CN113205010B - Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering - Google Patents
Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering Download PDFInfo
- Publication number
- CN113205010B CN113205010B CN202110418264.9A CN202110418264A CN113205010B CN 113205010 B CN113205010 B CN 113205010B CN 202110418264 A CN202110418264 A CN 202110418264A CN 113205010 B CN113205010 B CN 113205010B
- Authority
- CN
- China
- Prior art keywords
- frame
- target
- video
- data block
- sequence
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/49—Segmenting video sequences, i.e. computational techniques such as parsing or cutting the sequence, low-level clustering or determining units such as shots or scenes
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses an intelligent disaster-exploration on-site video frame efficient compression system based on target clustering, wherein a frame sequence extraction module is used for obtaining a monitoring video frame sequence; the video frame segmentation module is used for obtaining a foreground data block and a target data block; the video frame judgment module is used for obtaining a target frame set and a monitoring video background frame set; the target frame clustering module is used for carrying out clustering analysis on target frames in the monitoring video target frame set; the video frame compression module is used for obtaining target frame compression data and background frame compression data. The invention can eliminate information redundancy, improve the compression ratio, accelerate the compression speed and further improve the video data transmission speed.
Description
Technical Field
The invention relates to the technical field of intelligent disaster exploration emergency command site video information processing, in particular to a system and a method for efficiently compressing a video frame of an intelligent disaster exploration site based on target clustering.
Background
In the field of electric power emergency repair command, a plurality of monitoring devices based on intelligent sensing are deployed on site and are responsible for collecting site data, the site data are uploaded to a command center in real time through a wireless network, a central server decompresses, analyzes and excavates video data, the command center manages and controls the site in real time, and decision support is provided for emergency command.
At present, video monitoring is one of important means for collecting power emergency repair data, and real-time uploading of massive high-definition video data is completed, so that timeliness of emergency repair is directly limited. In a disaster site, in consideration of factors such as reduction of reliability of a wireless channel, limitation of broadband resources, reduction of video definition and the like, great challenges are brought to reliable real-time transmission of massive high-definition video data, and a high-magnification video frame compression method needs to be researched to reduce transmission of redundant information in video stream and realize reliable and rapid transmission of video stream data.
However, existing video compression methods, such as h.264/h.265, implement content compression based on content similarity, and have a large dependency on hardware storage resources and a low compression rate in an application process, which further limits a throughput rate in a video decompression process. Therefore, the research on the high-magnification video frame compression method suitable for the fixed position monitoring equipment in the intelligent disaster exploration site has important significance in ensuring the video stream compression rate, reducing the video data transmission amount, increasing the video decompression throughput, realizing the safe, reliable and rapid transmission of the video stream, and providing support for the auxiliary decision of the emergency repair command center.
Disclosure of Invention
The invention aims to provide a system and a method for efficiently compressing video frames of an intelligent disaster-exploration site based on target clustering, which can solve the problems of high dependence on hardware storage resources, low compression ratio and low video decompression throughput in the application process of the conventional video compression technology.
In order to achieve the purpose, the invention designs an intelligent disaster-exploration field video frame high-efficiency compression system based on target clustering, which is characterized in that: the video frame clustering system comprises a frame sequence extraction module, a video frame segmentation module, a video frame judgment module, a target frame clustering module and a video frame compression module;
the frame sequence extraction module is used for converting the monitoring video data in a preset time period into a corresponding monitoring video frame sequence;
the video frame segmentation module is used for carrying out image segmentation on the monitoring video frame sequence by adopting an improved watershed algorithm based on gray threshold segmentation to form a foreground data block and a target data block;
the video frame judging module is used for judging each frame of the monitoring video frame sequence according to the sequence of the frame sequence, judging whether each frame of the monitoring video frame sequence contains a target data block, wherein the frames containing the target data block form a monitoring video target frame set arranged according to a time sequence, the frames not containing the target data block form a monitoring video background frame set arranged according to the time sequence, the monitoring video target frame set contains a foreground data block and a target data block, and the monitoring video background frame set only contains the foreground data block;
the target frame clustering module is used for clustering and analyzing target frames in the monitoring video target frame set, dividing the target frames containing the same target into a subset, and performing segmentation processing on the monitoring video frame sequence according to a clustering result to enable each video frame to contain the same target;
the video frame compression module is used for compressing the video frame segments containing the same target by adopting a video frame compression method based on target frame combination to obtain target frame compression data; the video frame compression module is further used for compressing the monitoring video background frame set by adopting a video frame compression method based on background frame combination to obtain background frame compression data.
