A big data-based digital marketing method

Abstract

The application provides a kind of digital marketing method based on big data, comprising S1, obtaining the image to be identified of customer in front of advertisement player;S2, after the image to be identified is preprocessed, input into the neural network model trained in advance to carry out image recognition processing, obtain the first label of customer;S3, according to the first label, obtain the marketing advertisement video to be played;S4, marketing advertisement video is pushed to advertisement player and played.The application identifies the first label of customer by obtaining the image of customer in front of advertisement player, then according to the first label, obtains the corresponding marketing advertisement video to be played and plays, so as to effectively improve the pertinence of marketing advertisement video, improve the effectiveness of advertisement marketing.

Description

Technical Field

[0001] The present invention relates to the field of marketing, and in particular to a digital marketing method based on big data. Background Art

[0002] In the prior art, dynamic advertisements in shopping malls are generally displayed on display screens installed in various locations in the mall. The advertisements displayed on the display screens are generally played in a loop and cannot be played in a targeted manner based on the people in front of the display screens, resulting in insufficient targeting of marketing and affecting marketing effectiveness. Summary of the Invention

[0003] The purpose of the present invention is to disclose a digital marketing method based on big data to solve the problem that the advertising played on the display screen of a shopping mall is not targeted enough, which affects the marketing effect.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] A digital marketing method based on big data, including

[0006] S1, obtaining an image of a customer to be identified who is in front of the advertising player;

[0007] S2, after pre-processing the image to be identified, input it into a pre-trained neural network model for image recognition processing to obtain the customer's first label;

[0008] S3, obtaining a marketing advertisement video to be played according to the first tag;

[0009] S4, push the marketing advertisement video to the advertisement player for playback.

[0010] Preferably, the S1 includes:

[0011] The proximity sensor is used to determine whether there is a customer in the preset area in front of the advertising player. If so, the camera is used to obtain the customer's image to be identified.

[0012] Preferably, the S2 includes:

[0013] Preprocessing the image to be identified by a preset preprocessing method to obtain a preprocessed image;

[0014] The preprocessed image is input into the pre-trained neural network model for image recognition processing to obtain the customer's first label.

[0015] Preferably, the preprocessing of the image to be identified by a preset preprocessing method to obtain a preprocessed image includes:

[0016] Performing light adjustment processing on the image to be recognized to obtain a first image;

[0017] grayscale the first image to obtain a second image;

[0018] Perform noise reduction processing on the second image to obtain a preprocessed image.

[0019] Preferably, performing grayscale processing on the first image to obtain the second image includes:

[0020] Use the following formula to grayscale the first image:

[0021] gray(x,y)=w1×R(x,y)+w2×G(x,y)+w3×B(x,y)

[0022] Among them, (x,y) represents the coordinates of the pixel point, R(x,y), G(x,y), and B(x,y) represent the pixel values ​​of the pixel point with coordinates (x,y) in the red component image, green component image, and blue component image, respectively; w1, w2, and w3 represent preset scale coefficients; gray(x,y) represents the grayscale value of the pixel point with coordinates (x,y).

[0023] Preferably, the marketing advertisement video includes a second tag and a broadcast time.

[0024] Preferably, the S3 includes:

[0025] Obtaining the shooting time of the image to be identified;

[0026] Store the marketing advertisement video corresponding to the second tag that is the same as the first tag into the set U;

[0027] The marketing advertising video with the smallest difference between the broadcast time and the shooting time in the set U is used as the marketing advertising video to be played.

[0028] The present invention identifies the customer's first tag by acquiring an image of the customer in front of the advertising player, and then acquires the corresponding marketing advertising video to be played according to the first tag for playing, thereby effectively improving the pertinence of the marketing advertising video and improving the effectiveness of advertising marketing. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0030] Figure 1 , which is a diagram of an exemplary embodiment of a digital marketing method based on big data of the present invention. DETAILED DESCRIPTION

[0031] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0032] like Figure 1 In one embodiment shown, the present invention provides a digital marketing method based on big data, comprising

[0033] S1, obtaining an image of a customer to be identified who is in front of the advertising player;

[0034] S2, after pre-processing the image to be identified, input it into a pre-trained neural network model for image recognition processing to obtain the customer's first label;

[0035] S3, obtaining a marketing advertisement video to be played according to the first tag;

[0036] S4, push the marketing advertisement video to the advertisement player for playback.

[0037] The present invention identifies the customer's first tag by acquiring an image of the customer in front of the advertising player, and then acquires the corresponding marketing advertising video to be played according to the first tag for playing, thereby effectively improving the pertinence of the marketing advertising video and improving the effectiveness of advertising marketing.

[0038] Specifically, the neural network model can be a BP neural network model. The training data source can be real shopping guide data of shopping mall salespeople. For example, after serving a young person, the salesperson can record the young person's clothing, hairstyle, shoes, body shape, etc., and then label the young person with a first label.

[0039] Shopping guides accumulate a large amount of shopping guide data in their daily shopping activities. Therefore, the neural network model is trained with this big data to obtain a trained neural network model. The trained neural network model can obtain the first label of the customer based on the customer's image to be identified.

