An AI-based video classification method

Abstract

The application discloses an AI-based video classification method, and relates to the technical field of image recognition, and comprises the following steps: S0, in the video frame sequence disassembled frame by frame, a brightness chroma threshold acquisition module is arranged in an image input layer and a preprocessing channel, is used for extracting pixel-level brightness and chroma features, a time sequence consistency acquisition module is embedded in the middle layer of a frame sequence analysis channel, is used for capturing the time consistency of brightness and texture changes between adjacent frames, a time domain spectrum acquisition module is arranged in a video signal frequency analysis channel after feature extraction, is used for detecting interframe frequency oscillation and optical flow energy distribution, a morphological geometric template acquisition module is arranged in a high-level semantic modeling layer and a deconstruction layer, is used for identifying geometric boundaries, contours and morphological structure features in the video frame, S1, each acquisition module works and detects highlight saturation spot phenomena, and the application has the characteristics of targeted adjustment.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, specifically to an AI-based video classification method. Background Technology

[0002] In video classification based on AI image recognition, the mainstream approach is to break the video down into frames, extract features from each frame using a 2D CNN model, then use an RNN model to capture the temporal dynamics of the frame sequence, and finally classify the video after temporal fusion. When raindrops, snowflakes, lens smudges, or other objects suddenly enter the view of a video image, they can produce various types of bright saturation spots under light. Convolutional kernels and edge operators in image processing algorithms are most sensitive to these high-frequency changes, which may lead to abnormal amplification of feature maps and affect the image recognition results.

[0003] High-brightness saturation spots can be categorized into several types: scene optical interaction, lens optical system, post-decoding artifacts, and temporal motion-induced artifacts. Several methods can be used for high-brightness saturation spot detection, including brightness and chromaticity threshold judgment, temporal consistency judgment, temporal spectrum judgment, and morphological geometric template judgment. Each method employs multiple video image acquisition and analysis modules that work collaboratively to improve analysis efficiency and accuracy.

[0004] Existing technologies often employ a single method for detecting bright saturated spots, corresponding to one or more types of bright saturated spots. Relying on a single method is susceptible to data noise, leading to misjudgments. Directly using information synthesis results in inconsistencies because different detection methods have significantly different suitable frame acquisition frequencies and pixel spatial granularities for detecting different types of bright saturated spots. Furthermore, the lack of a unified standard for data feature extraction and modeling leads to compromised data consistency after normalization. Without targeted adjustments, this can deviate from normal feature extraction requirements. Therefore, designing a targeted, AI-based video classification method is essential. Summary of the Invention

[0005] The purpose of this invention is to provide an AI-based video classification method to solve the problems mentioned in the background section.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: an AI-based video classification method, comprising the following steps:

[0007] S0. In the video frame sequence that is disassembled frame by frame, the luminance and chrominance threshold acquisition module is deployed in the image input layer and the preprocessing channel to extract pixel-level luminance and chrominance features. The temporal consistency acquisition module is embedded in the middle layer of the frame sequence analysis channel to capture the temporal consistency of luminance and texture changes between adjacent frames. The temporal spectrum acquisition module is deployed in the video signal frequency analysis channel after feature extraction to detect inter-frame frequency oscillations and optical flow energy distribution. The morphological geometry template acquisition module is set in the high-level semantic modeling layer and the deconstruction layer to identify the geometric boundaries, contours and morphological structural features in the video frames.

[0008] S1. Each acquisition module works and detects the phenomenon of bright saturation spots. Each acquisition module outputs the detection results with the initial frame acquisition frequency and pixel spatial granularity, and the corresponding analysis module detects whether the phenomenon of bright saturation spots has occurred.

[0009] S2. When a certain acquisition module detects a certain type of bright saturation spot, other acquisition modules most related to the feature domain of that acquisition module are instructed to adjust their frame acquisition frequency and pixel space granularity, so that the acquisition modules that need to be adjusted are consistent with the current acquisition module in terms of frame acquisition frequency and pixel space granularity.

