 17
- 230000006399 behavior Effects 0.000 claims description 9
- 238000013500 data storage Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 5
- 230000006835 compression Effects 0.000 claims description 4
- 238000007906 compression Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims description 3
- 230000008676 import Effects 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a multi-level file index management method for video data, which relates to the technical field of file index management, and comprises the steps of acquiring video data corresponding to a plurality of video sources, sequentially carrying out frame cutting on all video data, dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source, carrying out characteristic processing on the video frames corresponding to each video source, further acquiring characteristic frame information corresponding to each video data, constructing a multi-level index file, mapping the video data corresponding to all video sources into the multi-level index file, further constructing a hierarchical search index corresponding to each video data, configuring a query platform, guiding the hierarchical search index of all video data into a terminal database corresponding to the query platform, inputting search information of the video data to be searched into the query platform, carrying out index matching according to the search information and the hierarchical search index, and further positioning target video data.
Description
Technical Field
The invention relates to the technical field of file index management, in particular to a multi-level file index management method for video data.
Background
The multi-level file index management refers to a method for organizing and managing a file index structure according to different levels, in which the index is not only a simple one-level index, i.e. a basic file name or attribute index, but is constructed and managed in a hierarchical manner according to the characteristics and attributes of the file.
With the massive generation and application of digital video data, how to efficiently manage and retrieve video files becomes one of the key contents of research of file index management technology, and the conventional file system and database management method generally have the problems of low retrieval efficiency, low retrieval precision and the like when facing large-scale video data, so that a management method combining multi-level file index management technology is needed to achieve more efficient and accurate retrieval of video data.
Disclosure of Invention
In order to solve the above problems, an object of the present invention is to provide a multi-level file index management method for video data.
The aim of the invention can be achieved by the following technical scheme: a multi-level file index management method for video data, comprising the steps of:
Step S1: acquiring video data corresponding to a plurality of video sources, sequentially carrying out frame cutting on all the video data, dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source, carrying out feature processing on the video frames corresponding to the video data of each video source, and further acquiring feature frame information corresponding to each video data;
Step S2: constructing a multi-level index file, mapping video data corresponding to all video sources into the multi-level index file, and further constructing a hierarchical retrieval index corresponding to each video data;
Step S3: configuring a query platform, guiding hierarchical search indexes of all video data into a terminal database correspondingly arranged in the query platform, inputting search information of the video data to be searched into the query platform, and performing index matching according to the search information and the hierarchical search indexes so as to locate target video data.
Further, the process of obtaining video data corresponding to a plurality of video sources, sequentially performing frame cutting on all video data, and dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source includes:
Numbering a plurality of video sources, wherein i=1, 2,3, … …, n and n are natural numbers larger than 0, the video source with the number i comprises a plurality of video data, all the video data comprising the video source with the number i are constructed into a video set, and are denoted as omega { i }, all the video data comprising the omega { i } are numbered, each video data is denoted as VD [ i ] [1], VD [ i ] [2], VD [ i ] [3], … …, VD [ i ] [ m ], and m is a natural number larger than 0, all the video data corresponding to all the video sources are obtained, and from the video source with i=1, all the video data comprising the video source are sequentially subjected to frame cutting, VD [1] [2], … …, VD [1] [ m ], and the operation is repeated until the frame cutting of all the video data corresponding to the video source with i=n is completed;
Dividing video data included in each video source into a plurality of corresponding video frames through frame cutting, and merging the number of the video source where the video frame is located, the number of the video data where the video frame is located and the corresponding time stamp on the video data where the video frame is located as a frame identifier of the current video frame.