[0040] Preferably, the S1 includes:

[0041] The proximity sensor is used to determine whether there is a customer in the preset area in front of the advertising player. If so, the camera is used to obtain the customer's image to be identified.

[0042] Specifically, the preset area may be the area directly in front of the advertisement player, and the size of the area may be adjusted according to actual conditions.

[0043] When no customer is detected, the advertisement player plays the advertisement in a loop, and after a customer is detected, it plays a targeted advertisement according to the first tag of the customer.

[0044] Preferably, the S2 includes:

[0045] Preprocessing the image to be identified by a preset preprocessing method to obtain a preprocessed image;

[0046] The preprocessed image is input into the pre-trained neural network model for image recognition processing to obtain the customer's first label.

[0047] Preferably, the preprocessing of the image to be identified by a preset preprocessing method to obtain a preprocessed image includes:

[0048] Performing light adjustment processing on the image to be recognized to obtain a first image;

[0049] grayscale the first image to obtain a second image;

[0050] Perform noise reduction processing on the second image to obtain a preprocessed image.

[0051] Performing illumination adjustment processing on the image to be recognized is beneficial for making the illumination distribution in the obtained first image more uniform, thereby suppressing overexposed positions and enhancing dark details, effectively improving the image quality.

[0052] Preferably, performing light adjustment processing on the image to be recognized to obtain the first image includes:

[0053] Obtain the L component image imgL of the image to be identified;

[0054] Perform the following adjustment processing on imgL to obtain the first adjusted image:

[0055]

[0056] Wherein, oimgL represents the first adjustment image, oimgL(u) represents the pixel value of pixel u in the first adjustment image, c1, c2, and c3 represent the first constant coefficient, the second constant coefficient, and the third constant coefficient, respectively, w1 represents the preset adjustment coefficient, δ represents the preset control coefficient, imgL(u) represents the pixel value of pixel u in imgL, aveimgL represents the average pixel value of all pixels in imgL, and vaz represents the variance of the pixel values ​​of all pixels in imgL;

[0057] Perform the following adjustment processing on imgL to obtain the second adjusted image:

[0058] Use edge detection algorithm to perform edge detection on imgL and obtain the set of edge pixel points SU;

[0059] Use the following formula to adjust the pixels in SU respectively:

[0060]

[0061] Wherein, timgL represents the second adjustment image, timgL(v) represents the pixel value of pixel v in timgL, shv represents the value function, if aveneiv-imgL(v) is greater than 0, then shv(aveneiv-imgL(v)) is 1, if aveneiv-imgL(v) is equal to 0, then shv(aveneiv-imgL(v)) is 0, if aveneiv-imgL(v) is less than 0, then shv(aveneiv-imgL(v)) is -1, neiv represents the set of pixel points in a neighborhood of a preset size centered on pixel point v, imgL(s) represents the pixel value of pixel point s in neiv in imgL, imgL(v) represents the pixel value of pixel point v in imgL, aveneiv represents the median of the pixel values ​​of all pixel points in neiv; Φ represents the total number of elements contained in neiv; w1 represents the preset adjustment coefficient;

[0062] The following formula is used to perform weighted summation on the first adjusted image and the second adjusted image to obtain the adjusted L component image:

[0063] finimgL=α×oimgL+β×timgL

[0064] Wherein, finimgL represents the adjusted L component image, α and β represent the preset weight parameters, α+β=1;

[0065] Convert finimgL back to RGB color space to get the first image.

[0066] Existing light adjustment processing methods generally use the same algorithm to process all pixels and then directly output the processed results. However, edge pixels in the image are often affected, resulting in the loss of some image edge information, affecting the accuracy of subsequent image recognition processing. Therefore, in the present invention, after obtaining edge pixels, another light adjustment algorithm is used to process them. The results of the first and second adjusted images are then combined to obtain the final first image. This effectively improves the retention of edge detail information in the first image, thereby improving the accuracy of the first label.

[0067] Preferably, the control coefficient is calculated as follows:

[0068] δ=ln(var)-ln(stdev)

[0069] Among them, var represents the standard deviation of all pixels in imgL, stdev represents the variance of the noise estimation of imgL, and δ represents the control coefficient.

[0070] Preferably, the value range of c1 is [98, 100], the value range of c2 is [9, 11], and the value range of c3 is [49, 51].

[0071] Preferably, performing grayscale processing on the first image to obtain the second image includes:

[0072] Use the following formula to grayscale the first image:

[0073] gray(x,y)=w1×R(x,y)+w2×G(x,y)+w3×B(x,y)

[0074] Among them, (x,y) represents the coordinates of the pixel point, R(x,y), G(x,y), and B(x,y) represent the pixel values ​​of the pixel point with coordinates (x,y) in the red component image, green component image, and blue component image, respectively; w1, w2, and w3 represent preset scale coefficients; gray(x,y) represents the grayscale value of the pixel point with coordinates (x,y).

[0075] Preferably, the marketing advertisement video includes a second tag and a broadcast time.