[0010] S3. As the frequency continues, the frame acquisition frequency and pixel spatial granularity of the acquisition module that needs to be adjusted are dynamically adjusted based on whether the same bright saturation spot phenomenon is detected again.

[0011] S4. When the frame acquisition frequency and pixel spatial granularity of each acquisition module are not uniform, the data normalization module is used to process the detection data and perform frame acquisition frequency and pixel spatial granularity normalization.

[0012] S5. Based on the frame-by-frame structural analysis results, the spatial morphological features of the video frames are extracted and judged using morphological and geometric template methods.

[0013] According to the above technical solution, in step S0, it is specified which detection method corresponds to each type of high-brightness saturated spot:

[0014] S0-1, High-brightness saturated spots in scene optical interaction, which are caused by reflection and refraction of the object surface and direct illumination from external light sources. Brightness and chromaticity threshold detection is the primary judgment method, while temporal consistency detection and temporal spectrum detection are auxiliary judgment methods.

[0015] S0-2, High-brightness saturation spots in the lens optical system, which are formed by lens stains, oil film, halo, ghosting and flare, are judged primarily by temporal consistency detection, with brightness and chromaticity threshold detection and morphological geometric template detection as auxiliary judgment methods.

[0016] S0-3. Decoding post-processing artifact-type bright saturation spots, namely pseudo-bright spots generated by compression ringing, block effect and oversharpening, uses time-domain spectrum detection as the main judgment method and brightness and chromaticity threshold detection as the auxiliary judgment method.

[0017] S0-4. Temporal motion-induced high-brightness saturation spots, i.e. instantaneous high-brightness areas caused by raindrops, snowdrops, stroboscopic flashes, and rapid movement, are judged primarily by temporal consistency detection, with temporal spectrum detection and brightness and chromaticity threshold detection as auxiliary judgment methods.

[0018] According to the above technical solution, in step S2, the unification of frame acquisition frequency and pixel spatial granularity specifically involves:

[0019] S2-1. Each acquisition module, by default, detects the type of bright saturated spot corresponding to its primary judgment method, and uses its default frame acquisition frequency. and pixel space granularity Output the detection results, where This refers to the number of data acquisition modules;

[0020] S2-2, Order No. The acquisition module detected a bright saturation spot phenomenon, and its default frame acquisition frequency was [missing information]. The default pixel space granularity is The initial probability of this type of bright saturation spot phenomenon occurring is: At this point, it is necessary to... The frame acquisition frequency and pixel spatial granularity of each acquisition module were adjusted. The default frame acquisition frequency and pixel spatial granularity before the adjustment were as follows: and This makes its adjusted frame acquisition frequency Adjusted pixel space granularity .

[0021] According to the above technical solution, the dynamic adjustment in S3 specifically includes:

[0022] S3-1, Order No. Each acquisition module maintains a frame acquisition frequency. and pixel space granularity continued During the time period, if Within the time period When the acquisition module detects the same type of bright saturation spot phenomenon again, observe the first... If each acquisition module simultaneously detects the same type of bright saturation spot phenomenon, and if so, the probability of this type of bright saturation spot phenomenon occurring is increased, with the adjusted probability being... ,in For the first When the first acquisition module is used as an auxiliary judgment method, it is related to the first... The probability increment caused by the detection of the same type of bright saturation spot phenomenon by each acquisition module, if the first acquisition module detects the same type of bright saturation spot phenomenon, If none of the acquisition modules simultaneously detected the same type of bright saturation spot phenomenon, then... ;

[0023] S3-2, If in Within the time period When the first acquisition module does not detect the same type of bright saturation spot phenomenon, the second... The frame acquisition frequency and pixel spatial granularity of each acquisition module are determined by... and To its initial value and Gradually recovering, making the current distance The elapsed time at the end of the time period is Then the frame acquisition frequency at this time and pixel space granularity The calculation formulas are as follows: when hour, ,when hour, ,when hour, ,when hour, ,in , This is the frequency conversion factor.

