An advertising data transmission method and system based on a cloud platform

Abstract

The present invention relates to the field of data transmission, and in particular to an advertisement data transmission method and system based on a cloud platform. The method includes the steps of: acquiring a merchant advertisement video; performing segmentation processing on each frame image of the merchant advertisement video to obtain foreground pixels and background pixels; denoting the sequence formed by all background pixels at any position in all frame images as a background pixel sequence; dividing the background pixel sequence into several background segments with different distributions; constructing a Gaussian distribution model for the background segments, calculating the number of intervals, dividing the Gaussian distribution model into the number of intervals, using the mean value of the interval to which the background pixel belongs as the updated value of the background pixel, and performing encoding processing on the updated values of the background pixels in the background segment by using a run-length encoding algorithm; encoding the foreground pixels; so as to realize the transmission of the merchant advertisement video. By reducing the value difference of the pixels, the compression effect of the run-length encoding is improved.

Description

Technical Field

[0001] The present invention relates to the field of data transmission, and in particular, to an advertisement data transmission method and system based on a cloud platform. Background Art

[0002] In order to improve the accuracy of advertisement placement, optimize the advertisement placement efficiency, and reduce the advertisement cost, enterprises often transmit advertisements to a cloud platform for management. Since there is a large amount of data such as videos, images, and audios in advertisement data, the transmission volume of advertisement data is large, which will result in a low transmission efficiency of advertisement data. In order to improve the transmission efficiency, it is necessary to compress the advertisement data.

[0003] As a compression coding algorithm, the run-length coding algorithm encodes and compresses data by recording the starting position and the continuous length of consecutive identical data. Therefore, the run-length coding algorithm has a strong compression ability for data with consecutive identical values. Since any pixel in a video has relatively similar values in the frame sequence, the run-length coding algorithm can be used to compress the video. Although the values of pixels in a video are relatively similar in the frame sequence, they are not exactly the same, and the diversity of their values will result in a poor compression effect of the run-length coding. How to improve the compression effect by reducing the diversity of pixel values in a video has become the research focus of the present invention.

[0004] The patent application document with the publication number CN103916448A discloses a data transmission method, system, and corresponding cloud transmission platform based on a cloud transmission platform. The method in the patent application document slices the data according to the location information of the transmission nodes to improve the data transmission efficiency. The method in the patent application document does not involve the content of run-length coding compression, so the method in the patent application document cannot solve the technical problems of the present solution. Summary of the Invention

[0005] In order to solve the problem of how to improve the compression effect by reducing the diversity of pixel values in a video, the present invention provides an advertisement data transmission method and system based on a cloud platform.

[0006] In a first aspect, the present invention provides an advertisement data transmission method based on a cloud platform, adopting the following technical solution:

[0007] An advertisement data transmission method based on a cloud platform includes the steps of:

[0008] Obtain a merchant advertisement video;

[0009] Perform segmentation processing on each frame image of the merchant advertisement video to obtain foreground pixels and background pixels;

[0010] Denote the sequence formed by all background pixels at any position in all frame images as the background pixel sequence; divide the background pixel sequence into several background segments with different distributions; construct a Gaussian distribution model for the background segments, and calculate the number of intervals: , represents the information reproduction rate within the preset area of the background pixels at this position, represents the information content within the preset area of the background pixels at this position, represents the peak value of the Gaussian distribution model, represents the mean of the differences between all two adjacent pixels in the background segment, represents the number of data value types in the background segment, norm() represents the linear normalization method, S1 represents the number of intervals, represents the ceiling symbol, divide the Gaussian distribution model into S1 intervals, use the mean value of the interval to which the background pixel belongs as the updated value of the background pixel, and encode the updated values of the background pixels in the background segment using the run - length encoding algorithm; encode the foreground pixels;

[0011] to achieve the transmission of merchant advertising videos.

[0012] The present invention analyzes the values of background pixels and foreground pixels in the frame sequence separately, effectively reducing the value differences of pixels in the frame sequence, providing a basis for improving the compression effect of run - length encoding in the follow - up; further, segment the values of pixels in the frame sequence according to the distribution characteristics, so that the pixel values in each segment all reflect the same object, further reducing the value differences and improving the compression effect of run - length encoding; further, further divide the segmented results of the values of pixels in the frame sequence into intervals, and replace each true value with the mean value of the interval, effectively eliminating the value differentiation and further improving the effect of run - length encoding; further, when dividing the intervals of the segmented results, consider information such as information importance and value differences, improving the accuracy of interval division, and providing a basis for improving the compression effect of run - length encoding in the follow - up.

[0013] Preferably, the segmentation of each frame image of the merchant advertising video to obtain foreground pixels and background pixels includes:

[0014] Process the merchant advertising video using the background subtraction method to obtain the foreground area and background area of each frame image;

[0015] Obtain the circumscribed rectangle of the foreground area with the largest area in all frame foreground areas as the reference rectangle;

[0016] Take the geometric center of the foreground area in each frame image as the center of the reference rectangle, and obtain the pixels within the reference rectangle, denoted as foreground pixels, and take the pixels in each frame image except the foreground pixels as background pixels.

[0017] The present invention efficiently and accurately segments the moving area and the background area through the background subtraction method, providing a data basis for improving the compression effect of run-length encoding.

[0018] Preferably, the segmentation of the background pixel sequence into several background segments with different distributions includes:

[0019] Taking any background pixel in the background pixel sequence as the center, obtaining a preset number of pixels as the reference pixels of this background pixel, fitting a Gaussian distribution model using the reference pixels of this background pixel, taking the difference between the Gaussian distribution models of every two adjacent background pixels as the clustering difference, performing clustering processing on all background pixels to obtain several categories, and taking the pixel segments formed by the background pixels in each category as the background segments.

[0020] Based on the feature that the pixel value distributions of the same object are similar, the present invention separates the corresponding value data of different objects, providing a basis for subsequent de-differentiation.

[0021] Preferably, the method for obtaining the difference between the Gaussian distribution models includes:

[0022] Multiplying the absolute value of the difference between the means of the two Gaussian distribution models by the absolute value of the difference between the variances of the two Gaussian distribution models to obtain the difference between the two Gaussian distribution models.

[0023] The present invention reflects the model difference through the variance and the mean difference. This calculation method is relatively simple and has higher implementation efficiency.

[0024] Preferably, the method for obtaining the information reproduction rate includes:

[0025] Taking the background pixels at this position in each frame of image as the center, obtaining a preset area, denoted as the local area of the background pixels at this position, calculating the similarity between the local areas of the background pixels at this position and the pixels at other positions, and taking the mean of the similarities between the local areas of the background pixels at this position and the local areas at all other positions as the information reproduction rate of the background pixels at this position.

[0026] The present invention reflects the repeated occurrence of information through similarity, and more accurately measures the importance of information.

[0027] Preferably, the method for obtaining the information content includes:

[0028] Taking the information entropy within the local area of the background pixels at this position as the information content.

[0029] Preferably, the division of the Gaussian distribution model into S1 intervals includes:

[0030] Evenly dividing the Gaussian distribution model into S1 regions with the same area, and obtaining the pixel value interval corresponding to each region.

[0031] Preferably, encoding the foreground pixels includes:

[0032] Denote the area composed of foreground pixels in each frame of image as the analyzed foreground area;

[0033] Align all frames of the analyzed foreground areas, and denote the sequence composed of all foreground pixels at any position in all frames of the analyzed foreground areas as the foreground pixel sequence;

[0034] Divide the foreground pixel sequence into several foreground segments with different distributions;

[0035] Encode the foreground segments.

[0036] Preferably, achieving the transmission of the merchant advertisement video includes:

[0037] Concatenate the encoding sequences of all background segments of the background pixel sequence to obtain the encoding sequence of the background pixel sequence, concatenate the encoding sequences of all foreground segments of the foreground pixel sequence to obtain the encoding sequence of the foreground pixel sequence, obtain the geometric center coordinates of the analyzed foreground areas, concatenate the geometric center coordinates of all analyzed foreground areas to obtain the trajectory sequence, obtain the starting frame numbers of the analyzed foreground areas, insert the starting frame numbers, the length and width of the reference rectangle before the trajectory sequence, insert a preset demarcation flag value after concatenating the trajectory sequence with the inserted starting frame numbers and the encoding sequences of all foreground pixel sequences, then concatenate with the encoding sequences of all background pixel sequences to obtain the transmission sequence, and perform transmission processing on the transmission sequence.

[0038] In a second aspect, the present invention provides an advertisement data transmission system based on a cloud platform, adopting the following technical solution:

[0039] An advertisement data transmission system based on a cloud platform includes: a processor and a memory, and the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned advertisement data transmission method based on a cloud platform is implemented.

