Video processing method and device

Abstract

The invention discloses a video processing method and device, and the method comprises the steps: obtaining a target feature and a time range interval of a to-be-processed video in response to a video processing task; obtaining a plurality of video clips according to the starting timestamp and the ending timestamp of the time range interval; respectively extracting a first key frame of each video clip according to the target features to obtain a first key frame sequence; based on a video stitching technology, integrating each video clip to obtain a target video, and extracting a second key frame sequence of the target video according to the target features; based on an image processing technology, performing comparison processing on the first key frame sequence and the second key frame sequence to obtain a target key frame; and performing video reconstruction processing on the target key frame to obtain a to-be-converted video, and obtaining image data obtained by converting the to-be-converted video. According to the method provided by the invention, the processing efficiency of the recorded video is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of video processing, and in particular to a method and device for processing recorded video. Background Art

[0002] With the continuous advancement of information technology, video processing technology has developed rapidly. Video processing technology not only improves the efficiency of video storage and transmission, but also makes video browsing more convenient.

[0003] In the prior art, in the process of extracting target image data from a complex video, the entire video data needs to be processed one by one, which results in low processing efficiency, causing the user to wait too long and affecting the user experience.

[0004] Therefore, how to optimize the processing flow of recorded videos is a technical problem that technical personnel in this field need to solve urgently. Summary of the invention

[0005] The present invention provides a method and device for processing recorded videos to solve the technical problem of low efficiency in complex video processing.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a method for processing recorded video, including:

[0007] In response to the video processing task, obtaining target features and time range intervals of the recorded video to be processed;

[0008] Obtaining multiple recorded video clips according to the start timestamp and the end timestamp of the time range interval;

[0009] Extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence;

[0010] Based on the video stitching technology, each of the recorded video clips is integrated to obtain a target video, and according to the target features, a second key frame sequence of the target video is extracted;

[0011] Based on image processing technology, the first key frame sequence and the second key frame sequence are compared to obtain a target key frame;

[0012] The target key frame is used to perform video reconstruction processing to obtain a video to be converted, and image data converted from the video to be converted is obtained.

[0013] As one of the preferred solutions, the multiple recorded video clips are obtained according to the start timestamp and the end timestamp of the time range interval, including:

[0014] Traversing the recorded video to be processed according to the start timestamp and the end timestamp;

[0015] During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained;

[0016] All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

[0017] As one of the preferred solutions, extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence includes:

[0018] Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature;

[0019] The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

[0020] As one of the preferred solutions, the video splicing technology is used to integrate each of the recorded video clips to obtain the target video, including:

[0021] Performing time alignment according to the first key frame of each of the recorded video segments to obtain a sequence of recorded video segments;

[0022] Extracting features from overlapping areas of adjacent video segments in the recorded video segment sequence to obtain adjacent corner points, and constructing a transformation matrix based on the adjacent corner points;

[0023] The overlapping areas are fused according to the transformation matrix to obtain a target video.

[0024] As one preferred solution, the comparing the first key frame sequence and the second key frame sequence based on the image processing technology to obtain the target key frame includes:

[0025] Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features;

[0026] Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair;

[0027] Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair;

[0028] A target key frame is obtained based on the similarity.

[0029] Another embodiment of the present invention provides a device for processing recorded video, including:

[0030] An acquisition module, used for acquiring target features and time range intervals of the video to be processed in response to the video processing task;

[0031] A preprocessing module, used for obtaining a plurality of recorded video clips according to a start timestamp and an end timestamp of the time range interval;

[0032] An extraction module, used for extracting the first key frame of each of the recorded video segments according to the target feature to obtain a first key frame sequence;

[0033] A splicing module, for integrating each of the recorded video clips based on a video splicing technology to obtain a target video, and extracting a second key frame sequence of the target video according to the target features;

[0034] A comparison module, configured to compare the first key frame sequence and the second key frame sequence based on image processing technology to obtain a target key frame;

[0035] The conversion module is used to perform video reconstruction processing with the target key frame to obtain the video to be converted, and obtain image data converted from the video to be converted.