The invention has the beneficial effects that:
the invention divides the video frame into the target data block set and the background data block set by the image segmentation, the target clustering and other technologies, and compresses the data blocks respectively, and the adoption of the separate compression method is favorable for eliminating information redundancy, improving the compression ratio, accelerating the compression speed and further improving the video data transmission speed because the similarity between the blocks in the target data block set is extremely high and the similarity between the blocks in the background data block set is also high.
Drawings
FIG. 1 is a schematic view of the structure of the present invention;
FIG. 2 is a flowchart of an efficient compression method for video frames in an intelligent disaster-exploration site based on target clustering according to an embodiment of the present invention;
FIG. 3 is a flowchart of an image region segmentation method based on threshold segmentation according to an embodiment of the present invention;
FIG. 4 is a flowchart of a video compression method based on H.264 encoding rules according to an embodiment of the present invention;
the system comprises a frame sequence extracting module 1, a video frame segmentation module 2, a video frame judgment module 3, a target frame clustering module 4, a video frame compression module 5 and a video reconstruction module 6.
Detailed Description
The invention is described in further detail below with reference to the following figures and specific examples:
as shown in fig. 1, the system for efficiently compressing the video frames in the intelligent disaster-exploring field based on the target clustering is characterized in that: the video frame clustering system comprises a frame sequence extraction module 1, a video frame segmentation module 2, a video frame judgment module 3, a target frame clustering module 4 and a video frame compression module 5;
the frame sequence extraction module 1 is used for converting the monitoring video data in a preset time period into a corresponding monitoring video frame sequence;
the video frame segmentation module 2 is used for carrying out image segmentation on the monitoring video frame sequence by adopting an improved watershed algorithm based on gray threshold segmentation to form a foreground data block and a target data block, and the operation process is simple and stable in the mode, and a better image segmentation result can be obtained;
the video frame judging module 3 is used for judging each frame of the monitoring video frame sequence according to the sequence of the frame sequence, judging whether each frame of the monitoring video frame sequence contains a target data block, wherein the frames containing the target data block form a monitoring video target frame set arranged according to a time sequence, the frames not containing the target data block form a monitoring video background frame set arranged according to the time sequence, the monitoring video target frame set contains a foreground data block and a target data block, the monitoring video background frame set only contains the foreground data block, and the video frames are divided into the foreground data block and the target data block according to an image segmentation result and are compressed separately, so that the compression speed can be effectively increased;
the target frame clustering module 4 is used for clustering analysis of target frames in the monitoring video target frame set, dividing the target frames containing the same target into a subset, performing segmentation processing on the monitoring video frame sequence according to a clustering result to enable each segment of video frame to contain the same target, clustering the target frames, and dividing the frame sequence containing the targets of the same class into a subset through clustering, so that the method is favorable for higher subsequent compression speed;
the video frame compression module 5 is used for compressing the video frame segments containing the same target by adopting a video frame compression method based on target frame combination to obtain target frame compression data; the video frame compression module 5 is further configured to perform compression processing on the monitoring video background frame set by using a video frame compression method based on background frame merging to obtain background frame compression data, and separately compress the target frame and the background frame, so as to reduce data redundancy and accelerate compression speed and efficiency.
In the above technical solution, it further includes a video reconstruction module 6, where the video reconstruction module 6 is configured to perform inverse quantization and decoding on the target frame compressed data and the background frame compressed data; according to different data block division rules, image frame recombination is carried out to obtain a complete target frame set and a complete background frame set; and then, by overlapping the boundary frames in the target frame data and the background frame data and deleting redundant data, the reconstruction of the whole monitoring video data can be completed.