[0024] According to the above technical solution, in step S4, the normalization processing of frame acquisition frequency and pixel spatial granularity specifically involves:

[0025] S4-1, Firstly in When the time period ends, the first The and the first The first acquisition frequency of each acquisition module is aligned, at which point both acquisition modules acquire data simultaneously. The next acquisition will only collect data that meets the specified frequency. and The detection data is collected only at the least common multiple of the time; other data are not collected.

[0026] S4-2, the first The and the first The maximum and minimum values ​​of the detection data from each acquisition module are mapped to... Within the specified range, the dimensional differences in pixel spatial granularity between different acquisition modules are eliminated.

[0027] According to the above technical solution, in S5, the determination of the type of bright saturated spot is specifically as follows: before each detection of a bright saturated spot phenomenon by each acquisition module, the geometric feature distribution of this type of bright spot detected by the morphological geometric template acquisition module is respectively... ,in The geometric feature distribution refers to the number of geometric morphological types involved in the high-brightness saturation spots. It indicates the intensity of the template's response in ring-shaped, strip-shaped, dot-shaped, and radial morphologies. When a certain type of high-brightness saturation spot phenomenon is detected, if this type of high-brightness saturation spot leads to... Changes have occurred, and Actual detection If an incremental change occurs, the probability of this type of bright saturation spot occurring will increase. ,in This is the geometric feature increment influence coefficient; if no detection is found... Incremental changes occur Combined with S3-1 The final probability is calculated when hour, The probability threshold is used to determine if this type of bright saturated spot is valid.

[0028] An AI-based video classification system includes an image recognition module, a multi-information integration module, and a high-brightness saturation spot judgment module. The image recognition module is used to collect brightness and chromaticity thresholds, temporal consistency, and temporal spectrum, and perform multi-dimensional detection in combination with morphological geometric templates. The multi-information integration module is used to perform unified frequency benchmark and granularity normalization processing on frame acquisition frequency and pixel spatial granularity, dynamically adjust the characteristics of the two output parameters according to subsequent detection results, and perform multi-source information integration processing. The high-brightness saturation spot judgment module is used to determine the type of high-brightness saturation spot based on the processed data.

[0029] According to the above technical solution, the image recognition module includes a luminance and chromaticity threshold acquisition module, a temporal consistency acquisition module, a temporal spectrum acquisition module, a morphological geometric template acquisition module, a luminance and chromaticity threshold analysis module, a temporal consistency analysis module, a temporal spectrum analysis module, and a morphological geometric template analysis module. The luminance and chromaticity threshold acquisition module, the temporal consistency acquisition module, and the temporal spectrum acquisition module are respectively used to acquire luminance and chromaticity thresholds, temporal consistency, and temporal spectrum. The morphological geometric template acquisition module is used to detect deconstructed morphological geometric templates in video images. The luminance and chromaticity threshold analysis module, the temporal consistency analysis module, and the temporal spectrum analysis module are respectively used to analyze the detection results of luminance and chromaticity thresholds, temporal consistency, and temporal spectrum. The morphological geometric template analysis module is used to analyze templates of various geometric shapes.

[0030] The multi-information integration module includes a frame acquisition frequency adjustment module, a pixel spatial granularity adjustment module, a data normalization module, a high-brightness saturation spot type correspondence module, and a time recording module. The time recording module is used to count the time when a certain type of high-brightness saturation spot reappears. The frame acquisition frequency adjustment module and the pixel spatial granularity adjustment module are used to adjust the frame acquisition frequency and pixel spatial granularity of each acquisition module, respectively. The data normalization module is used to normalize the frame acquisition frequency and pixel spatial granularity of each acquisition module.

[0031] The high-brightness saturation spot determination module includes a geometric morphology auxiliary acquisition module and a high-brightness saturation spot type determination module. The geometric morphology auxiliary acquisition module is used to analyze the morphological geometric template changes of the video image frame by frame to assist in the determination of the high-brightness saturation spot type. The high-brightness saturation spot type determination module is used to determine the high-brightness saturation spot type with the highest probability.

[0032] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: In order to address the interference of high-brightness saturated spots in videos on classification accuracy, the present invention proposes a closed-loop mechanism of multi-source classification detection and event triggering linkage: luminance and chromaticity thresholds, temporal consistency, temporal spectrum and morphological geometric templates are deployed at the input, temporal, spectral and morphological layers respectively, and the determination is jointly made according to the spot type-dominant + auxiliary strategy; once a suspicious event is detected in any channel, the relevant modules are linked to conduct collaborative observation within a short window.

[0033] By unifying heterogeneous methods to two controllable parameters—frame acquisition frequency and pixel spatial granularity—and performing adaptive alignment, dynamic adjustment, and alignment normalization, asynchronous and heteroscale distortion and statistical offset caused by direct fusion are avoided, thereby significantly reducing the false judgment rate and improving the temporal consistency and interpretability of video-level judgment. Attached Figure Description

[0034] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0035] Figure 1 This is a schematic diagram of the overall modular structure of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Please see Figure 1The present invention provides a technical solution: an AI-based video classification method, comprising the following steps:

[0038] S0. In the video frame sequence that is disassembled frame by frame, the luminance and chrominance threshold acquisition module is deployed in the image input layer and the preprocessing channel to extract pixel-level luminance and chrominance features. The temporal consistency acquisition module is embedded in the middle layer of the frame sequence analysis channel to capture the temporal consistency of luminance and texture changes between adjacent frames. The temporal spectrum acquisition module is deployed in the video signal frequency analysis channel after feature extraction to detect inter-frame frequency oscillations and optical flow energy distribution. The morphological geometry template acquisition module is set in the high-level semantic modeling layer and the deconstruction layer to identify the geometric boundaries, contours and morphological structural features in the video frames.

[0039] S1. Each acquisition module works and detects the phenomenon of bright saturation spots. Each acquisition module outputs the detection results with the initial frame acquisition frequency and pixel spatial granularity, and the corresponding analysis module detects whether the phenomenon of bright saturation spots has occurred.

[0040] S2. When a certain acquisition module detects a certain type of bright saturation spot, other acquisition modules most related to the feature domain of that acquisition module are instructed to adjust their frame acquisition frequency and pixel space granularity, so that the acquisition modules that need to be adjusted are consistent with the current acquisition module in terms of frame acquisition frequency and pixel space granularity.

[0041] S3. As the frequency continues, the frame acquisition frequency and pixel spatial granularity of the acquisition module that needs to be adjusted are dynamically adjusted based on whether the same bright saturation spot phenomenon is detected again.

[0042] S4. When the frame acquisition frequency and pixel spatial granularity of each acquisition module are not uniform, the data normalization module is used to process the detection data and perform frame acquisition frequency and pixel spatial granularity normalization.

[0043] S5. Based on the frame-by-frame structural analysis results, the spatial morphological features of the video frames are extracted and judged using morphological and geometric template methods.

[0044] To address the issue of inconsistent optimal observation granularity among different detection methods, this invention uses frame acquisition frequency and pixel spatial granularity as a unified parameter bundle. First, it unifies the parameters of the module most relevant to the current event upon triggering (S2). Then, it dynamically adjusts the parameters during continuous observation (S3). Finally, it achieves cross-module alignment through data normalization (S4). This process avoids normalization distortion and statistical bias caused by directly mixing asynchronous and heterogeneous scale data, ensuring the consistency and comparability of the fusion judgment.

[0045] Different detection methods for different speckle types (such as time-domain spectrum as the primary method for stroboscopic artifacts and time-series consistency as the primary method for lens system artifacts) should prioritize the evidence channel with the best physical / causal match during fusion, and then use other channels for verification to reduce the bias caused by strong correlation artifacts and improve the accuracy of classification.

[0046] In S0, it is specified which detection method corresponds to each type of high-brightness saturation spot:

[0047] S0-1, High-brightness saturated spots in scene optical interaction, which are caused by reflection and refraction of the object surface and direct illumination from external light sources. Brightness and chromaticity threshold detection is the primary judgment method, while temporal consistency detection and temporal spectrum detection are auxiliary judgment methods.

[0048] S0-2, High-brightness saturation spots in the lens optical system, which are formed by lens stains, oil film, halo, ghosting and flare, are judged primarily by temporal consistency detection, with brightness and chromaticity threshold detection and morphological geometric template detection as auxiliary judgment methods.

[0049] S0-3. Decoding post-processing artifact-type bright saturation spots, namely pseudo-bright spots generated by compression ringing, block effect and oversharpening, uses time-domain spectrum detection as the main judgment method and brightness and chromaticity threshold detection as the auxiliary judgment method.

[0050] S0-4, Temporal motion-induced high-brightness saturation spots, i.e. instantaneous high-brightness areas caused by raindrops, snowdrops, stroboscopic flashes and rapid movement, are judged primarily by temporal consistency detection, with temporal spectrum detection and brightness and chromaticity threshold detection as auxiliary judgment methods.

[0051] This invention deploys four types of acquisition and analysis modules (see steps S0, S0-1 to S0-4) in the input layer, temporal analysis layer, spectral analysis layer, and semantic structure layer, respectively, for luminance and chromaticity thresholding, temporal consistency, temporal spectrum, and morphological geometric templates. These modules are used to classify and detect high-brightness saturation spots caused by scene optical interactions, lens optical systems, post-decoding artifacts, and temporal motion-induced artifacts. Compared to single methods, cross-domain feature cross-verification effectively suppresses false positives caused by white backgrounds, oversharpening, flicker, and compression artifacts, significantly reducing the misclassification rate.

[0052] In S2, the unification of frame acquisition frequency and pixel spatial granularity is specifically as follows:

[0053] S2-1. Each acquisition module, by default, detects the type of bright saturated spot corresponding to its primary judgment method, and uses its default frame acquisition frequency. and pixel space granularity Output the detection results, where This refers to the number of data acquisition modules;

[0054] S2-2, Order No. The acquisition module detected a bright saturation spot phenomenon, and its default frame acquisition frequency was [missing information]. The default pixel space granularity is The initial probability of this type of bright saturation spot phenomenon occurring is: At this point, it is necessary to... The frame acquisition frequency and pixel spatial granularity of each acquisition module were adjusted. The default frame acquisition frequency and pixel spatial granularity before the adjustment were as follows: and This makes its adjusted frame acquisition frequency Adjusted pixel space granularity ;

[0055] In S3, the dynamic adjustment is specifically as follows:

[0056] S3-1, Order No. Each acquisition module maintains a frame acquisition frequency. and pixel space granularity continued During the time period, if Within the time period When the acquisition module detects the same type of bright saturation spot phenomenon again, observe the first... If each acquisition module simultaneously detects the same type of bright saturation spot phenomenon, and if so, the probability of this type of bright saturation spot phenomenon occurring is increased, with the adjusted probability being... ,in For the first When the first acquisition module is used as an auxiliary judgment method, it is related to the first... The probability increment caused by the detection of the same type of bright saturation spot phenomenon by each acquisition module, if the first acquisition module detects the same type of bright saturation spot phenomenon, If none of the acquisition modules simultaneously detected the same type of bright saturation spot phenomenon, then... ;

[0057] S3-2, If in Within the time period When the first acquisition module does not detect the same type of bright saturation spot phenomenon, the second... The frame acquisition frequency and pixel spatial granularity of each acquisition module are determined by... and To its initial value and Gradually recovering, making the current distance The elapsed time at the end of the time period is Then the frame acquisition frequency at this time and pixel space granularity The calculation formulas are as follows: when hour, ,when hour, ,when hour, ,when hour, ,in , For frequency conversion factors;

[0058] In S4, the normalization processing of frame acquisition frequency and pixel spatial granularity is specifically performed as follows:

[0059] S4-1, Firstly in When the time period ends, the first The and the first The first acquisition frequency of each acquisition module is aligned, at which point both acquisition modules acquire data simultaneously. The next acquisition will only collect data that meets the specified frequency. and The detection data is collected only at the least common multiple of the time; other data are not collected.

[0060] S4-2, the first The and the first The maximum and minimum values ​​of the detection data from each acquisition module are mapped to... Within the range, eliminate the dimensional differences in pixel spatial granularity between different acquisition modules;

[0061] In S5, the determination of the type of bright saturated spot is specifically as follows: before each detection of a bright saturated spot phenomenon by each acquisition module, the geometric feature distribution of this bright spot type detected by the morphological geometric template acquisition module is as follows: ,in The geometric feature distribution refers to the number of geometric morphological types involved in the high-brightness saturation spots. It indicates the intensity of the template's response in ring-shaped, strip-shaped, dot-shaped, and radial morphologies. When a certain type of high-brightness saturation spot phenomenon is detected, if this type of high-brightness saturation spot leads to... Changes have occurred, and Actual detection If an incremental change occurs, the probability of this type of bright saturation spot occurring will increase. ,in This is the geometric feature increment influence coefficient; if no detection is found... Incremental changes occur Combined with S3-1 The final probability is calculated when hour, The probability threshold is used to determine if this type of bright saturated spot is valid;

[0062] This method calculates the probability of occurrence of different types of high-brightness saturation spots by measuring changes in the geometric feature distribution of morphological templates, achieving quantitative and interpretable type determination. Instead of relying on single-frame brightness anomalies, this method captures the incremental responses of geometric features such as rings, stripes, dots, and radial patterns, effectively distinguishing different spot types caused by optical reflection, lens smudges, or motion artifacts. Compared to traditional threshold detection, this mechanism significantly reduces misjudgments caused by localized high-brightness false features and maintains temporal consistency and judgment stability in multi-frame fusion.

[0063] An AI-based video classification system includes an image recognition module, a multi-information integration module, and a high-brightness saturation spot judgment module. The image recognition module is used to collect brightness and chromaticity thresholds, temporal consistency, and temporal spectrum, and to perform multi-dimensional detection by combining morphological geometric templates. The multi-information integration module is used to perform unified frequency benchmark and granularity normalization processing on frame acquisition frequency and pixel spatial granularity, dynamically adjust the characteristics of the two output parameters according to the subsequent detection results, and perform multi-source information integration processing. The high-brightness saturation spot judgment module is used to determine the type of high-brightness saturation spot based on the processed data.

[0064] The image recognition module includes a luminance and chromaticity threshold acquisition module, a temporal consistency acquisition module, a temporal spectrum acquisition module, a morphological geometric template acquisition module, a luminance and chromaticity threshold analysis module, a temporal consistency analysis module, a temporal spectrum analysis module, and a morphological geometric template analysis module. The luminance and chromaticity threshold acquisition module, the temporal consistency acquisition module, and the temporal spectrum acquisition module are used to acquire luminance and chromaticity thresholds, temporal consistency, and temporal spectrum, respectively. The morphological geometric template acquisition module is used to detect deconstructed morphological geometric templates in video images. The luminance and chromaticity threshold analysis module, the temporal consistency analysis module, and the temporal spectrum analysis module are used to analyze the detection results of luminance and chromaticity thresholds, temporal consistency, and temporal spectrum, respectively. The morphological geometric template analysis module is used to analyze templates of various geometric shapes.

[0065] The multi-information integration module includes a frame acquisition frequency adjustment module, a pixel space granularity adjustment module, a data normalization module, a high-brightness saturation spot type correspondence module, and a time recording module. The time recording module is used to count the time when a certain type of high-brightness saturation spot reappears. The frame acquisition frequency adjustment module and the pixel space granularity adjustment module are used to adjust the frame acquisition frequency and pixel space granularity of each acquisition module, respectively. The data normalization module is used to normalize the frame acquisition frequency and pixel space granularity of each acquisition module.

[0066] The highlight saturation spot determination module includes a geometric morphology-assisted acquisition module and a highlight saturation spot type determination module. The geometric morphology-assisted acquisition module is used to analyze the morphological geometric template changes of the video image frame by frame to assist in the determination of the highlight saturation spot type. The highlight saturation spot type determination module is used to determine the highlight saturation spot type with the highest probability.

[0067] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another entity or operation, and do not necessarily imply any such actual relationship or order between these entities and operations. Furthermore, the terms "comprising," "including," and any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, and includes elements inherent to such a process, method, article, or apparatus.

[0068] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments and make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI-based video classification method, characterized in that: Comprise the following steps: S0, in the frame-by-frame disassembly video frame sequence, luminance chroma threshold acquisition module is arranged in the image input layer and the pretreatment channel, for extracting the pixel level luminance and chroma characteristics, the time sequence consistency acquisition module is embedded in the middle layer of the frame sequence analysis channel, for capturing the time consistency of the luminance and texture changes between adjacent frames, the time domain spectrum acquisition module is arranged in the frequency analysis channel of the video signal after feature extraction, for detecting the frequency oscillation and the light flow energy distribution between frames, the morphological geometric template acquisition module is set in the high-level semantic modeling layer and the deconstruction layer, for identifying the geometric boundary, contour and morphological structure characteristics in the video frame; S1, each acquisition module works and detects the highlight saturation spot phenomenon, each acquisition module outputs the detection result with the initial frame acquisition frequency and pixel space granularity, and detects whether the highlight saturation spot phenomenon occurs through the corresponding analysis module; S2, when a certain acquisition module detects that a certain type of highlight saturation spot occurs, the other type of acquisition module most related to the feature domain of the acquisition module is adjusted in frame acquisition frequency and pixel space granularity, so that the acquisition module to be adjusted and the current acquisition module are unified in frame acquisition frequency and pixel space granularity; S3, with the continuation of the frequency, according to whether the same highlight saturation spot phenomenon is continuously detected, the frame acquisition frequency and the pixel space granularity of the acquisition module to be adjusted are dynamically adjusted; S4, when the frame acquisition frequency and the pixel space granularity of each acquisition module are not unified, the detection data is processed by using the data normalization module, and the frame acquisition frequency and the pixel space granularity are normalized; S5, based on the frame-by-frame structure analysis result, the spatial morphological characteristics of the video frame are extracted and judged by using the morphological and geometric template method.

2. The AI-based video classification method of claim 1, wherein: In the S0, it is clear that which detection method corresponds to each type of highlight saturation spot, which is specifically: S0-1, scene optical interaction type highlight saturation spot, that is, the highlight saturation spot caused by object surface reflection, refraction and external light source direct incidence takes the luminance chroma threshold detection as the dominant judgment method, and the time sequence consistency detection and the time domain spectrum detection as the auxiliary judgment method; S0-2, lens optical system type highlight saturation spot, that is, the highlight saturation spot formed by lens stains, oil film, halo, ghosting and flare takes the time sequence consistency detection as the dominant judgment method, and the luminance chroma threshold detection and the morphological geometric template detection as the auxiliary judgment method; S0-3, decoding post-processing artifact type highlight saturation spot, that is, the pseudo highlight generated by compression ringing, block effect and over-sharpening takes the time domain spectrum detection as the dominant judgment method, and the luminance chroma threshold detection as the auxiliary judgment method; S0-4, time sequence motion induced type highlight saturation spot, that is, the transient highlight area caused by raindrops, snowflakes, stroboscopic light and rapid movement takes the time sequence consistency detection as the dominant judgment method, and the time domain spectrum detection and the luminance chroma threshold detection as the auxiliary judgment method.

3. The AI-based video classification method of claim 2, wherein: In the S2, the unification in the frame acquisition frequency and the pixel space granularity is specifically: S2-1, each acquisition module detects the highlight saturated spot type corresponding to the leading judgment mode by default, and acquires frames at its default frame acquisition frequency and pixel spatial granularity output the detection result, wherein is the number of acquisition modules; S2-2, Order No. The acquisition module detected a bright saturation spot phenomenon, and its default frame acquisition frequency was [missing information]. The default pixel space granularity is The initial probability of this type of bright saturation spot phenomenon occurring is: At this point, it is necessary to... The frame acquisition frequency and pixel spatial granularity of each acquisition module were adjusted. The default frame acquisition frequency and pixel spatial granularity before the adjustment were as follows: and This makes its adjusted frame acquisition frequency Adjusted pixel space granularity .

4. The AI-based video classification method of claim 3, wherein: In the S3, the dynamic adjustment is specifically: S3-1, Order No. Each acquisition module maintains a frame acquisition frequency. and pixel space granularity continued During the time period, if Within the time period When the acquisition module detects the same type of bright saturation spot phenomenon again, observe the first... If each acquisition module simultaneously detects the same type of bright saturation spot phenomenon, and if so, the probability of this type of bright saturation spot phenomenon occurring is increased, with the adjusted probability being... ,in For the first When the first acquisition module is used as an auxiliary judgment method, it is related to the first... The probability increment caused by the detection of the same type of bright saturation spot phenomenon by each acquisition module, if the first acquisition module detects the same type of bright saturation spot phenomenon, If none of the acquisition modules simultaneously detected the same type of bright saturation spot phenomenon, then... ; S3-2、if in the time period, the first acquisition module does not detect the same type of highlight saturation phenomenon, the frame acquisition frequency and the pixel spatial granularity of the first acquisition module are gradually restored to the initial values of the first acquisition module and the second acquisition module, respectively, and the distance between the first acquisition module and the second acquisition module is gradually reduced to the initial distance between the first acquisition module and the second acquisition module. wherein f1 and f2 are frequency conversion coefficients.​​​​​​​​​​​​​​​​​​​​ 5. The AI-based video classification method of claim 4, wherein: In the S4, the normalization processing of the frame acquisition frequency and the pixel space granularity is specifically: S4-1, Firstly in When the time period ends, the first The and the first The first acquisition frequency of each acquisition module is aligned, at which point both acquisition modules acquire data simultaneously. The next acquisition will only collect data that meets the specified frequency. and The detection data is collected only at the least common multiple of the time; other data are not collected. S4-2, the maximum and minimum of the detection data of the first and second acquisition modules are mapped into the interval of [0, 1], eliminating the dimensional difference of pixel space granularity between different acquisition modules. ​​​ 6. The AI-based video classification method of claim 5, wherein: The type of the highlight saturation spot is determined in S5, specifically, before each detection of the highlight saturation spot, the morphological geometric template acquisition module detects the geometric feature distribution of the type of the highlight saturation spot , wherein is the number of geometric morphologies involved in the highlight saturation spot, and the geometric feature distribution refers to the response intensity of the template on the annular, strip, point and radial morphologies. When a certain type of highlight saturation spot is detected, if the type of the highlight saturation spot causes to change, and actually detects to produce an incremental change, then the probability of the type of the highlight saturation spot occurring is , wherein is a geometric feature incremental influence coefficient, if no incremental change is detected , the final probability of is calculated in combination with S3-1, when , is a probability judgment threshold, and the type of the highlight saturation spot is determined to be correct.

7. An AI-based video classification system based on the method of any one of claims 1-6, characterized in that: The image recognition module is used for collecting luminance chroma threshold, time sequence consistency and time domain spectrum, and combining morphological geometric templates for multi-dimensional detection.

8. The AI-based video classification system of claim 7, wherein: The image recognition module includes luminance chroma threshold collection module, time sequence consistency collection module, time domain spectrum collection module, morphological geometric template collection module, luminance chroma threshold analysis module, time sequence consistency analysis module, time domain spectrum analysis module and morphological geometric template analysis module. The multi-information comprehensive module includes frame collection frequency adjustment module, pixel space granularity adjustment module, data normalization module, high-light saturated spot type corresponding module and time recording module. The high-light saturated spot judgment module includes geometric morphology auxiliary collection module and high-light saturated spot type judgment module.

Images