[0076] Preferably, the S3 includes:

[0077] Obtaining the shooting time of the image to be identified;

[0078] Store the marketing advertisement video corresponding to the second tag that is the same as the first tag into the set U;

[0079] The marketing advertising video with the smallest difference between the broadcast time and the shooting time in the set U is used as the marketing advertising video to be played.

[0080] Specifically, each marketing advertisement will be tagged with a second tag, and the second tag is used to match with the customer's first tag.

[0081] For example, when the first tag is "young white-collar worker," marketing advertisements with the second tag "young white-collar worker" are retrieved from all marketing advertisements. Since the ad player uses a loop when there are no customers in the preset area, each marketing advertisement has a corresponding broadcast time. The same second tag may correspond to several different marketing advertisements. Therefore, the broadcast time is used to determine the marketing advertisement to be played. The marketing advertisement video with the broadcast time closest to the capture time of the image to be identified is selected as the marketing advertisement to be played.

[0082] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

[0083] It should be noted that the functional units / modules in various embodiments of the present invention can be integrated into one processing unit / module.

[0084] In the block, each unit / module may exist physically separately, or two or more units / modules may be integrated into one unit / module. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of software functional units / modules.

[0085] Through the description of the above embodiments, it will be clear to those skilled in the art that it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to implement the functions described herein, or a combination thereof. For software implementation, part or all of the processes of the embodiments can be completed by instructing the relevant hardware through a computer program.

[0086] When implemented, the above program can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any media that facilitates the transmission of computer programs from one place to another. The storage medium can be any available medium that can be accessed by a computer. Computer-readable media can include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer.

Claims

1. A digital marketing method based on big data, characterized in that: include: S1, obtaining an image of a customer to be identified who is in front of the advertising player; S2, after pre-processing the image to be identified, input it into a pre-trained neural network model for image recognition processing to obtain the customer's first label; S3, obtaining a marketing advertisement video to be played according to the first tag; S4, push the marketing advertising video to the advertising player for playback; The S2 includes: Preprocessing the image to be identified by a preset preprocessing method to obtain a preprocessed image; Input the pre-processed image into the pre-trained neural network model for image recognition processing to obtain the customer's first label; The preprocessing of the image to be identified by a preset preprocessing method to obtain a preprocessed image includes: Performing light adjustment processing on the image to be recognized to obtain a first image; grayscale the first image to obtain a second image; performing noise reduction processing on the second image to obtain a preprocessed image; The step of performing light adjustment processing on the image to be recognized to obtain a first image includes: Get the L component image of the image to be recognized ; right Perform the following adjustment process to obtain a first adjusted image: ； in, represents the first adjusted image, represents the pixel value of pixel u in the first adjusted image, Represent the first constant coefficient, the second constant coefficient, and the third constant coefficient respectively. Indicates the preset adjustment coefficient, Indicates the preset control coefficient, Indicates that pixel u is The pixel values ​​in express The average value of all pixel values ​​in , express The variance of the pixel values ​​of all pixels in ; right Perform the following adjustment process to obtain a second adjusted image: Use edge detection algorithm to Perform edge detection to obtain a set of edge pixel points SU; Use the following formula to adjust the pixels in SU respectively: ； in, represents the second adjusted image, Indicates that the pixel v is The pixel values ​​in represents the value function, if Greater than 0, then is 1, if is equal to 0, then Is 0, if Less than 0, then = -1, neiv represents the set of pixels in a neighborhood of preset size centered on pixel v. Indicates that the pixel s in neiv is The pixel values ​​in Indicates that the pixel v is The pixel values ​​in Represents the median value of all pixel values ​​in neiv; Indicates the total number of elements contained in neiv; Indicates the preset adjustment coefficient; The following formula is used to perform weighted summation on the first adjusted image and the second adjusted image to obtain the adjusted L component image: ； in, represents the adjusted L component image, and Represents the preset weight parameter, ; Will Convert back to RGB color space to get the first image.

2. A digital marketing method based on big data according to claim 1, characterized in that: Said S1 comprises: The proximity sensor is used to determine whether there is a customer in the preset area in front of the advertising player. If so, the camera is used to obtain the customer's image to be identified.

3. The digital marketing method based on big data according to claim 1, characterized in that: The grayscale processing of the first image to obtain the second image includes: Use the following formula to grayscale the first image: ； in, Represents the coordinates of the pixel point, 、 、 Respectively represent the coordinates The pixel value of the pixel point in the red component image, the green component image, and the blue component image; 、 、 Indicates the preset scale factor; The coordinates are The grayscale value of the pixel.

4. The digital marketing method based on big data according to claim 1, characterized in that: The marketing advertisement video includes a second tag and a broadcast time.

5. The digital marketing method based on big data according to claim 4, characterized in that: The S3 includes: Obtaining the shooting time of the image to be identified; Store the marketing advertisement video corresponding to the second tag that is the same as the first tag into the set U; The marketing advertising video with the smallest difference between the broadcast time and the shooting time in the set U is used as the marketing advertising video to be played.

Images