In the above technical solution, the video frame segmentation module 2 adopts an improved watershed algorithm based on gray threshold segmentation to perform image segmentation on a monitored video frame sequence, and the specific method for forming the foreground data block and the target data block comprises:
firstly, performing median filtering denoising pretreatment on an input video frame sequence;
then, sequentially adopting an improved watershed model to perform image threshold segmentation on the filtered and denoised frame image sequence, wherein the segmentation process comprises the following steps: traversing an image, obtaining gradient data of each pixel of a frame image sequence by using a Sobel operator, performing flood treatment, starting flood treatment from the minimum gradient, marking basin regions, merging regions with the number of pixel points smaller than a given threshold, performing initial marking of the frame image sequence according to the gray level of the pixel points, wherein one marking point represents the central point of a region with the gray level within a preset range, then performing region merging according to the Euclidean distance between the marking points, generating new marking points after merging, performing super change, updating a marking region table, finally obtaining a binary image of the image, wherein the white region of the binary image is a target data block, the black region is a foreground data block, segmenting the image by using an improved watershed model, dividing the video frame into a foreground data block and a target data block according to a segmentation result, and separately compressing, thereby effectively accelerating the compression speed.
In the above technical solution, the specific process of the target frame clustering module 4 performing the segmentation processing on the monitored video frame sequence according to the clustering result is as follows: extracting SIFT (Scale-invariant feature transform) feature points of adjacent target data blocks in a monitored video frame sequence, and determining SIFT feature vectors of targets; determining the similarity between the adjacent target data blocks according to the matching number of SIFT feature points of the two target data blocks; extracting key frames according to a preset similarity threshold, and extracting all key frame sequences for the target data block; and taking the key frame as a dividing line, sequentially dividing the monitoring video frame sequence into a plurality of different video segments, wherein each segment contains the same target, and carrying out SIFT feature extraction on the images for target clustering.
In the above technical solution, the video frame compression module 5 is configured to perform quantization and coding on video frame segments containing the same target by using a video frame compression method based on target frame merging and using an h.264 coding rule to perform compression processing, so as to obtain target frame compression data.
The video frame compression module 5 is further configured to perform quantization and coding compression processing on the monitoring video background frame set by using a video frame compression method based on background frame merging and by using an h.264 coding rule, so as to obtain target frame compression data.
In the above technical solution, the video reconstruction module 6 is configured to perform inverse quantization and decoding on the target frame compressed data and the background frame compressed data according to an h.264 coding rule.
An intelligent disaster-exploration on-site video frame efficient compression method based on target clustering is characterized by comprising the following steps:
step 1: converting the monitoring video data in a preset time period into a corresponding monitoring video frame sequence;
step 2: performing image segmentation on the monitoring video frame sequence by adopting an improved watershed algorithm based on gray threshold segmentation to form a foreground data block and a target data block;
and step 3: judging each frame of the monitoring video frame sequence according to the sequence of the frame sequence, and judging whether each frame of the monitoring video frame sequence contains a target data block, wherein the frames containing the target data block form a monitoring video target frame set arranged according to a time sequence, the frames not containing the target data block form a monitoring video background frame set arranged according to the time sequence, the monitoring video target frame set contains a foreground data block and a target data block, and the monitoring video background frame set only contains the foreground data block;
and 4, step 4: performing cluster analysis on target frames in a monitoring video target frame set, dividing the target frames containing the same target into a subset, and performing segmentation processing on the monitoring video frame sequence according to a clustering result to enable each segment of video frame to contain the same target;
and 5: compressing the video frame segments containing the same target by adopting a video frame compression method based on target frame combination to obtain target frame compressed data; and compressing the monitoring video background frame set by adopting a video frame compression method based on background frame merging to obtain background frame compressed data.
The specific method for obtaining the sequence of the monitoring video frames in the step 1 is to select and download a field monitoring video, intercept a section of the field monitoring video as original video data, perform size scaling processing on images, and number the video frames in sequence according to a time sequence to obtain a video frame image sequence with uniform size.
In step 3, for the monitoring device of the fixed airplane position, the video images monitored normally in daily life are almost the same and can be regarded as background frames, when abnormal conditions such as foreign matter invasion, serious damage of the monitoring device or disaster occur in the monitoring area, new targets appear in the monitoring images, and the image frames can be regarded as target frames. Therefore, whether the video frames contain related targets, such as foreign matters of the power transmission line, typical disasters and the like, can be sequentially judged according to the segmentation result, and the video frames are determined to be target frames and background frames to form a target frame set and a background frame set;
in the step 5, the video segments containing the same target are compressed by a video frame compression method based on target frame merging, which includes the following steps:
firstly, dividing a target frame into a minimum data block containing a current target and a plurality of remaining background blocks by adopting a target-based dividing mode, and dividing all data frames containing the same target in a set by adopting the same dividing rule to obtain a divided data block set;
then, adopting a video compression method in an H.264 coding rule to carry out quantization processing on all frames in the data block, removing high-frequency components, and outputting a transformed signal set;
and finally, coding the data block by adopting a video compression method in an H.264 coding rule to form an I, P and B frame sequence, outputting the I, P and B frame sequence and compressing the video frame, wherein I \ P \ B are respectively three frames defined in an H264 protocol, the I frame is a complete coding frame, a frame which only contains a difference part code and is generated by the I frame is a P frame, and frames which refer to the front frame and the rear frame are B frames.
