Signaling matching method and device based on intelligent identification, and electronic equipment

Abstract

The invention provides a signaling matching method and device based on intelligent identification, and electronic equipment, and relates to the technical field of data processing, and the method comprises the steps: constructing the signaling receiving states of collection targets in a plurality of preset time periods into a time sequence information matrix, and carrying out the superposition of the time sequence information matrix to generate a second information matrix reflecting the long-term heat; the matrix element values are converted into gray level image pixel values, and a visual gray level image which is divided into columns according to an acquisition target is generated; the target corresponding to the highlight column in the image is accurately recognized based on the preset gray threshold, so that the high-frequency active user, namely the selected target, is intelligently screened out, the selected target serves as the hotspot short message data main body, hotspot data can be accurately judged, invalid data exchange between a high-speed cache and a low-speed cache is avoided, and the user experience is improved. According to the method, the high-value data is ensured to stably reside in the cache, the matching efficiency and real-time performance of the short message data and the signaling data are remarkably improved, and the computing resource consumption and the cache cost of the system are reduced while important short messages are ensured to be accurately reached.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of data processing, and particularly relates to a signaling matching method and device based on intelligent identification, and an electronic device. BACKGROUND

[0002] In order to better promote the development of gateway business, combine user online signaling and short message service, and judge the current use of mobile phone by user signaling data, it is necessary to perform high-speed real-time matching between short message data and signaling data to judge the short message pushing time, and push important short messages to users at appropriate time to improve the delivery rate of short messages. In the matching scenario of short message data and signaling data, as the scale of short message data and signaling data gradually increases, the calculation cost of matching gradually increases, and thus the real-time performance gradually decreases. Therefore, it is necessary to efficiently use cache to perform matching between short message data and signaling data.

[0003] However, although the cache is introduced in the prior art, the cache cost is high, and it is usually necessary to put hot short message data into the cache to perform signaling matching to improve the signaling matching efficiency. However, since the hot short message data cannot be accurately determined, it is necessary to continuously exchange data between the cache and the low-speed cache, which affects the efficiency of matching between short message data and signaling data.

[0004] Therefore, how to accurately determine the hot short message data is a problem to be solved at present. SUMMARY

[0005] The present disclosure provides a signaling matching method and device based on intelligent identification, and an electronic device. The main purpose is to solve the problem of how to accurately determine the hot short message data.

[0006] According to a first aspect of the present disclosure, a signaling matching method based on intelligent identification is provided, which comprises:

[0007] According to the receiving state of the signaling data corresponding to each of the plurality of collection targets in the first preset time period, a matrix construction process is performed to obtain a first information matrix corresponding to the first preset time period;

[0008] The first information matrix corresponding to each of the plurality of first preset time periods is added to obtain a second information matrix, wherein the maximum matrix element value in the second information matrix is not greater than the first time period number of the plurality of first preset time periods;

[0009] According to the first time period number and the second information matrix, a gray scale conversion process is performed to obtain a target gray scale diagram corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a gray scale value of each pixel point in the target gray scale diagram, and a plurality of pixel point columns of the target gray scale diagram each correspond to a plurality of collection targets.

[0010] The pixel point column in which the pixel point with the gray value greater than the preset gray threshold in the target gray image is determined as a target pixel point column, and the collection target corresponding to the target pixel point column is determined as a selected target.

[0011] The information data corresponding to the selected target is stored in the cache for signaling matching processing.

[0012] According to a second aspect of the present disclosure, a signaling matching device based on intelligent identification is provided, comprising:

[0013] The construction unit is configured to perform matrix construction processing on the reception states of the signaling data corresponding to the plurality of collection targets in the first preset time period, to obtain a first information matrix corresponding to the first preset time period.

[0014] The calculation unit is configured to add the first information matrices corresponding to the plurality of first preset time periods to obtain a second information matrix, wherein the maximum matrix element value in the second information matrix is not greater than the first time period number of the plurality of first preset time periods.

[0015] The conversion unit is configured to perform gray image conversion processing according to the first time period number and the second information matrix to obtain a target gray image corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a gray value of each pixel point in the target gray image, and each of the plurality of pixel point columns of the target gray image corresponds to a plurality of collection targets.

[0016] The determination unit is configured to determine a pixel point column in which a pixel point with a gray value greater than a preset gray threshold in a target gray image as a target pixel point column, and determine a collection target corresponding to the target pixel point column as a selected target.

[0017] The matching unit is configured to store information data corresponding to the selected target in the cache for signaling matching processing.

[0018] According to a third aspect of the present disclosure, an electronic device is provided, comprising:

[0019] at least one processor; and

[0020] a memory in communication with the at least one processor; wherein

[0021] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of the first aspect.

[0022] According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are configured to cause a computer to execute the method of the first aspect.

[0023] According to a fifth aspect of the present disclosure, there is provided a computer program product comprising a computer program configured to implement the method of the first aspect when executed by a processor.

[0024] The signaling matching method and device based on intelligent identification, and the electronic device provided by the present disclosure, by constructing the signaling receiving state of each collection target in multiple preset time periods into a time sequence information matrix, and superimposing a second information matrix reflecting long-term heat, and then converting the matrix element values into gray image pixel values to generate a visual gray image arranged according to the collection targets, and accurately identifying the target corresponding to the highlighted column in the image based on a preset gray threshold, thereby intelligently selecting a high-frequency active user as the selected target, and taking the selected target as the main body of the hot short message data, the hot data can be accurately determined, and the invalid data exchange between the high-speed cache and the low-speed cache is avoided, the high-value data is ensured to stably reside in the high-speed cache, the matching efficiency and real-time performance of the short message data and the signaling data are significantly improved, while the important short messages are accurately reached, the system computing resource consumption and the cache cost are reduced.

[0025] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0026] The accompanying drawings are used to better understand the present scheme and do not constitute a limitation on the present disclosure. Among them:

[0027] Figure 1 A flowchart of a signaling matching method based on intelligent identification provided by an embodiment of the present disclosure;

[0028] Figure 2 A flowchart of a target gray image provided by an embodiment of the present disclosure;

[0029] Figure 3 A target gray image provided by an embodiment of the present disclosure;

[0030] Figure 4 A schematic diagram of the principle of element elimination provided by an embodiment of the present disclosure;

[0031] Figure 5 A configuration platform schematic diagram of a preset information identification algorithm provided by an embodiment of the present disclosure;

[0032] Figure 6A flowchart of an information data acquisition process of a selected target according to an embodiment of the present disclosure is provided.