[0036] As one preferred solution, the preprocessing module is specifically used for:

[0037] Traversing the recorded video to be processed according to the start timestamp and the end timestamp;

[0038] During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained;

[0039] All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

[0040] As one preferred solution, the extraction module is specifically used for:

[0041] Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature;

[0042] The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

[0043] As one of the preferred solutions, the splicing module is specifically used for:

[0044] Performing time alignment according to the first key frame of each of the recorded video segments to obtain a sequence of recorded video segments;

[0045] Extracting features from overlapping areas of adjacent video segments in the recorded video segment sequence to obtain adjacent corner points, and constructing a transformation matrix based on the adjacent corner points;

[0046] The overlapping areas are fused according to the transformation matrix to obtain a target video.

[0047] As one preferred solution, the comparison module is specifically used for:

[0048] Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features;

[0049] Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair;

[0050] Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair;

[0051] A target key frame is obtained based on the similarity.

[0052] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0053] 1) The present invention responds to video processing tasks and obtains the target features and time range intervals of the recorded video to be processed. The method can quickly locate and process the required video content, avoiding unnecessary full video scanning, thereby significantly improving processing efficiency.

[0054] 2) The present invention extracts the first key frame of each recorded video clip according to the target features, and the second key frame sequence of the integrated target video, ensuring the accuracy and completeness of the key information, which is crucial for subsequent video comparison, analysis and reconstruction.

[0055] 3) The present invention performs video reconstruction based on the target key frame, ensuring the accuracy and coherence of the reconstructed video. This reconstruction not only retains important information in the original video, but also improves the quality of the video through optimization and adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flowchart of a method for processing recorded video in one embodiment of the present invention;

[0057] Figure 2It is a structural schematic diagram of a recorded video processing device in one embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0060] In the description of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are only for illustrative purposes, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0061] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0062] An embodiment of the present invention provides a method for processing recorded video. For details, see Figure 1 , Figure 1The flowchart of the method for processing recorded video in one embodiment of the present invention is shown, including steps S1-S6:

[0063] S1: In response to the video processing task, obtaining the target features and time range of the video to be processed;

[0064] Among them, video processing tasks are usually implemented through user input, system scheduling or event triggering. After responding to the task, according to the preset rules, algorithms or user-specified conditions, target features are extracted from the video to be processed. These features can be specific objects, colors, textures, motion patterns, etc.

[0065] In a specific embodiment of the present application, the security team of the shopping mall inputs a video processing task through the user interface of the video recording processing system, specifies the surveillance camera number (e.g., the camera at the entrance) and the approximate time range (e.g., 10 pm to 2 am) to be reviewed. The video recording processing system is constructed based on the above-mentioned video recording processing method.

[0066] The security team set target features in the system, including "human detection" (for identifying possible intruders) and "abnormal motion" (for detecting abnormal object movement). Based on the input time range, the system determined the starting timestamp (10 pm) and ending timestamp (2 am) of the video clips to be processed.

[0067] S2: obtaining a plurality of recorded video clips according to the start timestamp and the end timestamp of the time range interval;

[0068] Preferably, in an embodiment of the present application, the obtaining of multiple recorded video clips according to the start timestamp and the end timestamp of the time range interval includes:

[0069] Traversing the recorded video to be processed according to the start timestamp and the end timestamp;

[0070] During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained;

[0071] All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

[0072] It is understandable that recorded videos usually include multiple video clips, which may be generated due to different recording periods, scene changes, event triggers or manual segmentation. When processing such recorded videos, it is necessary to obtain specific video clips based on the start timestamp and the end timestamp.

[0073] The Start Timestamp defines the exact time point when the video clip starts, and the End Timestamp defines the exact time point when the video clip ends. Loading the video file to be processed into the video processing system usually involves reading the metadata of the video file, including frame rate, resolution, duration, etc.

[0074] It should be noted that traversing from the beginning to the end of the video according to the video playback order is usually achieved by reading the video content frame by frame. During the traversal process, check whether the timestamp of the current frame is equal to or exceeds the starting timestamp. Once such a frame is found, its position is recorded (using frame number or timestamp) and marked as the starting point of the starting video segment.