For the monitoring video background frame set, a video frame compression method based on background frame merging is adopted for compression, the compression method is also based on a video compression method specified in an H.264 coding rule, firstly, data block division is carried out, secondly, quantization and coding are carried out on the data block, compression is completed, and data transmission is carried out.
Referring to fig. 2, this embodiment is completed according to the following steps: extracting monitored video data according to a disaster site, and extracting a section of video as a video frame sequence (frame 1, frame 2, \ 8230; frame 1000) to be processed, wherein the length of the video frame sequence is 1000; dividing the video frame by adopting an image region division method based on threshold value division, and respectively dividing the ith frame into m i A data block; classifying all frames according to whether the image frames contain targets or not, and dividing the frames into a target frame set and a background frame set; for a target frame set, a video frame segmentation method based on SIFT feature point matching is adopted to segment target data blocks, the target data blocks are sequentially divided into p segments, and the p segments of the data blocks are sequentially encoded and compressed by adopting an H.264-based encoding rule to obtain compressed target data blocks; and for the background frame set, directly adopting an H.264 coding rule to carry out coding compression to obtain a compressed background data block. The specific implementation method comprises the following steps:
collecting the monitoring video data of the emergency repair site, intercepting a section of video as a video section to be processed, and then extracting a frame sequence of the video section to form an image frame sequence with the length of N.
Referring to fig. 3, the method for segmenting each frame of image by using an image region segmentation method based on threshold segmentation to obtain data blocks includes the following steps: first, a frame image is subjected to denoising processing. Performing median filtering denoising pretreatment on an input video frame sequence; and secondly, sequentially adopting an improved watershed model to perform image threshold segmentation on the frame image sequence. Then, outputting the image frame segmentation result; and finally, sequentially carrying out next frame image segmentation processing, repeating the steps until the last frame of the sequence is reached, and outputting a video frame sequence segmentation result.
And sequentially judging whether the video frames contain related targets such as foreign matters of a power transmission line, electric equipment, typical disasters and the like according to the segmentation result, and finally determining frames numbered as 155-185, 386-488 and 765-1000 as target data frames and frames numbered as 1-154, 186-385 and 489-764 as background data frames to form a target frame set and a background frame set.
For all image frames in the target frame set, a video frame segmentation method based on SIFT feature point matching is adopted to segment the target data blocks, and the process specifically comprises the following steps: extracting SIFT feature points of adjacent target blocks, and determining SIFT feature vectors of the targets; determining the similarity between the adjacent target blocks according to the matching number of the feature points of the two target blocks; extracting key frames, and extracting all key frame sequences for the target data block; and taking the key frame as a dividing line, and sequentially dividing the target data block into a plurality of different target sections. The third segment 765-1000 in the target frame set is finally refined into 2 video segments, namely {765-890} and {891-1000}.
Referring to fig. 4, video segments containing the same target are compressed by a video frame compression method based on target frame merging, and first, video frames belonging to the same type of target in a target frame set are divided by the same division rule by a data block division method based on the target; each data block is then quantized and encoded for video compression based on h.264 coding rules. And aiming at the background frame set, adopting a video frame compression method based on background frame merging to compress all data blocks in the background data block set, firstly adopting an average division principle to divide the data blocks of the background data, and then adopting H.264 to quantize and change the data blocks to compress the video. Resulting in compressed video data.
And performing video reconstruction on the compressed video data. According to the above process, two different types of compressed data can be obtained: target frame compressed data and background frame compressed data. The video reconstruction process is as follows: firstly, according to the H.264 coding rule, performing inverse quantization and decoding on two types of compressed data respectively; then, according to different data block division rules, image frame recombination is carried out to obtain a complete target frame set and a complete background frame set; and then, by overlapping the boundary frames in the target frame data and the background frame data and deleting redundant data, the reconstruction of the whole video segment can be completed.