[0033] Figure 7 A structural diagram of a signaling matching device based on intelligent identification according to an embodiment of the present disclosure is provided.

[0034] Figure 8 A schematic block diagram of an electronic device according to an embodiment of the present disclosure is provided. DETAILED DESCRIPTION

[0035] Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to assist in understanding, which should be considered in a descriptive sense only. Thus, it will be apparent to one of ordinary skill in the art that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, for the sake of brevity and clarity, descriptions of well-known functions and constructions are omitted from the following description.

[0036] A signaling matching method and device based on intelligent identification and an electronic device according to an embodiment of the present disclosure are described below with reference to the accompanying drawings.

[0037] The present disclosure relates to a signaling matching method based on intelligent identification, which significantly improves the real-time performance of large-scale signaling data and SMS matching through innovative data processing and image recognition technology. The traditional cache optimization problem based on Central Processing Unit (CPU) intensive calculation can be converted into an image recognition problem suitable for Graphics Processing Unit (GPU) / Neural Processing Unit (NPU) parallel processing. The specific steps of the method are described in detail below:

[0038] Figure 1 A flowchart of a signaling matching method based on intelligent identification according to an embodiment of the present disclosure is provided.

[0039] As shown in Figure 1 , the method comprises the following steps:

[0040] Step 101, according to the receiving state of the signaling data corresponding to each of the plurality of collection targets in the first predetermined time period, a matrix construction process is performed to obtain a first information matrix corresponding to the first predetermined time period.

[0041] In embodiments of the present disclosure, the first preset time period is a custom-determined continuous, fixed-length time period, for example, the first preset time period is defined as a natural day (24 hours). The target for collection refers to a user participating in signaling matching, and the unique identifier (such as a mobile phone number) of the user is taken as the basis for data association. The signaling data refers to real-time information reflecting the network activity state of the user (such as screen-on state, data transmission request, etc.). The reception state refers to whether signaling activity of the user equipment is detected in a unit divided according to a fixed time granularity (such as every 5 minutes) within the first preset time period. If signaling reception exists, it is marked as an effective state (value 1), otherwise as an invalid state (value 0).

[0042] The operation definition of the matrix construction process is that the time unit is taken as the row (a plurality of fixed time granularities within the first preset time period, such as 288 5-minute units per day), the target for collection is taken as the column, the signaling reception state (0 or 1) of each user in each time unit per day is filled into the corresponding position, and a binary matrix with 288 rows and N columns (N is the total number of users, i.e. the total number of targets for collection) is formed, that is, the first information matrix. The mathematical essence of the first information matrix is the discretization projection of user behavior in the time dimension.

[0043] Specifically, after the state of each user per day is represented by a vector with a length of 288, the state of N users per day can be represented in the form of a matrix, the number of rows of the matrix is 288, and the number of columns of the matrix is N. The first matrix information can be represented by but not limited to the following ways:

[0044]

[0045] Step 102, adding the first information matrix corresponding to each of the plurality of first preset time periods respectively to obtain a second information matrix, wherein the maximum matrix element value in the second information matrix is not greater than the number of first time periods of the plurality of first preset time periods.

[0046] In embodiments of the present disclosure, the number of first time periods refers to the total number of first preset time periods participating in the superposition calculation (for example, m days). The generation process of the second information matrix includes but is not limited to element-level addition of m first information matrices according to the same row and column positions. Since the matrix element value of the first information matrix is 0 or 1, the value range of each element in the superposition matrix is 0 to m (inclusive), indicating the total number of days of signaling activity of the corresponding user in the same time unit. The generation of the second information matrix is mathematically equivalent to the statistical quantification of the user historical behavior pattern, and the row and column dimensions thereof remain consistent with those of the first information matrix (288xN).

[0047] Specifically, when there are m days of data, the corresponding elements of the m matrices are added, and thus a matrix of 288 rows and N columns is obtained, and each element of the matrix has a value of 0-m. The second information matrix can be represented by, but is not limited to, the following manner:

[0048]

[0049] In step 103, a gray-scale image conversion process is performed according to the first time period quantity and the second information matrix to obtain a target gray-scale image corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a gray-scale value of each pixel point in the target gray-scale image, and a plurality of pixel point columns of the target gray-scale image each correspond to a plurality of collection targets.

[0050] In an embodiment of the present disclosure, it is necessary to convert the numerical second information matrix into a target gray-scale image that can be visually analyzed. The gray-scale image is a single-channel image format in computer vision, and the brightness value (gray-scale value) of each pixel point ranges from 0 (pure black) to 255 (pure white). The conversion rule is that each matrix element value in the second information matrix is linearly mapped to the interval of 0-255 as the gray-scale value of the corresponding pixel point in the gray-scale image.

[0051] The specific conversion method is dynamically adjusted according to the first time period quantity m:

[0052] When m<255 (i.e., a preset quantity threshold), scaling is performed on each matrix element value in the second information matrix by a preset formula, and a gray-scale value is output; when m≥255, a sliding window mechanism is adopted: only the first information matrix of the most recent 255 days is retained to participate in superposition, and when new data is added to the right end of the window, the oldest data at the left end of the window is automatically removed, so that the upper limit of the matrix element value is always 255. At this time, the gray-scale value can be directly taken as each matrix element value in the second information matrix.

[0053] The finally generated target gray-scale image has a height of 288 pixels (corresponding to a time unit) and a width of N pixels (each pixel point column strictly corresponds to one collection target, i.e., a user), forming a visual representation of a two-dimensional structure of time and user.

[0054] Specifically, regarding the construction process of the target gray-scale image, an embodiment of the present disclosure provides a construction flowchart of a target gray-scale image, as shown in Figure 2 and a target gray-scale image, as shown in Figure 3 wherein a matrix needs to be constructed according to user signaling information (i.e., the second information matrix is constructed), and then the matrix is converted into a gray-scale image.

[0055] In step 104, a pixel point column in which a pixel point with a gray-scale value greater than a preset gray-scale threshold is located in the target gray-scale image is determined as a target pixel point column, and a collection target corresponding to the target pixel point column is determined as a selected target.