Further, the process of performing feature processing on a plurality of video frames corresponding to video data of each video source to obtain feature frame information corresponding to each video data includes:
the feature processing of all video frames comprises feature extraction and feature vector construction;
Acquiring color features, edge features and texture features of each video frame through feature extraction;
The feature vector is constructed as: constructing a vector construction model for vectorization processing by a convolutional neural network technology, and taking video frames with the highest color feature ratio, the highest edge feature ratio and the highest texture feature ratio in video frames included in each video data as a first vector parameter, a second vector parameter and a third vector parameter corresponding to the vector construction model respectively;
Inputting the first vector parameter, the second vector parameter and the third vector parameter into a vector construction model, further constructing a conventional feature vector corresponding to color features, a conventional feature vector corresponding to edge features and a conventional feature vector corresponding to texture features, and setting vector distinguishing angle ranges of the respective conventional feature vectors, wherein the vector distinguishing angle ranges comprise a first positive interval range, a second positive interval range, a first negative interval range and a second negative interval range;
Inputting color features, edge features and texture features corresponding to all video frames into a vector construction model, further obtaining color features, edge features and texture feature vectors corresponding to each video frame, obtaining the degrees of included angles between the feature vectors of the color features corresponding to each video frame and corresponding conventional feature vectors, the degrees of included angles between the feature vectors of the edge features corresponding to each video frame and corresponding conventional feature vectors, and the degrees of included angles between the feature vectors of the texture features corresponding to each video frame and corresponding conventional feature vectors;
When the included angle degree is in a first positive interval range, the corresponding feature vector is marked as a class A feature vector, when the included angle degree is in a second positive interval range, the corresponding feature vector is marked as a class B feature vector, when the included angle degree is in a first negative interval range, the corresponding feature vector is marked as a class C feature vector, when the included angle degree is in a second negative interval range, the corresponding feature vector is marked as a class D feature vector, and all marked feature vectors of video frames corresponding to each video data are used as feature frame information of the corresponding video data.
Further, the process of constructing the multi-level index file includes:
Acquiring source address information corresponding to each video source, and abstracting video information corresponding to all video data contained in each video source and abstracting frame information of video frames contained in each video data;
the source address information comprises an address IP, an address gateway and an address number sequence;
The video information abstract is used for recording video compression information corresponding to each video data;
the frame information abstract comprises a frame length, a frame resolution and a frame rate;
Constructing a top layer index structure according to source address information corresponding to a video source, constructing a middle layer index structure which is equivalent to the number of current video data under the top layer index structure according to video information abstracts corresponding to each video data under the video source, and constructing a bottom layer index structure which is equivalent to the number of current video frames under the middle layer index structure according to frame information abstracts corresponding to each video frame under the video data;
And distributing blank files with corresponding numbers according to the numbers of the top-layer index structures, taking the middle-layer index structures and the bottom-layer index structures which are correspondingly included by the top-layer index structures as record items of the blank files, further constructing a total index file corresponding to each video source, setting an index traversal starting point, linking the total index file corresponding to each video source to the index traversal starting point, and further constructing a multi-level index file for traversing all video sources, all video data included by the video sources and all video frames included by the video data.
Further, the process of mapping the video data corresponding to all the video sources into the multi-level index file, and further constructing the hierarchical search index corresponding to each video data includes:
Mapping each video source into a top layer index structure in a multi-level index file, taking the number of each video source as a top layer index, mapping all video data included in each video source into a middle layer index structure in the multi-level index file, taking the number of each video data as a middle layer index, mapping all video frames included in each video data into a bottom layer index structure in the multi-level index file, taking the timestamp of each video frame as a bottom layer index, and integrating the top layer index, the middle layer index and the bottom layer index to construct a layered retrieval index corresponding to each video data.
Further, the process of configuring the query platform includes:
a query platform sends a link request to a preset server, and the server receives the link request and judges whether the IP address of the link request is in an address form corresponding to the server;
If yes, linking the query platform with the server;
If not, judging that the current query platform has system risk, acquiring a platform log recorded with the current system risk, inputting a plurality of log record items included in the platform log into a preset risk library, matching specific risk behaviors and corresponding risk detail information according to the content of each log record item by the risk library, synchronously matching risk restoration parameters corresponding to each risk detail information, restoring each specific risk behavior through the corresponding risk restoration parameters, further restoring the system risk of the current query platform, and continuing to link the current query platform and the server after the system risk restoration is completed;
and after the link between the query platform and the server is successful, the query platform is configured successfully, otherwise, the repairing of the specific risk behaviors and the link between the query platform and the server are continued until the link between the query platform and the server is successful.
Further, the process of guiding the hierarchical search index of all the video data into the terminal database correspondingly arranged on the query platform comprises the following steps:
the query platform is provided with a corresponding terminal database, the hierarchical search indexes of all video data corresponding to all video sources are led into the terminal database, an uploading interface corresponding to the query platform is distributed for each video source, each uploading interface is correspondingly associated with an interface sequence, and the hierarchical search indexes of all video data included by each video source are led into the terminal database through the corresponding uploading interfaces;
Before importing, judging whether an interface sequence corresponding to each uploading interface is tampered, if so, packaging the hierarchical search indexes of all video data of the corresponding uploading interface into a blank file, marking the blank file as an abnormal file, importing the abnormal file into a data security area corresponding to a terminal database, and if not, directly guiding the hierarchical search indexes of all video data of the uploading interface into a data storage area set by the terminal database for storage;
Traversing the hierarchical search index of each video data included in the abnormal file in the data security area, sequentially judging whether each hierarchical search index is changed or not, if so, updating the current hierarchical search index to the correct hierarchical search index before the change is not carried out, if not, not carrying out any operation on the current hierarchical search index until the update of the hierarchical search index with the error in the abnormal file is completed, marking the abnormal file as the security file after the update of the hierarchical search index with the error in the abnormal file is completed, and transferring and storing all the hierarchical search indexes included in the security file into the data storage area.
Further, the process of inputting the search information of the video data to be searched into the query platform, performing index matching according to the search information and the hierarchical search index, and locating the target video data includes:
The search information comprises one type of information and two types of information, the one type of information is used for positioning target video data, the one type of information comprises a top layer index and a middle layer index of the video data which are required to be searched currently, the two types of information are used for single-point positioning of a plurality of video frames included in the target video data, and the two types of information comprise bottom layer indexes of all video frames corresponding to the current target video data;
Inputting the edited search information into a query platform, performing index matching with a hierarchical search index of the query platform, taking the hierarchical search index meeting the requirement of a top-level index and a middle-level index corresponding to one type of information in a terminal database as a target index, positioning video data corresponding to the target index as target video data, and searching a plurality of corresponding video frames corresponding to the current target video data according to the bottom-level index corresponding to the two types of information after positioning the target video data.
Compared with the prior art, the invention has the beneficial effects that: the method comprises the steps of obtaining video data corresponding to a plurality of video sources, sequentially carrying out frame cutting, dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source, carrying out feature processing on all video frames of the video data corresponding to each video source, further obtaining feature frame information corresponding to each video data, constructing a multi-level index file, mapping the video data corresponding to all video sources into the multi-level index file, further constructing a hierarchical search index corresponding to each video data, guiding the hierarchical search index of all video data into a terminal database arranged on a configured query platform, inputting search information of the video data to be searched into the query platform, carrying out index matching according to the search information and the hierarchical search index, and further locating target video data, so that efficient and accurate search of the video data is realized to a certain extent.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
As shown in fig. 1, a multi-level file index management method for video data includes the steps of:
Step S1: acquiring video data corresponding to a plurality of video sources, sequentially carrying out frame cutting on all the video data, dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source, carrying out feature processing on the video frames corresponding to the video data of each video source, and further acquiring feature frame information corresponding to each video data;
Step S2: constructing a multi-level index file, mapping video data corresponding to all video sources into the multi-level index file, and further constructing a hierarchical retrieval index corresponding to each video data;
Step S3: configuring a query platform, guiding hierarchical search indexes of all video data into a terminal database correspondingly arranged in the query platform, inputting search information of the video data to be searched into the query platform, and performing index matching according to the search information and the hierarchical search indexes so as to locate target video data.
It should be further noted that, in a specific implementation process, the process of acquiring video data corresponding to a plurality of video sources, sequentially performing frame cutting on all video data, and further dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source includes:
Numbering a plurality of video sources, and marking the numbers as i, wherein i=1, 2,3, … … and n are natural numbers larger than 0, the video source with the number i correspondingly comprises a plurality of video data, and all video data correspondingly comprising the video source with the number i are constructed into a corresponding video set and marked as omega { i };
Numbering all video data contained in a video set omega { i }, each video data is recorded as VD [ i ] [1], VD [ i ] [2], VD [ i ] [3], … …, VD [ i ] [ m ], m is a natural number larger than 0, all video data corresponding to all video sources are obtained, all video data corresponding to i=3-n video sources are sequentially subjected to frame cutting from i=1 video source, VD [1] [2], … …, VD [1] [ m ], after the current video source corresponds to all video data, frame cutting of the video source corresponding to i=2 is continued, namely frame cutting of the video data VD [2] [1], VD [2] [ … …, VD [2] [ m ] is carried out, and frame cutting is repeated on all video data corresponding to the video source with i=3-n video source, wherein the video data corresponding to i=n comprises VD [ n ] [1] [2] [ VD [3] n ] [ … … ] [ VD ];
dividing video data included in each video source into a plurality of corresponding video frames through frame cutting, merging the number of the video source where the video frame is located, the number of the video data where the video frame is located and a corresponding time stamp on the video data where the video frame is located as a frame identifier of the current video frame, marking the frame identifier as Z, marking the time stamp as T, and then Z=i-m-T.