Further, the video frame compression result is finally output.
In summary, the video frame is divided into the target data block set and the background data block set by the technologies of image segmentation, target clustering and the like, and the target data block set and the background data block set are combined and compressed respectively.
Details not described in this specification are within the skill of the art that are well known to those skilled in the art.
Claims (4)
1. An intelligent disaster-exploration on-site video frame efficient compression system based on target clustering is characterized in that: the video frame clustering system comprises a frame sequence extraction module (1), a video frame segmentation module (2), a video frame judgment module (3), a target frame clustering module (4) and a video frame compression module (5);
the frame sequence extraction module (1) is used for converting the monitoring video data in a preset time period into a corresponding monitoring video frame sequence;
the video frame segmentation module (2) is used for carrying out image segmentation on the monitoring video frame sequence by adopting an improved watershed algorithm based on gray threshold segmentation to form a foreground data block and a target data block;
the video frame judging module (3) is used for judging each frame of the monitoring video frame sequence according to the sequence of the frame sequence, judging whether each frame of the monitoring video frame sequence contains a target data block, wherein the frames containing the target data block form a monitoring video target frame set arranged according to a time sequence, the frames not containing the target data block form a monitoring video background frame set arranged according to the time sequence, the monitoring video target frame set contains a foreground data block and a target data block, and the monitoring video background frame set only contains the foreground data block;
the target frame clustering module (4) is used for clustering and analyzing target frames in the monitoring video target frame set, dividing the target frames containing the same target into a subset, and performing segmentation processing on the monitoring video frame sequence according to a clustering result to enable each segment of video frame to contain the same target;
the video frame compression module (5) is used for quantizing and coding video frame segments containing the same target by adopting a video frame compression method based on target frame combination by utilizing an H.264 coding rule to realize compression processing so as to obtain target frame compression data; the video frame compression module (5) is also used for compressing the monitoring video background frame set by adopting a video frame compression method based on background frame combination to obtain background frame compression data; the method for compressing the video frame segments containing the same target by adopting a video frame compression method based on target frame combination comprises the following steps:
firstly, dividing a target frame into a minimum data block containing a current target and a plurality of remaining background blocks by adopting a target-based dividing mode, and dividing all data frames containing the same target in a set by adopting the same dividing rule to obtain a divided data block set;
then, adopting a video compression method in an H.264 coding rule to carry out quantization processing on all frames in the data block, removing high-frequency components, and outputting a transformed signal set;
finally, a video compression method in an H.264 coding rule is adopted to realize the coding of the data block, an I, P and B frame sequence is formed and output, and the compression of the video frame is realized;
for a monitoring video background frame set, compressing by adopting a video frame compression method based on background frame merging, wherein the compression method is also based on a video compression method specified in an H.264 coding rule, firstly dividing data blocks, secondly quantizing and coding the data blocks, completing compression and transmitting data;
the video reconstruction method comprises a video reconstruction module (6), wherein the video reconstruction module (6) is used for carrying out inverse quantization and decoding on target frame compressed data and background frame compressed data; according to different data block division rules, image frame recombination is carried out to obtain a complete target frame set and a background frame set; then, by overlapping the boundary frames in the target frame data and the background frame data, deleting redundant data, and completing the reconstruction of the whole monitoring video data;
the video frame segmentation module (2) adopts an improved watershed algorithm based on gray threshold segmentation to perform image segmentation on a monitoring video frame sequence to form a foreground data block and a target data block, and the specific method comprises the following steps:
firstly, performing median filtering denoising pretreatment on an input video frame sequence;
then, sequentially adopting an improved watershed model to perform image threshold segmentation on the filtered and denoised frame image sequence, wherein the segmentation process comprises the following steps: obtaining gradient data of each pixel of a frame image sequence by utilizing a Sobel operator, carrying out flooding treatment from the minimum gradient, marking a basin region, then carrying out merging treatment on regions with the number of pixel points smaller than a given threshold value, carrying out initial marking on the frame image sequence according to the gray level of the pixel points, wherein one marking point represents the central point of one region with the gray level within a preset range, then carrying out region merging according to the Euclidean distance between the marking points, generating new marking points for super change after merging, updating a marking region table, and finally obtaining a binary image of the image, wherein the white region of the binary image is a target data block, and the black region is a foreground data block;
the specific process of the target frame clustering module (4) for carrying out segmentation processing on the monitoring video frame sequence according to the clustering result is as follows: extracting SIFT feature points of adjacent target data blocks in a monitoring video frame sequence, and determining SIFT feature vectors of targets; determining the similarity between the adjacent target data blocks according to the matching number of SIFT feature points of the two target data blocks; extracting key frames according to a preset similarity threshold, and extracting all key frame sequences for the target data block; the key frame is taken as a dividing line, the monitoring video frame sequence is sequentially divided into a plurality of different video segments, and each segment contains the same target.