[0056] In the embodiments of the present disclosure, the preset gray threshold is a threshold value set by custom, which is an empirical threshold value set for distinguishing hot users, and can be dynamically configured according to the real-time requirement of the system. The pixel point column refers to a set of pixels in the vertical direction in the gray image, and each column is uniquely associated with a collection target. The identification process is as follows: all pixel point columns of the gray image are traversed, and if there is at least one pixel point in a column whose gray value exceeds the preset gray threshold, the column is determined as a target pixel point column. Since the high and low of the gray value directly reflects the historical signaling activity of the user (high gray value represents high frequency activity), the collection target corresponding to the target pixel point column marked is the selected target, that is, the hot user with a high probability of triggering the short message matching.

[0057] In step 105, the information data corresponding to the selected target is stored in the cache for signaling matching processing.

[0058] In the embodiments of the present disclosure, the information data corresponding to the selected target refers to the screen-on short message metadata (such as: user identifier, short message content, priority identifier, etc.) of the selected target. The cache refers to a temporary data storage area (such as: memory database or distributed cache system) with significantly higher access speed than the main memory. The storage strategy is: all associated short message data of the selected target are preloaded into the cache area. The signaling matching processing refers to searching the target user data in the cache first when the real-time signaling arrives, thereby skipping the low-speed storage query link. By continuously updating the selected target set, it is ensured that the active user data is always retained in the cache, and the matching response efficiency is greatly improved.

[0059] The present disclosure realizes the quality and efficiency improvement of the cache optimization strategy by encoding massive user signaling data into a gray image and identifying hot users by using computer vision technology. The present application transfers the computing load to the GPU / NPU driven image processing pipeline, significantly shortens the hot user identification time; at the same time, the intuitive correspondence between the brightness of the pixels in the gray image and the user activity makes the cache placement more accurate, effectively improves the cache hit rate, and finally guarantees the real-time requirement of the short message signaling matching under the super large scale data set.

[0060] The signaling matching method based on intelligent identification provided by the present disclosure constructs the signaling receiving state of each collection target in multiple preset time periods into a time sequence information matrix, and superimposes a second information matrix reflecting long-term heat; then, the matrix element values are converted into gray image pixel values to generate a visual gray image arranged according to the collection targets; the target corresponding to the highlighted column in the image is accurately identified based on a preset gray threshold, so that the high-frequency active user, i.e., the selected target, is intelligently screened out, the selected target is taken as the hotspot short message data subject, the hotspot data can be accurately determined, the invalid data exchange between the high-speed cache and the low-speed cache is avoided, the high-value data is ensured to stably reside in the high-speed cache, the matching efficiency and real-time performance of the short message data and the signaling data are significantly improved, the important short messages are accurately reached at the same time, and the system computing resource consumption and the cache cost are reduced.

[0061] In an implementation manner of the embodiment of the present disclosure, when the first information matrix is constructed, the following manner can also be used, but is not limited to: determining a first receiving state of a first collection target in a second preset time period, wherein the first collection target is any collection target in the multiple collection targets, the receiving state includes the first receiving state, and the second preset time period is obtained by equally dividing the first preset time period; determining a first matrix element value corresponding to each of the multiple second preset time periods according to the first receiving state, and performing column matrix construction processing according to the multiple first matrix element values to obtain a first matrix column corresponding to the first collection target; repeating the above steps until the matrix columns corresponding to the multiple collection targets are constructed, and performing data merging processing on the multiple matrix columns to obtain the first information matrix, wherein the number of rows of the first information matrix is the second time period number obtained by dividing the first preset time period into the second preset time periods.

[0062] In the embodiment of the present disclosure, the second preset time period is a more fine-grained time unit obtained by equally dividing the first preset time period. For example, the first preset time period of 24 hours is evenly divided into fixed-length segments, and if the second preset time period is defined as 5 minutes, a single day is divided into 288 continuous and equal-length second preset time periods. By equally dividing, a discretized time coordinate grid is established, and the continuous time flow is converted into a calculable discrete sequence.

[0063] The first collection target refers to a single user equipment selected from the multiple collection targets. The receiving state of the signaling data represents whether there is effective network activity of the user in a specific second preset time period. When the signaling data (such as a screen-on event, a data transmission request, etc.) of the first collection target is captured in a 5-minute unit, it is determined as an effective receiving state; if no signaling is detected, it is an invalid receiving state. The first receiving state refers to the state determination result of the first collection target in a single second preset time period.

[0064] The generation rule of the first matrix element value includes but is not limited to: mapping the first receiving state into a binary value. The valid receiving state corresponds to the value 1, and the invalid receiving state corresponds to the value 0. The mapping process is essentially to abstract the signal existence of the physical layer into a binary variable that can be processed mathematically. For example: if the user generates signaling in the period of 08:00-08:05, the corresponding second preset time period element value is recorded as 1; if there is no signaling in the period of 08:05-08:10, it is recorded as 0.

[0065] The process of column matrix construction processing includes but is not limited to: arranging the first matrix element values of the first collection target in the plurality of (288) second preset time periods in time sequence to form a longitudinal matrix with 288 rows and 1 column, i.e., the first matrix column. The first matrix column can be regarded as a time sequence vector of user behavior in mathematics, and the row index strictly corresponds to the time unit sequence number (for example, the first row represents 00:00-00:05, and the second row represents 00:05-00:10), and the column index is bound to the user (first collection target) identifier.

[0066] Specifically, in the time dimension, every 5 minutes is a basic unit, and the basic unit can also be defined as a time length of a finer granularity, such as 1 minute or 2 minutes, according to actual needs. When the user has signaling arrival in this time period, the value is 1, otherwise, the value is 0; 24 hours a day, 1440 minutes, which is equivalent to 288 units, at this time, the first matrix column can be represented by but not limited to the following way:

[0067]

[0068] The data merging processing refers to horizontally splicing the first matrix columns of a plurality of collection targets. Specifically, it includes but is not limited to: traversing all collection targets, independently generating the first matrix column of each collection target, and then arranging and combining the single-column matrices from left to right according to the user identifier sequence, and finally forming a first information matrix with 288 rows (consistent with the number of second preset time periods) and N columns (N is the total number of collection targets). The element value in the i-th row and the j-th column of the matrix uniquely represents the signaling state (0 or 1) of the j-th user in the i-th 5-minute unit of the day. Through the row-column structure, a two-dimensional mapping relationship of the time axis and the user set is established in the space dimension, so that the subsequent operation can directly locate the behavior characteristics of a specific user in a specific period through the matrix coordinates.

[0069] The disclosure realizes the structured reorganization of the original signaling data through the equal division of time units and the state binaryzation, and the effects include but are not limited to: converting the fragmented signaling events into regular mathematical matrices, eliminating the irregularity of time series; the binary numerical value representation significantly reduces the data storage complexity; the row and column index mechanism provides an efficient access path for subsequent statistical analysis by user (column direction) or by time period (row direction); the standardized matrix format can directly adapt to the parallel computing architecture of GPU / NPU, avoiding the parsing overhead of traditional text or log format.

[0070] In an implementable manner of an embodiment of the disclosure, when determining the first matrix element value, the following manner can be adopted but is not limited thereto: in a case where it is determined according to the first receiving state that the first collection target receives signaling data in a target second preset time period, a target first matrix element value corresponding to the target second preset time period is configured as a first numerical value, wherein the target second preset time period is any second preset time period in a plurality of second preset time periods; in a case where it is determined according to the first receiving state that the first collection target does not receive signaling data in the target second preset time period, the target first matrix element value is configured as a second numerical value.

[0071] In an embodiment of the disclosure, by binary state mapping, the continuity of the original signaling is converted into discrete numerical representation, creating a prerequisite for subsequent matrix operation and image conversion, the binary numerical system significantly reduces data complexity, avoiding the calculation redundancy brought by traditional multi-value coding (such as signal strength quantization); the Boolean operation of 0 / 1 logic and hardware circuit is essentially compatible, and the high-speed state writing can be directly realized by using the processor bit operation instruction; the unified numerical framework eliminates the heterogeneity of different user equipment signaling protocols, ensuring the standardization of the matrix data format. Through the minimalist coding strategy, an expandability foundation is laid for the subsequent parallel processing of massive data.

[0072] In an implementable manner of an embodiment of the disclosure, when converting the second information matrix into the target gray scale image, the following manner can be adopted but is not limited thereto: in a case where the first time period quantity is equal to a preset quantity threshold, each matrix element value of the second information matrix is determined as a gray scale value of each pixel point, to obtain the target gray scale image; in a case where the first time period quantity is less than the preset quantity threshold, all matrix element values of the second information matrix are normalized to obtain respective normalized element values of each matrix element value, and each normalized element value is determined as a gray scale value of each pixel point, to obtain the target gray scale image; in a case where the first time period quantity is greater than the preset quantity threshold, all matrix element values of the second information matrix are subjected to element elimination processing to obtain a plurality of processed element values, and the plurality of processed element values are determined as the gray scale values of each pixel point, to obtain the target gray scale image.

[0073] In the embodiments of the present disclosure, the preset quantity threshold is a self-defined parameter, for example, 255. The preset quantity threshold corresponds to the maximum value of the brightness of the pixels of the grayscale image (pure white), and the setting is based on the standard numerical range (0-255) of the 8-bit grayscale image in computer vision. The matrix element value of the second information matrix is the historical signaling active frequency of the user in each time unit (the value range is 0-m, and m is the number of the first preset time period participating in superposition).

[0074] The grayscale conversion processing is executed in three scenarios according to the comparison relationship between the first time period number m (that is, the number of historical data days) and the preset quantity threshold 255.

[0075] Scenario one: the data scale is equal to the threshold (m=255)

[0076] At this time, the second information matrix element value is directly used as the grayscale value. Because the matrix element value domain [0, 255] is completely overlapped with the pixel grayscale value domain, each matrix element value can be used as the brightness value of the corresponding pixel point in the grayscale image without conversion. For example, if a matrix element value is 180, the pixel point in the generated grayscale image presents a medium brightness (180 / 255≈70.6% brightness).

[0077] Scenario two: the data scale is less than the threshold (m<255)

[0078] The matrix element value is linearly scaled to the interval [0, 255] through normalization processing to obtain a normalized element value, and the normalized element value is used as the pixel grayscale value. The influence of the data scale difference on the brightness representation is eliminated, so that the grayscale images generated in different periods are comparable.

[0079] Scenario three: the data scale is greater than the threshold (m>255)

[0080] Element elimination processing is performed: a sliding window mechanism with a length of 255 is used to update the second information matrix. The window always retains the first information matrix of the most recent 255 days, and when new data is added to the right end of the window, the oldest data matrix at the left end of the window is removed at the same time. The processed element value obtained after the processing is the superposition value of the 255-day data under the new window, and the value range is restored to [0, 255], which is then directly used as the pixel grayscale value. For example, after the window eliminates 10-day old data, a matrix element value decreases from 260 to 250, which is directly used as the grayscale value.

[0081] The generation of the target grayscale image is essentially to establish a spatial mapping of the matrix coordinates to the image coordinates, and the row index (288 time units) of the second information matrix corresponds to the image row coordinates, and the column index (N users) corresponds to the image column coordinates. Finally, a single-channel grayscale image is output, the width of the target grayscale image is the number of users N, the height is 288 pixels, and the brightness of each pixel is determined by one of the three scenarios.

[0082] The disclosure solves the interference problem of data scale change on gray representation by dynamic adaptation strategy. In the initial stage of data accumulation (m<255), the normalization processing ensures the reasonable distribution of image brightness; when the data reaches the ideal scale (m=255), the zero conversion loss preserves the original statistical characteristics; in the face of continuously growing data (m>255), the sliding window mechanism automatically filters historical noise, making the gray image focus on recent high-value behavior patterns. Through adaptive conversion, it ensures that the subsequent image recognition link is always based on the most time-effective feature expression, providing a stable and reliable input source for hot user identification.

[0083] In one implementation manner of the embodiment of the disclosure, when the normalization processing is performed on all matrix element values of the second information matrix, the following manner can be used but is not limited to: dividing the preset number threshold by the number of the first time period to obtain a normalization coefficient; multiplying all matrix element values of the second information matrix by the normalization coefficient respectively to obtain the normalization element value corresponding to each matrix element value.

[0084] In the embodiment of the disclosure, the normalization processing is mainly applied to the scene where the number of historical data days (the number of the first time period m) is less than the preset number threshold (255), and the goal is to linearly map the original statistical frequency to the standard gray interval, solving the brightness representation distortion problem caused by insufficient data accumulation.

[0085] The normalization coefficient, as a scaling factor, is generated by division operation, including but not limited to: taking the ratio of the preset number threshold 255 to the number of the first time period m, that is, k=255 / m. The normalization coefficient represents the gray increment corresponding to the unit historical frequency in mathematics. For example: when m=100 days, k=255 / 100=2.55, which means that for every 1 day of active record of a user, the corresponding pixel brightness is increased by 2.55 gray units. The calculation of the normalization coefficient can be completed once in the initial stage, avoiding repeated element-by-element operation in the future.

[0086] The product processing refers to multiplying each matrix element value ω (ω∈[0,m]) of the second information matrix by the normalization coefficient k: Ω=ω×k, where Ω is the output normalization element value. Since ω≤m, Ω≤255, ensuring that the result strictly falls within the valid interval of the gray value. For example: a certain matrix element value ω=80, when m=100, the normalization coefficient k=2.55, then Ω=80×2.55=204. This value is directly used as the gray value of the corresponding pixel point in the target gray image.

[0087] Specifically, when the accumulated user information is less than 255 days, the value of each element of the user signaling information matrix is 0-m, not 0-255. At this time, the normalization can be performed by using but not limited to formula (1):

[0088]

[0089] Where Ω is the normalized element value, ω is the matrix element value, m is the number of elements in the first time period, and p is the preset quantity threshold.

[0090] This disclosure employs a linear scaling strategy to ensure consistent brightness scales for grayscale images across different data accumulation stages. By introducing a normalization coefficient k, the data scale m is transformed into a computable physical quantity, eliminating hotspot misjudgments caused by historical time variations. The ergodicity of the product operation ensures distortion-free transformation of the statistical characteristics of each user behavior pattern. A unified output value range (0-255) guarantees the stability of the input format for subsequent image recognition algorithms. Its lightweight computation makes it particularly suitable for initial system deployments or scenarios with dynamically changing user scales, providing a robust data preprocessing foundation for hotspot user identification.

[0091] In one possible implementation of this disclosure, when performing element removal processing on all matrix element values ​​of the second information matrix, it can also be implemented in the following ways, but is not limited to: sorting multiple first preset time periods in chronological order to obtain a time period sequence, and selecting a first preset time period with a preset quantity threshold as the selected time period in the time period sequence in chronological order from late to early; subtracting the second information matrix from the information matrices to be removed corresponding to each of the multiple time periods to be removed to obtain a target information matrix, wherein the time periods to be removed are the first preset time periods in the time period sequence excluding the selected time periods, and the information matrices to be removed are the first information matrices corresponding to the time periods to be removed; and determining each target matrix element value of the target information matrix as the processed element value.

[0092] In the embodiments of this disclosure, the element removal process is triggered when the number of historical data days (the number of first time periods m) exceeds a preset threshold (255). The data window is dynamically updated through time-series filtering and matrix operations to ensure that the grayscale image only reflects recent user behavior characteristics.

[0093] The construction of the time period sequence is based on a strict chronological order, including but not limited to: arranging the m first preset time periods (each a complete natural day) that participate in generating the second information matrix from earliest to latest according to their actual occurrence dates, forming a continuous historical timeline. For example: the sequence [D1, D2, ..., D...]. m ] represents the data from day 1 to day m, where D m The most recent day.

[0094] The extraction rule for the selected time period is as follows: select 255 days (preset threshold) in reverse order from the end of the time period sequence (i.e., the latest end of the time sequence). Specific operations include, but are not limited to: locating the D at the end of the sequence. m(the latest date); take 254 adjacent dates in succession, and construct a subset [D m-254 , D m-253 , ..., D m ]; this subset is the selected time period. This process is essentially to retain the data window of the last 255 days, which conforms to the business logic that new data is more important than old data.

[0095] The time period to be removed refers to the set of dates in the time period sequence that are not selected. When m > 255, the number of time periods to be removed is m-255, that is, the expired data is removed in succession from the beginning of the sequence (the earliest date). For example, when m = 300, the time period to be removed is the first 45 days (D 45 ).

[0096] The to-be-removed information matrix refers to the original first information matrix (binary day matrix) corresponding to the to-be-removed time period. Each to-be-removed time period has and only has one such matrix, and the row and column structures are completely consistent with the current second information matrix (288 rows x N columns).

[0097] The subtraction process is the reverse of the superposition process, that is, all to-be-removed information matrices corresponding to the to-be-removed time period are subtracted from the current second information matrix (generated by m days of data superposition) day by day. This operation is equivalent to removing the contribution of expired data to the statistical value. For example, the historical active value of a user in a certain time unit is originally 280 (that is, there is signaling for 280 days), and the value at this position in 10 to-be-removed information matrices is removed (if 10 days are all active, then reduce 10), and the result is reduced to 270.

[0098] The processed element value is the target matrix element value in the target information matrix. Each target matrix element value represents the number of signaling active days of the corresponding user in the corresponding time unit within the selected time period (the last 255 days). Because the length of the time window is always 255, the value range is also returned to [0, 255], which can be directly used as a gray value.

[0099] Specifically, in order to facilitate understanding of the embodiments of the present disclosure, the embodiments of the present disclosure also provide a principle diagram of element removal, as shown in Figure 4 When the accumulated users reach 255 days, although they can be converted by the above formula, considering that old historical data has weaker representativeness than new data, old historical data needs to be removed gradually, so a sliding window with a length of 255 is selected to replace new and old data, thereby ensuring the effectiveness of the data.

[0100] First step: when new data arrives, add it to the rightmost side of the sliding window;

[0101] Second step: determine whether the total number of data in the window exceeds 255;

[0102] Third step: if more than 255, then the leftmost data of the window is discarded.

[0103] The present disclosure realizes dynamic maintenance of data window through timing screening and matrix difference, and solves the problem of historical data overload in long-term operation. The specific effects include but are not limited to: time sequence reverse sequence selection ensures that the latest data is losslessly reserved; matrix subtraction accurately eliminates the influence of expired data, avoiding the resource overhead of full recalculation; the element value after processing automatically adapts to the gray value range, ensuring the stability of the input of the image recognition link. The system always maintains high sensitivity to recent user behavior patterns in continuous operation, providing a continuous and effective input source for hot user identification.

[0104] In an implementable manner of an embodiment of the present disclosure, when determining the selected target, the following manner can also be used, but is not limited to: inputting the target gray image into a preset information recognition algorithm for image recognition processing to obtain bright pixel points, wherein the bright pixel points are pixel points with a gray value greater than a preset gray threshold; determining, based on the information recognition algorithm, a pixel point column in which each bright pixel point is located as a target pixel point column, and determining the collection target corresponding to the target pixel point column as the selected target.

[0105] In an embodiment of the present disclosure, the preset information recognition algorithm refers to a general image processing program deployed on a GPU / NPU acceleration platform. This algorithm does not depend on a specific neural network architecture, but refers to a computer vision routine that can perform pixel-level brightness analysis. Its basic function is to traverse each pixel point of the input image and output a feature label according to a preset rule.

[0106] The operation flow of image recognition processing is defined as follows:

[0107] Receive the target gray image as input. The image is a single-channel matrix with a width of N users and a height of 288 pixels.

[0108] Compare the gray value with the preset gray threshold (configurable parameter, typical value is 128) pixel by pixel.

[0109] If the pixel gray value is greater than the threshold, it is marked as a bright pixel point; otherwise, it is marked as a dark area pixel.

[0110] For example, when the threshold is 150, the pixel with a gray value of 200 is identified as a bright pixel point, while the pixel with a gray value of 100 is filtered. This process is a variant of binary segmentation, which discretizes the continuous gray space into two active regions of high / low.

[0111] A pixel column refers to a set of pixels aligned in the vertical direction in an image. In the coordinate system of the target gray image, each column of pixels (from the 1st row to the 288th row) corresponds to a collection target (i.e., a user equipment) in a strict manner, and the column binding relationship is determined in advance by the matrix construction process: the user identifier is mapped to the image column index in a fixed order.

[0112] The determination rule of the target pixel column is that if there is at least one bright pixel in a column, the entire column is marked as a target column. For example, in the jth column corresponding to user A, the 50th row pixel gray value is 210 (> threshold), regardless of the brightness of the remaining pixels in the column, the jth column is determined to be a target pixel column. This loose strategy ensures that a user is included in the candidate set as long as he is active at any historical time unit.

[0113] The generation of the selected target is achieved by reverse mapping of the column index: according to the image horizontal coordinate j of the target pixel column, the jth collection target in the initial user identifier sequence is retrieved and marked as the selected target. For example, when the 5th column is the target column, the 5th registered user in the system user ID list is selected. This process establishes a deterministic conversion chain between the image column coordinates and the user database index.

[0114] In order to facilitate understanding, the disclosure embodiment provides a configuration platform diagram of a preset information recognition algorithm, as shown in Figure 5 , and an acquisition flowchart of information data corresponding to the selected target, as shown in Figure 6 , wherein the cloud platform with GPU+NPU+CPU processes the gray image, the artificial intelligence algorithm is deployed on the platform as support, the computer vision and neural network program processes the gray image, and the relatively bright points in the gray image are identified at a high speed:

[0115] The U-Net neural network can be used to process the gray image, and specifically, the following methods can be used but are not limited to:

[0116] Input layer: gray image of (H, W, 1).

[0117] Encoder (downsampling): 4-5 downsampling stages, each stage containing: two 3x3 convolutions + batch normalization (Batch Normalization, BN) + rectified linear unit (Rectified Linear Unit, ReLU) and 2x2 maximum pooling, output high-level semantic features.

[0118] Decoder (upsampling): each stage is upsampled by transposed convolution or interpolation, and the features of the corresponding layer of the encoder are spliced (jump connection), two 3x3 convolutions + BN + ReLU.

[0119] Output layer: 1x1 convolution + Sigmoid activation function.

[0120] After obtaining the relatively bright point in the image, the column where the point is located corresponds to the user, i.e., the user of the relatively hot point.

[0121] In summary, the embodiments of the present disclosure can achieve the following technical effects:

[0122] 1. The present disclosure constructs the signaling receiving state of each collection target in multiple preset time periods into a time sequence information matrix, and superimposes a second information matrix reflecting long-term heat; then converts the matrix element value into a gray image pixel value to generate a visual gray image columned by the collection target; based on a preset gray threshold, the target corresponding to the highlighted column in the image is accurately identified, thereby intelligently selecting the high-frequency active user as the selected target, and the selected target is taken as the hot short message data subject, which can accurately determine the hot data, avoid invalid data exchange between the high-speed cache and the low-speed cache, ensure that the high-value data stably resides in the high-speed cache, significantly improve the matching efficiency and real-time performance of the short message data and the signaling data, while ensuring the accurate reach of important short messages, reduce the system computing resource consumption and cache cost.

[0123] 2. The present disclosure calculates based on the computing power of CPU+GPU+NPU, and fully utilizes the computing power advantage of new technology.

[0124] 3. The present disclosure processes the gray image after data conversion, which can convert the processing flow into a computer vision problem based on neural network artificial intelligence technology, reducing the complexity of data processing.

[0125] Corresponding to the above-mentioned signaling matching method based on intelligent identification, the present disclosure also proposes a signaling matching device based on intelligent identification. Since the device embodiment of the present disclosure corresponds to the above-mentioned method embodiment, for the details not disclosed in the device embodiment, the above-mentioned method embodiment can be referred to, and the present disclosure will not be described in detail.

[0126] Figure 7 A structure diagram of a signaling matching device based on intelligent identification provided by the present disclosure is shown in Figure 7 , which includes:

[0127] The construction unit 71 is configured to perform matrix construction processing according to the receiving state of the signaling data of each of the multiple collection targets in the first preset time period, to obtain a first information matrix corresponding to the first preset time period.

[0128] The calculation unit 72 is configured to add the first information matrix corresponding to each of the multiple first preset time periods to obtain a second information matrix, wherein the maximum matrix element value in the second information matrix is not greater than the number of first time periods of the multiple first preset time periods.

[0129] The conversion unit 73 is configured to perform a grayscale map conversion process on the first time period quantity and the second information matrix to obtain a target grayscale map corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a grayscale value of each pixel point in the target grayscale map, and a plurality of pixel point columns of the target grayscale map each correspond to a plurality of collection targets.

[0130] The determination unit 74 is configured to determine a pixel point column in which a pixel point with a grayscale value greater than a preset grayscale threshold value is located in the target grayscale map as a target pixel point column, and determine a collection target corresponding to the target pixel point column as a selected target.

[0131] The matching unit 75 is configured to store information data corresponding to the selected target in a cache for signaling matching processing.

[0132] Further, in a possible implementation manner of the embodiment of the present disclosure, the construction unit 71 is further configured to:

[0133] determine a first reception state of the first collection target in a second preset time period, wherein the first collection target is any collection target in the plurality of collection targets, the reception state includes the first reception state, and the second preset time period is obtained by averaging division of the first preset time period;

[0134] determine a first matrix element value corresponding to each of the plurality of second preset time periods according to the first reception state, and perform column matrix construction processing according to the plurality of first matrix element values to obtain a first matrix column corresponding to the first collection target.

[0135] repeat the above steps until a matrix column corresponding to each of the plurality of collection targets is constructed, and perform data merging processing on the plurality of matrix columns to obtain the first information matrix, wherein a row quantity of the first information matrix is a second time period quantity obtained by dividing the first preset time period into the second preset time periods.

[0136] Further, in a possible implementation manner of the embodiment of the present disclosure, the construction unit 71 is further configured to:

[0137] in a case where it is determined according to the first reception state that the first collection target receives signaling data in a target second preset time period, configure a target first matrix element value corresponding to the target second preset time period as a first numerical value, wherein the target second preset time period is any second preset time period in the plurality of second preset time periods;

[0138] in a case where it is determined according to the first reception state that the first collection target does not receive signaling data in the target second preset time period, configure the target first matrix element value as a second numerical value.

[0139] Further, in a possible implementation manner of the embodiment of the present disclosure, the conversion unit 73 is further configured to:

[0140] In a case where the number of the first time periods is equal to the preset number threshold, each matrix element value of the second information matrix is determined as a gray value of each pixel point, to obtain the target gray map;

[0141] In a case where the number of the first time periods is less than the preset number threshold, all matrix element values of the second information matrix are normalized to obtain a respective normalized element value corresponding to each matrix element value, and each normalized element value is determined as a gray value of each pixel point, to obtain the target gray map;

[0142] In a case where the number of the first time periods is greater than the preset number threshold, all matrix element values of the second information matrix are element-eliminated to obtain a plurality of processed element values, and the plurality of processed element values are determined as gray values of each pixel point, to obtain the target gray map.

[0143] Further, in a possible implementation of the embodiment of the present disclosure, the conversion unit 73 is further configured to:

[0144] divide the preset number threshold by the number of the first time periods to obtain a normalization coefficient;

[0145] multiply all matrix element values of the second information matrix by the normalization coefficient respectively to obtain a respective normalized element value corresponding to each matrix element value.

[0146] Further, in a possible implementation of the embodiment of the present disclosure, the conversion unit 73 is further configured to:

[0147] sort the plurality of first preset time periods according to a time sequence to obtain a time period sequence, and select a first preset time period of the preset number threshold as a selected time period in the time period sequence according to a time sequence from late to early;

[0148] subtract the second information matrix from a respective to-be-eliminated information matrix corresponding to each to-be-eliminated time period to obtain a target information matrix, wherein the to-be-eliminated time period is a first preset time period in the time period sequence except the selected time period, and the to-be-eliminated information matrix is a first information matrix corresponding to the to-be-eliminated time period;

[0149] determine each target matrix element value of the target information matrix as a processed element value.

[0150] Further, in a possible implementation of the embodiment of the present disclosure, the determination unit 74 is further configured to:

[0151] input the target gray map into a preset information recognition algorithm for image recognition processing to obtain a bright pixel point, wherein the bright pixel point is a pixel point with a gray value greater than a preset gray threshold.

[0152] Based on the information recognition algorithm, the pixel column where each bright pixel is located is determined as the target pixel column, and the acquisition target corresponding to the target pixel column is determined as the selected target.

[0153] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of the embodiments of this disclosure, and the principle is the same. Therefore, the embodiments of this disclosure are not limited thereto.

[0154] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0155] Figure 8 A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0156] like Figure 8 As shown, device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in ROM (Read-Only Memory) 802 or a computer program loaded from storage unit 808 into RAM (Random Access Memory) 803. RAM 803 can also store various programs and data required for the operation of device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via bus 804. I / O (Input / Output) interface 805 is also connected to bus 804.

[0157] Multiple components in device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of monitors, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0158] The computing unit 801 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a CPU (Central Processing Unit), a GPU (Graphic Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, a DSP (Digital Signal Processor), and any appropriate processor, controller, microcontroller, etc. The computing unit 801 performs various methods and processes described above, such as the intelligent recognition based signaling matching method. For example, in some embodiments, the intelligent recognition based signaling matching method can be implemented as a computer software program, which is tangibly embodied in a machine-readable medium, such as the storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded onto the RAM 803 and executed by the computing unit 801, one or more steps of the methods described above can be performed. Alternatively, in other embodiments, the computing unit 801 can be configured to perform the aforementioned intelligent recognition based signaling matching method by any other appropriate means, such as by means of firmware.

[0159] Various implementations of the systems and techniques described above herein can be realized in digital electronic circuitry, integrated circuitry, a Field Programmable Gate Array (FPGA), an Application-Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a System on Chip (SOC), a Complex Programmable Logic Device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0160] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0161] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable storage medium can include but are not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium would include one or more lines of electrical wire, portable computer diskette, hard disk, RAM, ROM, EPROM (Electrically Programmable Read-Only-Memory), or flash memory, fiber optics, CD-ROM (Compact Disc Read-Only Memory), optical storage device, magnetic storage device, or any suitable combination of the foregoing.

[0162] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (Cathode Ray Tube) or LCD (Liquid Crystal Display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0163] The systems and techniques described herein can be implemented in a computing system that includes a back end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front end component, e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described herein, or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, and a blockchain network.

[0164] The computer system can include clients and servers. The clients and servers are generally remote from each other and typically interact through a communication network. The relationship of client and server is one of communication and distribution, with the server receiving requests from the client and transmitting responses via the communication network. The client and server can be distinguished from each other in that the server is a file and data management system that provides requested files and / or data to the client. The server transmits data, e.g., data files, to the client, and the client processes the received data into an output format usable by the client. The server and client communicate using any of a number of known protocols, such as the hypertext transfer protocol (HTTP), the simple mail transfer protocol (SMTP), or the file transfer protocol (FTP). Other protocols can also be used.

[0165] It should be noted that artificial intelligence is a discipline that studies enabling computers to simulate some thinking processes and intelligent behaviors of people (such as learning, reasoning, thinking, planning, etc.), which has both hardware and software technologies. Artificial intelligence hardware technology generally includes technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing, etc.; artificial intelligence software technology mainly includes computer vision technology, speech recognition technology, natural language processing technology, and machine learning / deep learning, big data processing technology, knowledge graph technology, etc. several major directions.

[0166] It should be understood that the various forms of flow shown above can be used to reorder, add or delete steps. For example, each step described in the present disclosure can be executed in parallel, sequentially or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, which is not limited herein.

[0167] The above detailed description does not limit the scope of the disclosure. Various modifications, combinations, sub-combinations and alternatives can be made to the detailed description. Any modification, equivalent replacement and improvement etc. made within the spirit and principle of the disclosure shall be included in the scope of the disclosure.

Claims

1. A signaling matching method based on intelligent identification, characterized in that, The method comprises the following steps: According to the receiving state of the signaling data corresponding to each of the plurality of collection targets within the first preset time period, matrix construction processing is performed to obtain a first information matrix corresponding to the first preset time period; The first information matrix corresponding to each of the plurality of first preset time periods is added to obtain a second information matrix, wherein the maximum matrix element value in the second information matrix is not greater than the first time period number of the plurality of first preset time periods; According to the first time period number and the second information matrix, gray scale graph conversion processing is performed to obtain a target gray scale graph corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a gray scale value of each pixel point in the target gray scale graph, and each pixel point column of the target gray scale graph corresponds to the plurality of collection targets; In the target gray scale graph, the pixel point column in which the pixel point with a gray scale value greater than a preset gray scale threshold is determined as a target pixel point column, and the collection target corresponding to the target pixel point column is determined as a selected target; The information data corresponding to the selected target is stored in the cache for signaling matching processing. 2.The smart recognition based signaling matching method of claim 1, wherein, The method comprises the following steps: Determine the first receiving state of the first collection target within the second preset time period, wherein the first collection target is any collection target in the plurality of collection targets, the receiving state includes the first receiving state, and the second preset time period is obtained by averaging the first preset time period; According to the first receiving state, determine the first matrix element value corresponding to each of the plurality of second preset time periods, and perform column matrix construction processing according to the plurality of first matrix element values to obtain a first matrix column corresponding to the first collection target; Repeat the above steps until the matrix column corresponding to each of the plurality of collection targets is constructed, and perform data merging processing on the plurality of matrix columns to obtain the first information matrix, wherein the row number of the first information matrix is the second time period number obtained by dividing the first preset time period into the second preset time period. 3.The smart recognition based signaling matching method of claim 2, wherein, The method comprises the following steps: In the case where it is determined according to the first receiving state that the first collection target receives the signaling data within the target second preset time period, the target first matrix element value corresponding to the target second preset time period is configured as a first value, wherein the target second preset time period is any second preset time period in the plurality of second preset time periods; In the case where it is determined according to the first receiving state that the first collection target does not receive the signaling data within the target second preset time period, the target first matrix element value is configured as a second value. 4.The smart recognition based signaling matching method of claim 1, wherein, The method comprises the following steps: In a case where the first time period quantity is equal to a preset quantity threshold, each matrix element value of the second information matrix is determined as the gray value of each pixel point, and the target gray scale diagram is obtained; In a case where the first time period quantity is less than the preset quantity threshold, all the matrix element values of the second information matrix are normalized to obtain a corresponding normalized element value of each matrix element value, and each normalized element value is determined as the gray value of each pixel point, and the target gray scale diagram is obtained; In a case where the first time period quantity is greater than the preset quantity threshold, all the matrix element values of the second information matrix are element-eliminated to obtain a plurality of processed element values, and the plurality of processed element values are determined as the gray value of each pixel point, and the target gray scale diagram is obtained. 5.The smart recognition based signaling matching method of claim 4, wherein, In a case where the first time period quantity is less than the preset quantity threshold, all the matrix element values of the second information matrix are normalized to obtain a corresponding normalized element value of each matrix element value, and each normalized element value is determined as the gray value of each pixel point, and the target gray scale diagram is obtained. The preset quantity threshold is divided by the first time period quantity to obtain a normalization coefficient; The second information matrix and a plurality of to-be-eliminated information matrices corresponding to a plurality of to-be-eliminated time periods are subtracted to obtain a target information matrix, wherein the to-be-eliminated time period is a first preset time period in the time period sequence except the selected time period, and the to-be-eliminated information matrix is a first information matrix corresponding to the to-be-eliminated time period; 6.The smart recognition based signaling matching method of claim 4, wherein, Each target matrix element value of the target information matrix is determined as the processed element value. In the target gray scale diagram, the pixel point column in which the pixel point with a gray value greater than a preset gray scale threshold is determined as a target pixel point column, and a collection target corresponding to the target pixel point column is determined as a selected target, comprising: The target gray scale diagram is input into a preset information recognition algorithm for image recognition processing to obtain bright pixel points, wherein the bright pixel points are pixel points with a gray value greater than a preset gray scale threshold; Based on the information recognition algorithm, the pixel point column in which each bright pixel point is located is determined as a target pixel point column, and a collection target corresponding to the target pixel point column is determined as a selected target. 7.The smart recognition based signaling matching method of claim 1, wherein, Comprising: ​ ​ 8. An intelligent recognition based signaling matching device, characterized by, ​ The construction unit is configured to perform matrix construction processing on the reception states of the signaling data corresponding to the plurality of collection targets respectively in a first preset time period, to obtain a first information matrix corresponding to the first preset time period. The calculation unit is configured to perform addition processing on the first information matrices corresponding to the plurality of first preset time periods respectively, to obtain a second information matrix, wherein a maximum matrix element value in the second information matrix is not greater than a first time period number of the plurality of first preset time periods. The conversion unit is configured to perform gray scale map conversion processing on the first time period number and the second information matrix, to obtain a target gray scale map corresponding to the second information matrix, wherein each matrix element value in the second information matrix is converted into a gray scale value of each pixel point in the target gray scale map, and a plurality of pixel point columns of the target gray scale map correspond to the plurality of collection targets respectively. The determination unit is configured to determine a pixel point column in which a pixel point with a gray scale value greater than a preset gray scale threshold in the target gray scale map as a target pixel point column, and determine a collection target corresponding to the target pixel point column as a selected target. The matching unit is configured to store information data corresponding to the selected target into a cache for signaling matching processing.

9. An electronic device, comprising: comprise: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A computer program product, characterised in that, comprise a computer program, which, when executed by a processor, implements the method according to any one of claims 1-7.

Images