The meaning of z=i-m-T is: the m-th video data in the i-th video source corresponds to a video frame with a time stamp of T, the video data comprises a plurality of video frames, and the length units of the video frames are as follows: s, the length of the video frames is set to be 2S, and the corresponding frame start time of each video frame is taken as the time stamp of the corresponding video frame.
It should be further noted that, in a specific implementation process, the feature processing is performed on a plurality of video frames corresponding to video data of each video source, and a process of obtaining feature frame information corresponding to each video data further includes:
the feature processing of all video frames comprises feature extraction and feature vector construction;
Obtaining color features, edge features and texture features of each video frame through feature extraction, wherein the color features are performed by calculating a color histogram corresponding to the video frame, the edge features are performed through a Canny edge detection algorithm, and the texture features are performed through a local binary pattern;
The feature vector is constructed as follows: constructing a vector construction model for vectorization processing by a convolutional neural network technology, using a first vector parameter which is the vector construction model and has the highest color feature ratio in a video frame included in each video data, using a second vector parameter which is the vector construction model and has the highest edge feature ratio in the video frame included in each video data, and using a third vector parameter which is the vector construction model and has the highest texture feature ratio in the video frame included in each video data;
inputting the first vector parameter, the second vector parameter and the third vector parameter into a vector construction model, and further constructing a conventional feature vector corresponding to the color feature, a conventional feature vector corresponding to the edge feature and a conventional feature vector corresponding to the texture feature;
Setting vector distinguishing angle ranges of the color features, the edge features and the texture features corresponding to conventional feature vectors, wherein the vector distinguishing angle ranges comprise a first positive interval range, a second positive interval range, a first negative interval range and a second negative interval range;
Wherein the first positive interval range is the angle range of 0-90 degrees on the right side of the normal feature vector, the second positive interval range is the angle range of 91-80 degrees on the right side of the normal feature vector, the first negative interval range is the angle range of 0-90 degrees on the left side of the normal feature vector, and the second negative interval range is the angle range of 91-180 degrees on the left side of the normal feature vector;
it should be noted that the vector discrimination angle ranges of the color feature, the edge feature and the texture feature corresponding to the conventional feature vector are all the above cases;
Inputting color features, edge features and texture features corresponding to all video frames into a vector construction model, further obtaining color features, edge features and texture feature vectors corresponding to each video frame, obtaining the degrees of included angles between the feature vectors of the color features corresponding to each video frame and corresponding conventional feature vectors, obtaining the degrees of included angles between the feature vectors of the edge features corresponding to each video frame and corresponding conventional feature vectors, and obtaining the degrees of included angles between the feature vectors of the texture features corresponding to each video frame and corresponding conventional feature vectors;
When the included angle degree is in a first forward interval range, the corresponding feature vector is marked as a class A feature vector, and when the included angle degree is in a second forward interval range, the corresponding feature vector is marked as a class B feature vector;
When the included angle degree is in a first negative interval range, the corresponding feature vector is marked as a C-type feature vector, and when the included angle degree is in a second negative interval range, the corresponding feature vector is marked as a D-type feature vector;
And taking all the marked feature vectors of the video frames corresponding to each video data as the feature frame information of the corresponding video data.
It should be further noted that, in a specific implementation process, the process of constructing the multi-level index file includes:
Acquiring source address information corresponding to each video source, acquiring video information abstracts corresponding to all video data included by each video source, and acquiring frame information abstracts of video frames included by each video data;
The source address information is used for recording the source of each video source, and comprises an address IP, an address gateway and an address digital sequence;
The video information abstract is used for recording video compression information corresponding to each video data, and the video compression information is used for reflecting the overview content of each video data;
the frame information abstract comprises a frame length, a frame resolution and a frame rate;
Constructing a top layer index structure according to source address information corresponding to a video source, constructing a middle layer index structure corresponding to the top layer index structure and equivalent to the number of current video data according to video information abstracts corresponding to each video data under the video source, and constructing a bottom layer index structure corresponding to the middle layer index structure and equivalent to the number of current video frames according to frame information abstracts corresponding to each video frame under the video data;
distributing a corresponding number of blank files according to the number of the top-layer index structures, taking the middle-layer index structure and the bottom-layer index structure which are correspondingly included in the top-layer index structures as record items of the blank files, and further constructing a total index file corresponding to each video source;
setting an index traversal starting point, and linking a total index file corresponding to each video source to the index traversal starting point, so as to construct a multi-level index file for traversing all video sources, all video data included by the video sources and all video frames included by the video data.
It should be further noted that, in a specific implementation process, the process of mapping the video data corresponding to all the video sources into the multi-level index file, and further constructing the hierarchical search index corresponding to each video data includes:
Mapping each video source into a top-level index structure in a multi-level index file according to the sequence of the serial numbers from small to large, taking the serial number of each video source as a top-level index, mapping all video data included in each video source into a middle-level index structure in the multi-level index file according to the sequence of the serial numbers from left to right and from top to bottom, taking the serial number of each video data as a middle-level index, mapping all video frames included in each video data into a bottom-level index structure in the multi-level index file, and taking the time stamp of each video frame as a bottom-level index;
And integrating the top-level index, the middle-level index and the bottom-level index to construct a hierarchical retrieval index corresponding to each video data, wherein the hierarchical retrieval index is used for subsequent rapid retrieval of the video data.
It should be further noted that, in a specific implementation process, the process of configuring the query platform includes:
the method comprises the steps that a query platform is arranged for carrying out query operation of video data corresponding to different video sources and video frames corresponding to each video data, a link request is sent to a preset server by the query platform, the server receives the link request and judges whether an IP address of the link request is in an address form corresponding to the server;
If yes, linking the query platform with the server;
If not, judging that the current query platform has system risk, acquiring a platform log recorded with the current system risk, inputting a plurality of log record items included in the platform log into a preset risk library, matching specific risk behaviors and corresponding risk detail information according to the content of each log record item by the risk library, synchronously matching risk restoration parameters corresponding to each risk detail information, restoring each specific risk behavior through the corresponding risk restoration parameters, further restoring the system risk of the current query platform, and continuing to link the current query platform and the server after the system risk restoration is completed;
and after the link between the query platform and the server is successful, the query platform is configured successfully, otherwise, the repairing of the specific risk behaviors and the link between the query platform and the server are continued until the link between the query platform and the server is successful.
It should be further noted that, in a specific implementation process, the process of guiding the hierarchical search index of all video data into the terminal database correspondingly set by the query platform includes:
The query platform is provided with a corresponding terminal database, a data import period corresponding to the terminal database is set, the data import period is recorded as T Introduction of ,T Introduction of ＝[t1,t2, T 1 is the starting time of the data import period, and T 2 is the ending time of the data import period;
in the data import period, the hierarchical search indexes of all video data corresponding to all video sources are led into a terminal database, an uploading interface corresponding to a query platform is allocated for each video source, each uploading interface is correspondingly associated with an interface sequence, and the hierarchical search indexes of all video data included in each video source are led into the terminal database through the corresponding uploading interfaces;
Before importing, judging whether an interface sequence corresponding to each uploading interface is tampered, if so, packaging the hierarchical search indexes of all video data of the corresponding uploading interface into a blank file, marking the blank file as an abnormal file, importing the abnormal file into a data security area corresponding to a terminal database, and if not, directly guiding the hierarchical search indexes of all video data of the uploading interface into a data storage area set by the terminal database for storage;
Traversing the hierarchical search index of each video data included in the abnormal file in the data security area, sequentially judging whether each hierarchical search index is changed or not, if so, indicating that the hierarchical search index is wrong, updating the current hierarchical search index to the correct hierarchical search index before the change is not carried out, and if not, carrying out no operation on the current hierarchical search index; until the updating of the hierarchical search index with errors in all abnormal files is completed;
and after the hierarchical search indexes with errors in all the abnormal files are updated, marking the abnormal files as safe files, and transferring and storing all the hierarchical search indexes contained in the safe files into a data storage area.
It should be further noted that, in a specific implementation process, the process of inputting the search information of the video data to be searched into the query platform, performing index matching according to the search information and the hierarchical search index, and further locating the target video data includes:
setting a search period, and recording the search period as T Retrieval and T Retrieval ＝[t Starting from the beginning ,t Terminal (A) , wherein T Starting from the beginning is the starting time of the search period, T Terminal (A) is the ending time of the search period, and editing search information corresponding to video data to be searched at the starting time corresponding to T Starting from the beginning , and the search information comprises one type of information and two types of information;
The information is used for positioning target video data, and comprises a top-level index and a middle-level index of the video data to be searched currently;
the second-class information is used for single point positioning of a plurality of video frames included in the target video data, and the second-class information comprises bottom layer indexes of all video frames corresponding to the current target video data;
Inputting the edited search information into a query platform, performing index matching by the query platform according to the search information and the hierarchical search indexes, taking the hierarchical search indexes meeting the first category of information corresponding to the top-level index and the middle-level index in the terminal database as target indexes, and positioning video data corresponding to the target indexes as target video data;
after the target video data to be searched is positioned, searching a plurality of corresponding video frames corresponding to the current target video data according to the bottom layer index corresponding to the second class information, and further completing single point positioning of each video frame.
It should be noted that, the quick positioning of the video data to be searched is realized through the first type of information, the positioned video data is used as the target video data, and all video frames corresponding to the target video data are further positioned through the second type of information, so that the more detailed video characteristics of the video data are obtained.
The above embodiments are only for illustrating the technical method of the present invention and not for limiting the same, and it should be understood by those skilled in the art that the technical method of the present invention may be modified or substituted without departing from the spirit and scope of the technical method of the present invention.
Claims (8)
1. A multi-level file index management method for video data, comprising the steps of:
Step S1: acquiring video data corresponding to a plurality of video sources, sequentially carrying out frame cutting on all the video data, dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source, carrying out feature processing on the video frames corresponding to the video data of each video source, and further acquiring feature frame information corresponding to each video data;
Step S2: constructing a multi-level index file, mapping video data corresponding to all video sources into the multi-level index file, and further constructing a hierarchical retrieval index corresponding to each video data;
Step S3: configuring a query platform, guiding hierarchical search indexes of all video data into a terminal database correspondingly arranged in the query platform, inputting search information of the video data to be searched into the query platform, and performing index matching according to the search information and the hierarchical search indexes so as to locate target video data.
2. The method for multi-level file index management of video data according to claim 1, wherein the process of obtaining video data corresponding to a plurality of video sources, sequentially performing frame cutting on all video data, and dividing the video data corresponding to each video source into a plurality of video frames corresponding to each video source comprises:
Numbering a plurality of video sources, wherein i=1, 2,3, … …, n and n are natural numbers larger than 0, the video source with the number i comprises a plurality of video data, all the video data comprising the video source with the number i are constructed into a video set, and are denoted as omega { i }, all the video data comprising the omega { i } are numbered, each video data is denoted as VD [ i ] [1], VD [ i ] [2], VD [ i ] [3], … …, VD [ i ] [ m ], and m is a natural number larger than 0, all the video data corresponding to all the video sources are obtained, and from the video source with i=1, all the video data comprising the video source are sequentially subjected to frame cutting, VD [1] [2], … …, VD [1] [ m ], and the operation is repeated until the frame cutting of all the video data corresponding to the video source with i=n is completed;
Dividing video data included in each video source into a plurality of corresponding video frames through frame cutting, and merging the number of the video source where the video frame is located, the number of the video data where the video frame is located and the corresponding time stamp on the video data where the video frame is located as a frame identifier of the current video frame.
3. The method for multi-level file index management of video data according to claim 2, wherein the process of performing feature processing on a plurality of video frames corresponding to video data from each video source to obtain feature frame information corresponding to each video data comprises:
the feature processing of all video frames comprises feature extraction and feature vector construction;
Acquiring color features, edge features and texture features of each video frame through feature extraction;
The feature vector is constructed as: constructing a vector construction model for vectorization processing by a convolutional neural network technology, and taking video frames with the highest color feature ratio, the highest edge feature ratio and the highest texture feature ratio in video frames included in each video data as a first vector parameter, a second vector parameter and a third vector parameter corresponding to the vector construction model respectively;
Inputting the first vector parameter, the second vector parameter and the third vector parameter into a vector construction model, further constructing a conventional feature vector corresponding to color features, a conventional feature vector corresponding to edge features and a conventional feature vector corresponding to texture features, and setting vector distinguishing angle ranges of the respective conventional feature vectors, wherein the vector distinguishing angle ranges comprise a first positive interval range, a second positive interval range, a first negative interval range and a second negative interval range;
Inputting color features, edge features and texture features corresponding to all video frames into a vector construction model, further obtaining color features, edge features and texture feature vectors corresponding to each video frame, obtaining the degrees of included angles between the feature vectors of the color features corresponding to each video frame and corresponding conventional feature vectors, the degrees of included angles between the feature vectors of the edge features corresponding to each video frame and corresponding conventional feature vectors, and the degrees of included angles between the feature vectors of the texture features corresponding to each video frame and corresponding conventional feature vectors;
When the included angle degree is in a first positive interval range, the corresponding feature vector is marked as a class A feature vector, when the included angle degree is in a second positive interval range, the corresponding feature vector is marked as a class B feature vector, when the included angle degree is in a first negative interval range, the corresponding feature vector is marked as a class C feature vector, when the included angle degree is in a second negative interval range, the corresponding feature vector is marked as a class D feature vector, and all marked feature vectors of video frames corresponding to each video data are used as feature frame information of the corresponding video data.
4. A multi-level file index management method for video data according to claim 3, wherein the process of constructing the multi-level index file comprises:
Acquiring source address information corresponding to each video source, and abstracting video information corresponding to all video data contained in each video source and abstracting frame information of video frames contained in each video data;
the source address information comprises an address IP, an address gateway and an address number sequence;
The video information abstract is used for recording video compression information corresponding to each video data;
the frame information abstract comprises a frame length, a frame resolution and a frame rate;
Constructing a top layer index structure according to source address information corresponding to a video source, constructing a middle layer index structure which is equivalent to the number of current video data under the top layer index structure according to video information abstracts corresponding to each video data under the video source, and constructing a bottom layer index structure which is equivalent to the number of current video frames under the middle layer index structure according to frame information abstracts corresponding to each video frame under the video data;
And distributing blank files with corresponding numbers according to the numbers of the top-layer index structures, taking the middle-layer index structures and the bottom-layer index structures which are correspondingly included by the top-layer index structures as record items of the blank files, further constructing a total index file corresponding to each video source, setting an index traversal starting point, linking the total index file corresponding to each video source to the index traversal starting point, and further constructing a multi-level index file for traversing all video sources, all video data included by the video sources and all video frames included by the video data.
5. The method of claim 4, wherein mapping video data corresponding to all video sources into a multi-level index file, and constructing a hierarchical index corresponding to each video data comprises:
Mapping each video source into a top layer index structure in a multi-level index file, taking the number of each video source as a top layer index, mapping all video data included in each video source into a middle layer index structure in the multi-level index file, taking the number of each video data as a middle layer index, mapping all video frames included in each video data into a bottom layer index structure in the multi-level index file, taking the timestamp of each video frame as a bottom layer index, and integrating the top layer index, the middle layer index and the bottom layer index to construct a layered retrieval index corresponding to each video data.
6. The multi-level file index management method for video data of claim 5, the method is characterized in that the process of configuring the query platform comprises the following steps:
a query platform sends a link request to a preset server, and the server receives the link request and judges whether the IP address of the link request is in an address form corresponding to the server;
If yes, linking the query platform with the server;
If not, judging that the current query platform has system risk, acquiring a platform log recorded with the current system risk, inputting a plurality of log record items included in the platform log into a preset risk library, matching specific risk behaviors and corresponding risk detail information according to the content of each log record item by the risk library, synchronously matching risk restoration parameters corresponding to each risk detail information, restoring each specific risk behavior through the corresponding risk restoration parameters, further restoring the system risk of the current query platform, and continuing to link the current query platform and the server after the system risk restoration is completed;
and after the link between the query platform and the server is successful, the query platform is configured successfully, otherwise, the repairing of the specific risk behaviors and the link between the query platform and the server are continued until the link between the query platform and the server is successful.
7. The method for multi-level file index management for video data according to claim 6, wherein the step of guiding the hierarchical search index of all video data into the terminal database provided correspondingly to the query platform comprises:
the query platform is provided with a corresponding terminal database, the hierarchical search indexes of all video data corresponding to all video sources are led into the terminal database, an uploading interface corresponding to the query platform is distributed for each video source, each uploading interface is correspondingly associated with an interface sequence, and the hierarchical search indexes of all video data included by each video source are led into the terminal database through the corresponding uploading interfaces;
Before importing, judging whether an interface sequence corresponding to each uploading interface is tampered, if so, packaging the hierarchical search indexes of all video data of the corresponding uploading interface into a blank file, marking the blank file as an abnormal file, importing the abnormal file into a data security area corresponding to a terminal database, and if not, directly guiding the hierarchical search indexes of all video data of the uploading interface into a data storage area set by the terminal database for storage;
Traversing the hierarchical search index of each video data included in the abnormal file in the data security area, sequentially judging whether each hierarchical search index is changed or not, if so, updating the current hierarchical search index to the correct hierarchical search index before the change is not carried out, if not, not carrying out any operation on the current hierarchical search index until the update of the hierarchical search index with the error in the abnormal file is completed, marking the abnormal file as the security file after the update of the hierarchical search index with the error in the abnormal file is completed, and transferring and storing all the hierarchical search indexes included in the security file into the data storage area.
8. The method of claim 7, wherein inputting search information of video data to be searched into the search platform, performing index matching with the hierarchical search index according to the search information, and locating target video data comprises:
The search information comprises one type of information and two types of information, the one type of information is used for positioning target video data, the one type of information comprises a top layer index and a middle layer index of the video data which are required to be searched currently, the two types of information are used for single-point positioning of a plurality of video frames included in the target video data, and the two types of information comprise bottom layer indexes of all video frames corresponding to the current target video data;
Inputting the edited search information into a query platform, performing index matching with a hierarchical search index of the query platform, taking the hierarchical search index meeting the requirement of a top-level index and a middle-level index corresponding to one type of information in a terminal database as a target index, positioning video data corresponding to the target index as target video data, and searching a plurality of corresponding video frames corresponding to the current target video data according to the bottom-level index corresponding to the two types of information after positioning the target video data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411045767.6A CN118839031A (en) | 2024-08-01 | 2024-08-01 | Multi-level file index management method for video data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411045767.6A CN118839031A (en) | 2024-08-01 | 2024-08-01 | Multi-level file index management method for video data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118839031A true CN118839031A (en) | 2024-10-25 |
Family
ID=93147260
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411045767.6A Pending CN118839031A (en) | 2024-08-01 | 2024-08-01 | Multi-level file index management method for video data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118839031A (en) |
-
2024
- 2024-08-01 CN CN202411045767.6A patent/CN118839031A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7409401B2 (en) | Method and system for supporting multivalue attributes in a database system | |
| CN115145906B (en) | Preprocessing and completion method for structured data | |
| WO2009031915A1 (en) | Method and a system for storing, retrieving and extracting information on the basis of low-organised and decentralised datasets | |
| US20200341903A1 (en) | Data caching, dynamic code generation, and data visualization technology | |
| CN109213820A (en) | Method for realizing fusion use of multiple types of databases | |
| CN108280159A (en) | A method of converting chart database to relational database | |
| CN113641653A (en) | Historical data migration method and system based on domestic dream database | |
| CN115391439B (en) | Document data export method, device, electronic equipment and storage medium | |
| CN116452123A (en) | Method and device for generating characteristic value of inventory item and computer equipment | |
| CN114385587B (en) | Construction method and query method for relational database version snapshot | |
| CN115237914A (en) | Tamper-resistant index structure and construction, storage and query methods thereof | |
| CN115329504A (en) | BOM construction method based on complex product structure | |
| CN118839031A (en) | Multi-level file index management method for video data | |
| CN111917861A (en) | Knowledge storage method and system based on block chain and knowledge graph and application thereof | |
| CN104376000A (en) | Webpage attribute determination method and webpage attribute determination device | |
| CN117093556A (en) | Log classification method, device, computer equipment and computer readable storage medium | |
| CN116955469A (en) | Service alarm tracing method based on blood margin analysis | |
| CN115470223A (en) | Data lake data incremental consumption method based on two-layer time identification | |
| CN111143582B (en) | Multimedia resource recommendation method and device for updating association words in double indexes in real time | |
| CN114090558A (en) | Data quality management method and device for database | |
| CN116501788B (en) | Storehouse lake integrated data management and control platform | |
| CN115861688B (en) | Medicine packaging and container appearance modeling identification and counting method and system | |
| CN115510144B (en) | Method and system for capturing real-time change data of database | |
| CN117176550B (en) | Integrated operation maintenance method and system based on fault identification | |
| CN112463890B (en) | Cross-system data sharing method based on block chain and machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: No. 122 Xingguo Avenue, Xingguo Town, Yangxin County, Huangshi City, Hubei Province 435000 Applicant after: Hubei Yifeng Digital Technology Co.,Ltd. Address before: No. 122, Xingguo Avenue, Xingguo Town, Yangxin County, Huangshi City, Hubei Province, 435200 Applicant before: Hubei Yifeng Digital Technology Co.,Ltd. Country or region before: China |
|
| SE01 | Entry into force of request for substantive examination |