2. The intelligent disaster-exploration on-site video frame efficient compression system based on target clustering as claimed in claim 1, wherein: the video reconstruction module (6) is used for carrying out inverse quantization and decoding on the target frame compressed data and the background frame compressed data according to an H.264 coding rule.
3. An intelligent disaster-exploration site video frame efficient compression method based on target clustering by using the system of claim 1, which is characterized in that:
step 1: converting monitoring video data in a preset time period into a corresponding monitoring video frame sequence;
step 2: performing image segmentation on the monitoring video frame sequence by adopting an improved watershed algorithm based on gray threshold segmentation to form a foreground data block and a target data block;
and step 3: judging each frame of the monitoring video frame sequence according to the sequence of the frame sequence, and judging whether each frame of the monitoring video frame sequence contains a target data block, wherein the frames containing the target data block form a monitoring video target frame set arranged according to a time sequence, the frames not containing the target data block form a monitoring video background frame set arranged according to the time sequence, the monitoring video target frame set contains a foreground data block and a target data block, and the monitoring video background frame set only contains the foreground data block;
and 4, step 4: performing cluster analysis on target frames in a monitoring video target frame set, dividing the target frames containing the same target into a subset, and performing segmentation processing on the monitoring video frame sequence according to a clustering result to enable each segment of video frame to contain the same target;
and 5: compressing the video frame segments containing the same target by adopting a video frame compression method based on target frame combination to obtain target frame compressed data; and compressing the monitoring video background frame set by adopting a video frame compression method based on background frame combination to obtain background frame compressed data.
4. The intelligent disaster-exploration on-site video frame efficient compression method based on target clustering as claimed in claim 3, wherein: the specific method for obtaining the sequence of the monitoring video frames in the step 1 is to select and download a field monitoring video, intercept a section of the field monitoring video as original video data, perform size scaling processing on images, and number the video frames in sequence according to a time sequence to obtain a video frame image sequence with uniform size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110418264.9A CN113205010B (en) | 2021-04-19 | 2021-04-19 | Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110418264.9A CN113205010B (en) | 2021-04-19 | 2021-04-19 | Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113205010A CN113205010A (en) | 2021-08-03 |
| CN113205010B true CN113205010B (en) | 2023-02-28 |
Family
ID=77027417
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110418264.9A Active CN113205010B (en) | 2021-04-19 | 2021-04-19 | Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113205010B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114157873B (en) * | 2021-11-25 | 2024-08-23 | 中国通信建设第四工程局有限公司 | Video compression method and video compression system |
| CN115278188B (en) * | 2022-07-28 | 2025-06-24 | 腾讯科技（深圳）有限公司 | A video processing method and related device |
| CN116453068A (en) * | 2023-03-01 | 2023-07-18 | 智慧互通科技股份有限公司 | Road side parking behavior identification method and system based on video compression domain |
| CN116320427B (en) * | 2023-03-31 | 2025-12-12 | 重庆大学 | Low-loss video compression method and monitoring system for drone-based crowd surveillance |
| CN118400577B (en) * | 2024-05-23 | 2025-03-28 | 威海凯思信息科技有限公司 | A video generation method and device based on AIGC |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101610411A (en) * | 2009-07-16 | 2009-12-23 | 中国科学技术大学 | Method and system for video sequence hybrid encoding and decoding |
| CN105407353A (en) * | 2014-09-11 | 2016-03-16 | 腾讯科技（深圳）有限公司 | Image compression method and apparatus |
| CN106203277A (en) * | 2016-06-28 | 2016-12-07 | 华南理工大学 | Fixed lens real-time monitor video feature extracting method based on SIFT feature cluster |