[0040] By adopting the above technical solution, generate a computer program for the above-mentioned advertisement data transmission method based on a cloud platform, and store it in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0041] The present invention has the following technical effects:

[0042] The present invention analyzes the value of background pixels in the frame sequence and the value of foreground pixels in the frame sequence separately, effectively reducing the value difference of pixels in the frame sequence, and providing a basis for improving the compression effect of run-length encoding in the follow-up;

[0043] Further, the pixel values in the frame sequence are segmented according to the distribution characteristics, so that the pixel values of each segment reflect the same object, further reducing the value differences and improving the compression effect of the run-length encoding;

[0044] Further, the segmented results of the pixel values in the frame sequence are further divided into intervals, and the mean value of the interval is used to replace each real value, effectively eliminating the value differences and further improving the effect of the run-length encoding;

[0045] Further, when dividing the intervals of the segmented results, information such as information importance and value differences is considered, improving the accuracy of the interval division and providing a basis for further improving the compression effect of the run-length encoding. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0047] Figure 1 is a flowchart of the method in a method for transmitting advertisement data based on a cloud platform according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] Hereinafter, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0049] It should be understood that when the claims, specifications and drawings of the present invention use terms such as "first", "second", etc., they are only used to distinguish different objects and not to describe a specific order. The terms "including" and "comprising" used in the specifications and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0050] An embodiment of the present invention discloses a method for transmitting advertisement data based on a cloud platform, referring to Figure 1 , including steps S1 - S5:

[0051] S1: Obtain the merchant advertisement video.

[0052] Specifically, obtain the merchant advertisement video.

[0053] S2: Segment each frame image of the merchant advertisement video to obtain foreground pixels and background pixels.

[0054] It should be noted that there are several moving objects in the merchant advertisement video. For the same moving object, only its position changes in each frame, and its color value changes little. Therefore, the pixel values of the same moving object in consecutive frames are similar. The color of non-moving objects also changes little in consecutive frames. Therefore, in order to improve the compression effect, moving objects and non-moving objects need to be considered separately. First, separate the moving objects and non-moving objects in each frame image of the merchant advertisement video. In this implementation, the foreground pixels reflect the moving objects, and the background pixels reflect the non-moving objects.

[0055] Preferably, as an example, segmenting each frame image of the merchant advertisement video to obtain foreground pixels and background pixels includes:

[0056] Process the merchant advertisement video using the background subtraction method to obtain the foreground region and background region of each frame image;

[0057] Obtain the circumscribed rectangle of the foreground region with the largest area in all frame foreground regions as the reference rectangle;

[0058] Take the geometric center of the foreground region in each frame image as the center of the reference rectangle, and obtain the pixels within the reference rectangle, denoted as foreground pixels. Take the pixels other than the foreground pixels in each frame image as background pixels.

[0059] S3: Denote the sequence of all background pixels at any position in all frame images as the background pixel sequence, segment the background pixel sequence into several background segments with different distributions, and construct a Gaussian distribution model for the background segments; calculate the number of intervals, divide the Gaussian distribution model into the number of intervals, take the mean value of the interval to which the background pixel belongs as the updated value of the background pixel, and use the run-length encoding algorithm to encode the updated values of the background pixels in the background segments; encode the foreground pixels.

[0060] It should be noted that in order to reduce the diversity of the values of background pixels at a position in the frame sequence, it is necessary to de-differentiate different values with small differences, so as to increase the continuation length of consecutive identical data and thus improve the compression effect.

[0061] S30: Denote the sequence of all background pixels at any position in all frame images as the background pixel sequence, segment the background pixel sequence into several background segments with different distributions, and construct a Gaussian distribution model for the background segments; calculate the number of intervals, divide the Gaussian distribution model into S1 intervals, take the mean value of the interval to which the background pixel belongs as the updated value of the background pixel, and use the run-length encoding algorithm to encode the updated values of the background pixels in the background segments.

[0062] S300: Denote the sequence composed of all background pixels at any position in all frame images as the background pixel sequence.

[0063] S301: Segment the background pixel sequence into several background segments with different distributions, and construct the Gaussian distribution model of the background segments.

[0064] It should be noted that due to phenomena such as video splicing in the merchant advertisement video, the background pixels at one position may reflect different objects in different frames. Normally, the pixel values of the same object generally have similar distribution characteristics, and the pixel values of different objects have different distribution characteristics. Therefore, the pixel segments corresponding to different objects can be segmented by the distribution characteristics.

[0065] Preferably, as an example, segmenting the background pixel sequence into several background segments with different distributions and constructing the Gaussian distribution model of the background segments includes:

[0066] Taking any background pixel in the background pixel sequence as the center, obtaining a preset number of pixels as the reference pixels of this background pixel, fitting the Gaussian distribution model with the reference pixels of this background pixel, taking the difference between the Gaussian distribution models of every two adjacent background pixels as the clustering difference, performing clustering processing on all background pixels to obtain several categories, and taking the pixel segments composed of the background pixels in each category as the background segments.

[0067] Fit the Gaussian distribution model of all pixels in each background segment. In this embodiment, it is illustrated by taking the preset number as 11. Other embodiments can take other values, and this embodiment does not limit it.

[0068] It should be added that the difference between two Gaussian distribution models is obtained by multiplying the absolute value of the difference between the means of the two Gaussian distribution models by the absolute value of the difference between the variances of the two Gaussian distribution models.

[0069] S302: Calculate the number of intervals.

[0070] It should be noted that in order to reduce the diversity of pixel values in the frame sequence, similar values need to be de-differentiated. Since the information in the video is inevitably damaged during the process of de-differentiating the values, in order to reduce the damage to the important information in the video, the process of de-differentiating needs to be controlled.

[0071] Preferably, as an example, calculating the number of intervals includes:

[0072]

[0073] Among them, represents the information reproduction rate within the preset area of the background pixels at this position, represents the information content within the preset area of the background pixels at this position. represents the peak value of the Gaussian distribution model. represents the standard deviation of all background pixels in the background segment. represents the number of data value types in the background segment, norm() represents the linear normalization method, and S1 represents the number of intervals. represents the ceiling symbol.

[0074] It can be understood that The larger it is, the greater the information content within the preset area of the background pixels at this position, and the more times it appears repeatedly. This indicates that the information within the preset area of the background pixels at this position is more likely to be the advertising theme information. Therefore, this information loss has a greater impact on the advertisement. Thus, more intervals should be divided to make the values within the intervals more similar, thereby reducing the information loss during de-differentiation. reflects the degree of concentration of pixel values in the background segment. The larger this value is, the smaller the difference in values in this background segment. The number of interval divisions can be appropriately reduced to enhance the intensity of de-differentiation and provide a data compression effect. reflects the degree of difference in the data in the background segment. The greater the degree of difference, the number of interval divisions should be appropriately reduced to enhance the intensity of de-differentiation and improve the compression effect.

[0075] It should be added that the method for obtaining the information reproduction rate includes:

[0076] Taking the background pixels at this position in each frame of the image as the center, obtaining a preset area, denoted as the local area of the background pixels at this position, calculating the similarity between the background pixels at this position and the local areas of each pixel at other positions, and taking the average of the similarities between the local areas of the background pixels at this position and all other positions as the information reproduction rate of the background pixels at this position. In this embodiment, the size of the preset area is taken as as an example for description. Other embodiments can take other values, and this embodiment does not make specific limitations.

[0077] The method for obtaining the information content includes:

[0078] Taking the information entropy within the local area of the background pixels at this position as the information content.

[0079] S303: Divide the Gaussian distribution model into S1 intervals, take the mean value of the interval to which the background pixels belong as the updated value of the background pixels, and use the run-length encoding algorithm to encode the updated values of the background pixels in the background segment.

[0080] Preferably, as an example, taking the mean value of the interval to which the background pixels belong as the updated value of the background pixels, and using the run-length encoding algorithm to encode the updated values of the background pixels in the background segment includes:

[0081] The Gaussian distribution model is evenly divided into S1 regions with the same area, and the pixel value range corresponding to each region is obtained.

[0082] The mean value of all background pixels in the interval to which the background pixels belong is used as the updated value of the background pixels, and the run-length variation algorithm is used to encode the updated values of the background pixels in the background segment to obtain the encoded sequence of the background segment.

[0083] It can be understood that dividing the interval according to the area can take into account the influence of the number of pixels, so that the degree of differentiation of the values with a large number of pixels is relatively small, effectively reducing information loss; using the mean value to replace the values of each background pixel effectively solves the problem of the differentiation of the values of background pixels, thereby effectively extending the length of the same continuous value and improving the compression effect.

[0084] S31: Encode the foreground pixels.

[0085] Preferably, as an example, encoding the foreground pixels includes:

[0086] The region formed by the foreground pixels in each frame of the image is denoted as the analysis foreground region; the analysis foreground region with the closest distance in adjacent frames is used as the corresponding analysis foreground region;

[0087] Align the corresponding analysis foreground regions of all frames, and the sequence formed by all foreground pixels at any position in the corresponding analysis foreground regions of all frames is denoted as the foreground pixel sequence;

[0088] The foreground pixel sequence is divided into several foreground segments according to the method of dividing the background segment;

[0089] Encode each foreground segment according to the encoding method of the background segment to obtain the encoded sequence of each foreground segment.

[0090] It can be understood that the corresponding foreground region represents the region where the same moving object is located in different frames. Since the values of the same moving object in the frame sequence are highly similar, the loss of video information caused by the de-differentiation process is not too large. However, through the de-differentiation process, the continuation length of continuous identical values can be greatly extended. Therefore, the compression effect can be greatly improved through the de-differentiation process.

[0091] S4: To realize the transmission of the merchant advertising video.

[0092] Preferably, as an example, to realize the transmission of the merchant advertising video, it includes:

[0093] Concatenate the encoding sequences of all background segments of the background pixel sequence to obtain the encoding sequence of the background pixel sequence, concatenate the encoding sequences of all foreground segments of the foreground pixel sequence to obtain the encoding sequence of the foreground pixel sequence, obtain the geometric center coordinates of the analyzed foreground region, concatenate all the geometric center coordinates of the analyzed foreground regions to obtain a trajectory sequence, obtain the starting frame number of the analyzed foreground region, insert the starting frame number, the length and width of the reference rectangle before the trajectory sequence, insert a preset delimiter flag value after concatenating the trajectory sequence with the inserted starting frame number and the encoding sequences of all foreground pixel sequences, and then concatenate with the encoding sequences of all background pixel sequences to obtain a transmission sequence, and perform transmission processing on the transmission sequence. In this embodiment, the ASCALL code of the semicolon is used as the decomposition flag value, and other embodiments can take other values, which are not specifically limited in this embodiment.

[0094] S5: Perform decoding processing on the transmission sequence.

[0095] Preferably, as an example, performing decoding processing on the transmission sequence includes:

[0096] Using the delimiter flag value, separate the encoding sequence of the background pixel sequence from other sequences;

[0097] Separate the trajectory sequence, the starting frame number, the encoding sequences of all foreground pixel sequences, and the length and width of the reference rectangle in the other sequences; separate the encoding sequences of all foreground pixel sequences according to the continuation frame number, perform decoding processing on the encoding sequences of each foreground pixel sequence using the run-length encoding algorithm to obtain each foreground pixel sequence, and use the starting frame number, each geometric center coordinate in the trajectory sequence, and the length and width of the reference rectangle to locate the frame where each foreground pixel in the foreground pixel sequence is located and the coordinates in the frame.

[0098] For the convenience of explanation, the following uses an example to explain the content of locating the frame where each foreground pixel in the foreground pixel sequence is located and the coordinates in the frame using the starting frame number, each geometric center coordinate in the trajectory sequence, and the length and width of the reference rectangle: Taking the positioning method of the second foreground pixel in the second foreground pixel sequence as an example, assuming the starting frame number is 2, then the second foreground pixel in the second foreground pixel sequence is a pixel in the third frame image, and the geometric center coordinate corresponding to the second foreground pixel sequence for the analyzed foreground region is the second element in the trajectory sequence. Assuming the second element in the trajectory sequence is , and the length and width of the reference rectangle are 11 and 11 respectively, then the coordinate position of the second foreground pixel is , thus locating the frame where the second foreground pixel in the second foreground pixel sequence is located and the coordinates in the frame.

[0099] The encoded sequences of each background pixel sequence are decoded using the run-length encoding algorithm to obtain the background pixel sequence, and the frame in which each background pixel in the background pixel sequence is located and its coordinates in the frame are located.

[0100] An embodiment of the present invention also discloses an advertising data transmission system based on a cloud platform, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an advertising data transmission method based on the present invention is implemented.

[0101] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0102] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and the program can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high-bandwidth memory, a hybrid storage cube, etc., or any other medium that can be used to store the required information and can be accessed by an application program, a module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device.

[0103] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

[0104] The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A cloud platform-based advertising data transmission method, characterized in that: Includes steps: Get the merchant's advertising video; Segment each frame of the merchant advertisement video to obtain foreground pixels and background pixels; The sequence of all background pixels at any position in all frame images is recorded as the background pixel sequence; the background pixel sequence is divided into several background segments with different distributions; a Gaussian distribution model of the background segment is constructed, and the number of intervals is calculated: , Indicates the information recurrence rate within the preset area of ​​the background pixel at that position, Indicates the information content of the background pixel at that location within the preset area. represents the peak value of the Gaussian distribution model, represents the mean of the differences between all two adjacent pixels in the background segment, Indicates the number of data values ​​in the background segment, norm() indicates the linear normalization method, S1 indicates the number of intervals, Indicates the rounding up symbol, divides the Gaussian distribution model into S1 intervals, takes the mean value of the interval to which the background pixel belongs as the update value of the background pixel, and uses the run-length encoding algorithm to encode the update value of the background pixel in the background segment; Encode foreground pixels; To achieve merchant advertising video transmission.

2. The method for transmitting advertising data based on a cloud platform according to claim 1, characterized in that: The segmentation process of each frame of the merchant advertisement video to obtain foreground pixels and background pixels includes: The merchant advertisement video is processed using background subtraction method to obtain the foreground area and background area of ​​each frame image; Obtain the bounding rectangle of the largest foreground area in all frame foreground areas as the reference rectangle; The geometric center of the foreground area in each frame image is taken as the center of the reference rectangle, the pixels within the reference rectangle are obtained and recorded as foreground pixels, and the pixels other than the foreground pixels in each frame image are taken as background pixels.

3. The method for transmitting advertising data based on a cloud platform according to claim 1, characterized in that: The step of dividing the background pixel sequence into a plurality of background segments with different distributions includes: Taking any background pixel in the background pixel sequence as the center, a preset number of pixels are obtained as the reference pixels of the background pixel, and a Gaussian distribution model is fitted using the reference pixels of the background pixel. The difference in the Gaussian distribution model between every two adjacent background pixels is used as the clustering difference. All background pixels are clustered to obtain several categories, and the pixel segments composed of background pixels in each category are used as background segments.

4. The method for transmitting advertising data based on a cloud platform according to claim 3, characterized in that: The method for obtaining the difference of the Gaussian distribution model includes: The difference between the two Gaussian distribution models is obtained by multiplying the absolute value of the difference between the means of the two Gaussian distribution models by the absolute value of the difference between the variances of the two Gaussian distribution models.

5. The method for transmitting advertising data based on a cloud platform according to claim 1, characterized in that: The method for obtaining the information recurrence rate comprises: Taking the background pixel at that position in each frame image as the center, obtain a preset area, record it as the local area of ​​the background pixel at that position, calculate the similarity between the background pixel at that position and the local areas of pixels at other positions, and take the average of the similarities between the background pixel at that position and the local areas of all other positions as the information recurrence rate of the background pixel at that position.

6. The method for transmitting advertising data based on a cloud platform according to claim 5, characterized in that: The method for obtaining the information content comprises: The information entropy in the local area of ​​the background pixel at that position is taken as the information content.

7. The method for transmitting advertising data based on a cloud platform according to claim 1, characterized in that: The Gaussian distribution model is divided into S1 intervals, including: The Gaussian distribution model is evenly divided into S1 regions of equal area, and the pixel value interval corresponding to each region is obtained.

8. The method for transmitting advertising data based on a cloud platform according to claim 1, characterized in that: The encoding of the foreground pixels comprises: The area formed by the foreground pixels in each frame of the image is recorded as the analysis foreground area; Aligning the analyzed foreground areas of all frames, and recording a sequence consisting of all foreground pixels at any position in the analyzed foreground areas of all frames as a foreground pixel sequence; Divide the foreground pixel sequence into several foreground segments with different distributions; Encode the foreground segment.

9. The method for transmitting advertising data based on a cloud platform according to claim 8, characterized in that: The method of realizing the transmission of merchant advertisement video includes: The coding sequences of all background segments of the background pixel sequence are spliced ​​together to obtain the coding sequence of the background pixel sequence, the coding sequences of all foreground segments of the foreground pixel sequence are spliced ​​together to obtain the coding sequence of the foreground pixel sequence, the geometric center coordinates of the analyzed foreground area are obtained, the geometric center coordinates of all analyzed foreground areas are spliced ​​together to obtain a trajectory sequence, the starting frame number of the analyzed foreground area is obtained, the starting frame number, the length and width of the reference rectangle are inserted before the trajectory sequence, the trajectory sequence with the inserted starting frame number is spliced ​​together with the coding sequences of all foreground pixel sequences, and then a preset demarcation mark value is inserted, and then it is spliced ​​together with the coding sequences of all background pixel sequences to obtain a transmission sequence, and the transmission sequence is transmitted.

10. An advertising data transmission system based on a cloud platform, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a cloud platform-based advertising data transmission method according to any one of claims 1 to 9 is implemented.

Images