[0075] If the start timestamp is not precisely aligned to the beginning of a certain frame, it is necessary to determine the start point by interpolation or selecting the closest frame.

[0076] Similarly, continue to traverse the video until you find a frame with a timestamp equal to or greater than the end timestamp, record the position of this frame, and mark it as the end point of the end recording video segment. If the end timestamp is not exactly aligned to the end of a frame, you need to determine the end point by interpolation or selecting the closest frame.

[0077] According to the determined starting point and ending point, all video segments between the starting video segment and the ending video segment are extracted from the video to be processed.

[0078] S3: extracting the first key frame of each of the recorded video segments according to the target feature to obtain a first key frame sequence;

[0079] Preferably, in an embodiment of the present application, extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence includes:

[0080] Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature;

[0081] The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

[0082] In this step, before starting to extract key frames, it is necessary to clarify what the target features are. For example, in a surveillance video, the target feature may be a human figure (for detecting human activities), a license plate number (for vehicle tracking), or a specific color (for detecting specific objects). According to the target features, select a suitable detection algorithm. These algorithms can be based on computer vision, such as target detection, object recognition, motion detection, etc., which are not described in detail in this application. During the analysis process, the algorithm checks whether the target features exist in each frame of the image and records its location, size, shape and other attributes.

[0083] Specifically, the first key frames extracted from each recorded video clip are integrated according to their time sequence in the original video, which will form a new video sequence, namely the first key frame sequence, in which each key frame represents the target feature at a certain moment in the original video.

[0084] S4: Based on the video splicing technology, each of the recorded video clips is integrated to obtain a target video, and according to the target features, a second key frame sequence of the target video is extracted;

[0085] Preferably, in an embodiment of the present application, extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence includes:

[0086] Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature;

[0087] The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

[0088] Among them, video stitching technology is usually used to combine multiple video clips into a coherent video. These clips may come from different cameras, different time periods or different scenes.

[0089] It is understandable that before video stitching, each recorded video clip is preprocessed to ensure coherence and consistency during stitching, and the preprocessing process includes adjusting the frame rate, resolution, color balance, etc. The preprocessed video clips are stitched together in chronological order or other logical order (such as scene changes) to generate a target video.

[0090] Specifically, a preset detection algorithm is applied to analyze the target video to identify key frames containing target features. Unlike the extraction of the first key frame (performed from each individual video clip), the extraction of the second key frame is performed on the integrated target video. The detection algorithm traverses each frame of the target video and checks whether it contains the target feature. If it does, the frame is marked as a key frame. All frames marked as key frames are integrated into a sequence according to their time order in the target video, namely the second key frame sequence.

[0091] S5: Based on image processing technology, compare the first key frame sequence and the second key frame sequence to obtain a target key frame;

[0092] Preferably, in an embodiment of the present application, the comparing the first key frame sequence and the second key frame sequence based on the image processing technology to obtain the target key frame includes:

[0093] Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features;

[0094] Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair;

[0095] Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair;

[0096] A target key frame is obtained based on the similarity.

[0097] Specifically, a feature extraction method in image processing technology is used to process each key frame respectively, extract its unique image features, and obtain two sets of feature sets, which respectively represent the first key frame features and the second key frame features.

[0098] Among them, the feature extraction method may include but is not limited to edge detection, corner detection, spot detection, histogram statistics, texture analysis, etc. For each key frame, multiple types of features can be extracted to fully describe its image content.

[0099] In this step, the feature matching algorithm is applied to compare the two sets of features and find similar feature pairs. A set of matching feature pairs is obtained, each of which contains similar features that appear in both the first key frame and the second key frame. Among them, feature matching algorithms include Brute Force Matching, FLANN (Fast Library for Approximate Nearest Ne i ghbors) matching, etc. In the matching process, a threshold can be set to control the strictness of the matching to avoid false matching.

[0100] Specifically, based on the matching feature pair, the similarity between the first key frame feature and the second key frame feature is calculated, and the similarity can be represented by a similarity score or a similarity matrix, which is used to represent the overall similarity between the two sets of features. The similarity calculation may be based on a distance metric between features, such as Euclidean distance, Manhattan distance, cosine similarity, etc. In one embodiment of the present invention, the similarity can also be calculated for each matching feature pair, and the average or weighted average is taken as the overall similarity, and by setting a similarity threshold, the key frames whose similarity exceeds the threshold are regarded as target key frames.

[0101] S6: Performing video reconstruction processing with the target key frame to obtain a video to be converted, and acquiring image data converted from the video to be converted.

[0102] In this step, the target key frames need to be sorted according to timestamps or frame numbers to ensure that they are played in the correct order in the reconstructed video.

[0103] In one embodiment of the present invention, if the time interval between target key frames is large, in order to generate a smoother video, an image interpolation technique may be used to generate intermediate frames, which usually involves image processing algorithms such as inter-frame difference estimation, motion estimation and compensation, etc., to generate transition frames between two key frames.

[0104] The sorted key frames (and intermediate frames that may be generated by interpolation) are combined in time order and encoded using an appropriate video encoding format (such as H.264, H.265, etc.) to generate a complete video file, i.e., the video to be converted.

[0105] The image data can be extracted frame by frame from the video to be converted by reading each frame of the video file and saving it as an independent image file (such as JPEG, PNG, etc.) or storing it in memory. After the image data is extracted, further image processing can be performed on it, such as image enhancement, denoising, color correction, etc., to meet specific application requirements.

[0106] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0107] 1) The present invention responds to video processing tasks and obtains the target features and time range intervals of the recorded video to be processed. The method can quickly locate and process the required video content, avoiding unnecessary full video scanning, thereby significantly improving processing efficiency.

[0108] 2) The present invention extracts the first key frame of each recorded video clip according to the target features, and the second key frame sequence of the integrated target video, ensuring the accuracy and completeness of the key information, which is crucial for subsequent video comparison, analysis and reconstruction.

[0109] 3) The present invention performs video reconstruction based on the target key frame, ensuring the accuracy and coherence of the reconstructed video. This reconstruction not only retains important information in the original video, but also improves the quality of the video through optimization and adjustment.

[0110] An embodiment of the present invention provides a method for processing recorded video. For details, see Figure 2 , Figure 2 The structure diagram of the apparatus for processing recorded video in one embodiment of the present invention is shown, comprising: an acquisition module 11, a preprocessing module 12, an extraction module 13, a splicing module 14, a comparison module 15, and a conversion module 16, wherein:

[0111] An acquisition module 11 is used to acquire target features and a time range of a video to be processed in response to a video processing task;

[0112] A preprocessing module 12, configured to obtain a plurality of recorded video clips according to a start timestamp and an end timestamp of the time range interval;

[0113] An extraction module 13 is used to extract the first key frame of each of the recorded video segments according to the target feature to obtain a first key frame sequence;

[0114] A splicing module 14 is used to integrate each of the recorded video clips based on the video splicing technology to obtain a target video, and extract a second key frame sequence of the target video according to the target feature;

[0115] A comparison module 15 is used to compare the first key frame sequence and the second key frame sequence based on image processing technology to obtain a target key frame;

[0116] The conversion module 16 is used to perform video reconstruction processing with the target key frame to obtain a video to be converted, and obtain image data converted from the video to be converted.

[0117] Preferably, in one embodiment of the present application, the preprocessing module is specifically used for:

[0118] Traversing the recorded video to be processed according to the start timestamp and the end timestamp;

[0119] During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained;

[0120] All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

[0121] Preferably, in one embodiment of the present application, the extraction module is specifically used for:

[0122] Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature;

[0123] The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

[0124] Preferably, in one embodiment of the present application, the splicing module is specifically used for:

[0125] Performing time alignment according to the first key frame of each of the recorded video segments to obtain a sequence of recorded video segments;

[0126] Extracting features from overlapping areas of adjacent video segments in the recorded video segment sequence to obtain adjacent corner points, and constructing a transformation matrix based on the adjacent corner points;

[0127] The overlapping areas are fused according to the transformation matrix to obtain a target video.

[0128] Preferably, in one embodiment of the present application, the comparison module is specifically used for:

[0129] Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features;

[0130] Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair;

[0131] Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair;

[0132] A target key frame is obtained based on the similarity.

[0133] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for processing recorded video, characterized in that: include: In response to the video processing task, obtaining target features and time range intervals of the recorded video to be processed; Obtaining multiple recorded video clips according to the start timestamp and the end timestamp of the time range interval; Extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence; Based on the video stitching technology, each of the recorded video clips is integrated to obtain a target video, and according to the target features, a second key frame sequence of the target video is extracted; Based on image processing technology, the first key frame sequence and the second key frame sequence are compared to obtain a target key frame; The target key frame is used to perform video reconstruction processing to obtain a video to be converted, and image data converted from the video to be converted is obtained.

2. The method for processing recorded video according to claim 1, characterized in that: The step of obtaining a plurality of recorded video clips according to the start timestamp and the end timestamp of the time range interval includes: Traversing the recorded video to be processed according to the start timestamp and the end timestamp; During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained; All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

3. The method for processing recorded video according to claim 1, characterized in that: The step of extracting the first key frame of each of the recorded video segments according to the target features to obtain a first key frame sequence includes: Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature; The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

4. The method for processing recorded video according to claim 1, characterized in that: The video splicing technology is used to integrate each of the recorded video clips to obtain a target video, including: Performing time alignment according to the first key frame of each of the recorded video segments to obtain a sequence of recorded video segments; Extracting features from overlapping areas of adjacent video segments in the recorded video segment sequence to obtain adjacent corner points, and constructing a transformation matrix based on the adjacent corner points; The overlapping areas are fused according to the transformation matrix to obtain a target video.

5. The method for processing recorded video according to claim 1, characterized in that: The comparing the first key frame sequence and the second key frame sequence based on the image processing technology to obtain the target key frame includes: Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features; Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair; Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair; A target key frame is obtained based on the similarity.

6. A video processing device, characterized in that: include: An acquisition module, used for acquiring target features and time range intervals of the recorded video to be processed in response to the video processing task; A preprocessing module, used for obtaining a plurality of recorded video clips according to a start timestamp and an end timestamp of the time range interval; An extraction module, used for extracting the first key frame of each of the recorded video segments according to the target feature to obtain a first key frame sequence; A splicing module, for integrating each of the recorded video clips based on a video splicing technology to obtain a target video, and extracting a second key frame sequence of the target video according to the target features; A comparison module, configured to compare the first key frame sequence and the second key frame sequence based on image processing technology to obtain a target key frame; The conversion module is used to perform video reconstruction processing with the target key frame to obtain the video to be converted, and obtain image data converted from the video to be converted.

7. The recorded video processing device according to claim 6, characterized in that: The preprocessing module is specifically used for: Traversing the recorded video to be processed according to the start timestamp and the end timestamp; During the traversal process, a start video clip including the start timestamp and an end video clip including the end timestamp are obtained; All recorded video segments between the start recorded video segment and the end recorded video segment are extracted as the recorded video segments.

8. The recorded video processing device according to claim 6, characterized in that: The extraction module is specifically used for: Analyzing each of the recorded video clips using a preset detection algorithm to obtain a first key frame including the target feature; The first key frames obtained from each of the recorded video segments are integrated into a first key frame sequence in chronological order.

9. The recorded video processing device according to claim 6, characterized in that: The splicing module is specifically used for: Performing time alignment according to the first key frame of each of the recorded video segments to obtain a sequence of recorded video segments; Extracting features from overlapping areas of adjacent video segments in the recorded video segment sequence to obtain adjacent corner points, and constructing a transformation matrix based on the adjacent corner points; The overlapping areas are fused according to the transformation matrix to obtain a target video.

10. The recorded video processing device according to claim 6, characterized in that: The comparison module is specifically used for: Extracting features from the first key frame sequence and the second key frame sequence respectively to obtain first key frame features and second key frame features; Based on a feature matching algorithm, feature matching is performed on the first key frame feature and the second key frame feature to obtain a matching feature pair; Calculating the similarity between the first key frame feature and the second key frame feature according to the matching feature pair; A target key frame is obtained based on the similarity.

Images