| WO2018187622A1 (en) * | 2017-04-05 | 2018-10-11 | Lyrical Labs Holdings, Llc | Video processing and encoding |
| CN111901604A (en) * | 2020-09-29 | 2020-11-06 | 创新奇智(南京)科技有限公司 | Video compression method, video reconstruction method, corresponding devices, camera and video processing equipment |
| CN112085031A (en) * | 2020-09-11 | 2020-12-15 | 河北工程大学 | Object detection method and system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2443827A4 (en) * | 2009-06-14 | 2016-11-23 | Rafael Advanced Defense Sys | SYSTEMS AND METHODS FOR CONTINUOUS READING AND VIDEO ARCHIVING WITH GEOGRAPHIC ANCHORING OF FRAME CONTENT |
| WO2013037069A1 (en) * | 2011-09-15 | 2013-03-21 | Libre Communications Inc. | Method, apparatus and computer program product for video compression |
| CN103093481B (en) * | 2013-01-28 | 2015-11-18 | 中国科学院上海微系统与信息技术研究所 | A kind of based on moving target detecting method under the static background of watershed segmentation |
-
2021
- 2021-04-19 CN CN202110418264.9A patent/CN113205010B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101610411A (en) * | 2009-07-16 | 2009-12-23 | 中国科学技术大学 | Method and system for video sequence hybrid encoding and decoding |
| CN105407353A (en) * | 2014-09-11 | 2016-03-16 | 腾讯科技（深圳）有限公司 | Image compression method and apparatus |
| CN106203277A (en) * | 2016-06-28 | 2016-12-07 | 华南理工大学 | Fixed lens real-time monitor video feature extracting method based on SIFT feature cluster |
| WO2018187622A1 (en) * | 2017-04-05 | 2018-10-11 | Lyrical Labs Holdings, Llc | Video processing and encoding |
| CN112085031A (en) * | 2020-09-11 | 2020-12-15 | 河北工程大学 | Object detection method and system |
| CN111901604A (en) * | 2020-09-29 | 2020-11-06 | 创新奇智(南京)科技有限公司 | Video compression method, video reconstruction method, corresponding devices, camera and video processing equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113205010A (en) | 2021-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113205010B (en) | Intelligent disaster-exploration on-site video frame efficient compression system and method based on target clustering | |
| US12445625B2 (en) | Coding video frame key points to enable reconstruction of video frame | |
| Zhang et al. | MSFC: Deep feature compression in multi-task network | |
| CN119031147B (en) | Video coding and decoding acceleration method and system based on learning task perception mechanism | |
| CN106231214B (en) | An Approximate Lossless Compression Method for High-speed CMOS Sensor Images Based on Adjustable Macroblocks | |
| CN105825530B (en) | Distributed lossy encoding and decoding method for coastal hyperspectral images based on region of interest | |
| US20110262049A1 (en) | Vision-based compression | |
| JPH10126775A (en) | Coding method for image signal | |
| CN110677624B (en) | Monitoring video-oriented foreground and background parallel compression method based on deep learning | |
| KR20220045920A (en) | Method and apparatus for processing images/videos for machine vision | |
| CN116896638A (en) | A data compression and encoding technology for power transmission and inspection scenarios | |
| Chen et al. | An entropy minimization histogram mergence scheme and its application in image compression | |
| CN113473142B (en) | Video encoding method, video decoding method, video encoding device, video decoding device, electronic equipment and storage medium | |
| JP2022069398A (en) | Video coding device and method, video decoding device and method, and codec system | |
| US20230239474A1 (en) | Video coding method, apparatus, and device, and medium | |
| CN115459780A (en) | Data compression method, data decompression method and related equipment | |
| Kumar et al. | Vector quantization with codebook and index compression | |
| CN113784147A (en) | A high-efficiency video coding method and system based on convolutional neural network | |
| CN113542771A (en) | Video high-efficiency compression processing method based on content weight | |
| CN104053012A (en) | A video encoding and decoding method and device based on a dictionary | |
| Corte-Real et al. | A very low bit rate video coder based on vector quantization | |
| CN110290382A (en) | A method for encoding and decoding residual data | |
| CN110545445A (en) | Video compression method and device, server and computer readable storage medium | |
| CN113691818A (en) | Video target detection method, system, storage medium and computer vision terminal | |
| Li et al. | Enlarging context with low cost: Efficient arithmetic coding with trimmed convolution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |