Intelligent service method and system

Abstract

The invention relates to the technical field of intelligent service platforms, and particularly discloses an intelligent service method and system, and the method comprises the steps: obtaining the real-time personnel information of each service unit based on an information collector installed in a public area, and constructing a crowd portrait of the service unit; determining a service demand table according to the crowd portrait; the service demand table comprises service types and service quotas; a service request sent by the service unit is received in real time, a service instruction is generated, the service request is analyzed, and the generation process of the service demand table is updated; and periodically pre-scheduling the service resources according to the service demand tables of all the service units. According to the method, the demand of the customer is predicted according to the available public customer information, the service resources are prepared in advance according to the prediction result, compared with a traditional scheme, the preparation process is advanced, and the labor cost is improved, but the service efficiency is greatly improved from the perspective of customers.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent service platforms, and in particular to an intelligent service method and system. Background Art

[0002] Intelligent services refer to the use of technologies such as artificial intelligence (AI), big data, the Internet of Things (IoT), and cloud computing to improve the automation, personalization, and intelligence of services, thereby meeting the diverse needs of users and improving user experience and service efficiency.

[0003] Intelligent services are widely used in the service industry. For example, hotels, KTVs and restaurants are equipped with intelligent systems to receive demand information sent by customers remotely and then provide services quickly. However, the existing service architecture does not have advance communication. Services can only be provided after the customer sends a request, which causes a certain delay. How to improve service efficiency is the technical problem that the technical solution of the present invention aims to solve. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent service method and system to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An intelligent service method, comprising:

[0007] Based on the information collectors installed in public areas, real-time personnel information of each service unit is obtained to build a crowd portrait of the service unit;

[0008] Determine a service demand table based on the crowd portrait; the service demand table includes service type and service amount;

[0009] The process of receiving service requests sent by service units in real time, generating service instructions, analyzing the service requests, and updating the service demand table;

[0010] Regularly pre-schedule service resources according to the service demand tables of all service units;

[0011] The crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

[0012] Furthermore, the step of acquiring real-time personnel information of each service unit based on information collectors installed in public areas and constructing a crowd portrait of the service unit includes:

[0013] For any service unit, the usage information of the service unit is obtained according to the information input port; the usage information includes the number of users and the usage time period;

[0014] Obtaining personal characteristics of people entering and leaving the service unit based on cameras installed in public areas;

[0015] According to the entry and exit situations corresponding to each personnel characteristic, the personnel characteristics of the service unit during the use period are counted to obtain a crowd portrait.

[0016] Furthermore, the step of obtaining the personal characteristics of people entering and leaving the service unit based on cameras installed in the public area includes:

[0017] Track people and locate areas with people based on cameras installed in public areas;

[0018] Performing contour recognition on the personnel area, and determining an upper garment area and a lower garment area according to the relative position of each contour in the personnel area;

[0019] Identify items in the tops and bottoms areas, query the mode price range of the identification results, and use the mode price range as a person feature;

[0020] The mode price range refers to the mode price range of all the recognition results.

[0021] Furthermore, the step of determining a service demand table based on the crowd portrait includes:

[0022] Read all the mode price ranges in the crowd portrait;

[0023] Determining the consumption characteristics of the population based on the mode price range;

[0024] Input the consumption characteristics of the crowd into the trained neural network model to obtain the service type and service amount, and construct a service demand table;

[0025] Among them, the training process of the neural network model is:

[0026] Select service units within a preset time and space range to obtain their population portraits and real consumption records;

[0027] Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges;

[0028] Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts;

[0029] Train the neural network model based on the sample set;

[0030] The process of determining the consumption characteristics of a population is as follows:

[0031]

[0032] In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α i represents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N},|S|=k represents the set {A1,A2,…,A N} Randomly select a subset of k mode value intervals, ∩ i∈S (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N},|S|=k} means to find the mean of the midpoint values of the intersection.

[0033] Furthermore, the steps of receiving a service request sent by a service unit in real time, generating a service instruction, analyzing the service request, and updating the service requirement table include:

[0034] Receive the service type and service quota sent by the service unit in real time, upload the service unit and its service type and service quota to the main control terminal, and generate service instructions;

[0035] The service type and service amount of the service unit are counted as real consumption records and the sample set is updated.

[0036] Furthermore, the step of pre-scheduling service resources according to the service demand tables of all service units at regular intervals includes:

[0037] Every preset period, obtain the service demand table of all service units in the previous period;

[0038] Combining the service units according to their positional relationships and merging the corresponding service demand tables;

[0039] Pre-schedule service resources according to the merged service demand table.

[0040] The technical solution of the present invention also provides an intelligent service system, which includes:

[0041] The crowd portrait construction module is used to obtain real-time personnel information of each service unit based on information collectors installed in public areas and construct crowd portraits of service units;

[0042] A service demand determination module is used to determine a service demand table based on the crowd portrait; the service demand table includes service types and service amounts;

[0043] A service instruction generation module is used to receive service requests sent by service units in real time, generate service instructions, analyze the service requests, and update the generation process of the service demand table;

[0044] Resource pre-scheduling module, used to pre-scheduling service resources according to the service demand tables of all service units at regular intervals;

[0045] The crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

[0046] Furthermore, the crowd portrait construction module includes:

[0047] A usage information query unit is used to obtain usage information of any service unit according to the information input port; the usage information includes the number of users and the usage time period;

[0048] A personnel feature detection unit, configured to obtain the personnel features of persons entering and exiting the service unit based on cameras installed in public areas;

[0049] The personnel feature statistics unit is used to count the personnel features of the service unit during the use period according to the entry and exit conditions corresponding to each personnel feature, and obtain a crowd portrait as the crowd portrait.

[0050] Furthermore, the personnel feature detection unit includes:

[0051] The personnel tracking subunit is used to track people and locate personnel areas based on cameras installed in public areas;

[0052] A contour recognition subunit, configured to perform contour recognition on the personnel area and determine an upper garment area and a lower garment area according to the relative position of each contour in the personnel area;

[0053] The item recognition subunit is used to recognize items in the top and bottom areas, query the mode price range of the recognition results, and use the mode price range as a person feature;

[0054] The mode price range refers to the mode price range of all the recognition results.

[0055] Furthermore, the service demand determination module includes:

[0056] Price range reading unit, used to read all the majority price ranges in the crowd portrait;

[0057] a consumption characteristics determination unit, configured to determine the consumption characteristics of a group of people based on the mode price range;

[0058] Construct an execution unit to input the consumption characteristics of the crowd into the trained neural network model, obtain the service type and service amount, and construct a service demand table;

[0059] Among them, the training process of the neural network model is:

[0060] Select service units within a preset time and space range to obtain their population portraits and real consumption records;

[0061] Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges;

[0062] Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts;

[0063] Train the neural network model based on the sample set;

[0064] The process of determining the consumption characteristics of a population is as follows:

[0065]

[0066] In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α i represents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N},|S|=k represents the set {A1,A2,…,A N} Randomly select a subset of k mode value intervals, ∩ i∈S (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N},|S|=k} means to find the mean of the midpoint values of the intersection.

[0067] Compared with the existing technology, the beneficial effects of the present invention are: the present invention predicts customer needs based on available public customer information, and prepares service resources in advance based on the prediction results. Compared with traditional solutions, the preparation process is advanced. Although it increases labor costs, from the customer's perspective, it greatly improves service efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention.

[0069] Figure 1 This is a flowchart of the intelligent service method.

[0070] Figure 2 This is the first sub-flow chart of the intelligent service method.

[0071] Figure 3 This is the second sub-flow chart of the intelligent service method.

[0072] Figure 4 This is the third sub-flow diagram of the intelligent service method.

[0073] Figure 5 This is the fourth sub-flow chart of the intelligent service method.

[0074] Figure 6 This is a structural diagram of the intelligent service system. DETAILED DESCRIPTION

[0075] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0076] Figure 1 This is a flowchart of an intelligent service method. In an embodiment of the present invention, an intelligent service method includes:

[0077] Step S100: acquiring real-time personnel information of each service unit based on information collectors installed in public areas, and constructing a crowd portrait of the service unit;

[0078] The technical solution of the present invention is applied to private entertainment service places, such as restaurants, hotels and KTVs that have separate private rooms. Information inside the private rooms cannot be obtained, and they belong to the customers' private areas. Areas such as corridors, halls and front desks are called public areas. Information collectors are installed in public areas to manage the security of the entire place. In existing places, information collectors are also infrastructure; the information collectors include information entry equipment installed at the front desk and cameras that fully cover the public areas.

[0079] It should be noted that the information collector in the public area itself has the collection authority, and what is obtained is the information in the public area, not personal information. Of course, the manager of the information collector cannot disclose the obtained information at will. There are relevant regulations. In the technical solution of the present invention, in order to clarify the authority issue, it is necessary to clearly inform the user of the location of the information collector. When the user still chooses to go to the public space, it is considered that he has explicitly given the information collection authority, but not the information dissemination authority.

[0080] Among them, the smallest service unit providing services is called a service unit. For a hotel, it is each room; for a restaurant, it is each private room; for a KTV, it is each private room. Based on the collection of public information by the information collector, we can obtain the number of people in each service unit and the characteristics of each person, which is called a crowd portrait.

[0081] It must be pointed out that the above content must be based on known and accessible data, such as the number of consumers entered (some involve specific identity information, such as hotels, which are all in compliance with regulations) and information on the movement of people in public areas, etc. This information must be data that is tacitly accessible by both parties, and information that customers refuse to be obtained cannot be obtained.

[0082] Step S200: determining a service demand table based on the crowd portrait; the service demand table includes service types and service amounts;

[0083] The obtained crowd portrait is used to reflect the demand characteristics of the crowd in the box. By identifying the crowd portrait, the service demand table can be determined. The service demand table includes service type and service quota. The service type is used to represent which service. The service quota is an upper probability, which represents the amount of service the user needs. For example, drinking water demand is a service type, and water consumption is the corresponding service quota. Since the numerical units of different service types are different, they are collectively referred to as service quotas.

[0084] Specifically, the mapping relationship between crowd portraits and service needs can be obtained with the help of neural network model training, or a table can be created based on historical data. When needed, the crowd portrait can be used to query the table.

[0085] It is worth mentioning that the simplest way to profile a population is a set consisting of ages, that is, a set consisting of the ages of all personnel in the service unit. This is a numerical group. On this basis, the parameter of gender can also be introduced. Both gender and age parameters will affect service demand. Count the combinations of different genders and ages, and then record the corresponding historical demands. After obtaining the record table, when someone uses the service unit, obtain the gender and age of the person, and query the record table for the closest historical demand, which can be used as a service demand table.

[0086] In summary, the crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

[0087] Step S300: receiving a service request sent by a service unit in real time, generating a service instruction, analyzing the service request, and updating the generation process of a service demand table;

[0088] Step S300 is the actual application process, which receives a service request sent by any service unit, generates a service instruction, provides services to it, and at the same time, records the provided services to obtain historical demands; historical demands are used to update the generation process of the service demand table; in the above solution, it is to update the record table.

[0089] Step S400: pre-scheduling service resources according to the service demand tables of all service units on a regular basis;

[0090] Service resources are pre-scheduled at preset intervals, such as one hour, which can greatly improve service speed (reduce resource preparation time). It should be noted that since customers are constantly changing, the population portrait of each service unit changes over time, and the corresponding service demand table also changes over time. Therefore, the service demand tables of all service units are different at preset intervals. Therefore, the plan for pre-scheduling service resources is not static, and the pre-scheduling plan at different times is different.

[0091] Specifically, the pre-scheduling process is to prepare service resources in advance based on the predicted demand of each service unit (forecast demand table). The accuracy of the predicted demand is relatively high, but it cannot reach 100%. Therefore, some advance preparation work will not be implemented, which increases the labor cost to a certain extent. However, from the customer's perspective, the service efficiency is extremely high and the service quality enjoyed is high.

[0092] Figure 2This is a first sub-flow diagram of the intelligent service method. The steps of acquiring real-time personnel information of each service unit based on information collectors installed in public areas and constructing a crowd portrait of the service unit include:

[0093] Step S101: For any service unit, obtain usage information of the service unit according to the information input port; the usage information includes the number of users and usage time period;

[0094] Step S102: acquiring personal characteristics of people entering and leaving the service unit based on cameras installed in public areas;

[0095] Step S103: The personnel characteristics of the service unit during the use period are counted according to the entry and exit situations corresponding to each personnel characteristic to obtain a crowd portrait.

[0096] In an example of the technical solution of the present invention, a specific description is given of the process of constructing a crowd portrait. For any service unit, the usage information of the service unit is obtained based on the information entry port. The usage information is how many people use it in what time period, which is called the number of users and usage time period; the information entry port is built into the main control terminal, which is generally a computer placed at the front desk in the service place.

[0097] By capturing real-time video from cameras installed in public areas and identifying the video, we can obtain the personal characteristics of people entering and leaving each service unit. We can then calculate the number of people inside the service unit during the usage period and count their personal characteristics, which is called crowd profiling.

[0098] Among them, the video recognition solution uses the existing target tracking algorithm to quickly determine whether a person is entering or leaving the service unit.

[0099] Specifically, the step of obtaining the personal characteristics of people entering and leaving the service unit based on cameras installed in the public area includes:

[0100] Track people and locate areas with people based on cameras installed in public areas;

[0101] Performing contour recognition on the personnel area, and determining the upper garment area and the lower garment area according to the relative position of each contour in the personnel area;

[0102] Identify items in the tops and bottoms areas, query the mode price range of the identification results, and use the mode price range as a person feature;

[0103] The mode price range refers to the mode price range of all the recognition results.

[0104] In an example of the technical solution of the present invention, by tracking people based on cameras installed in public areas, the personnel area can be located, and the contour of the personnel area can be recognized. The upper garment area and the lower garment area can be further located, and objects can be recognized in the upper garment area and the lower garment area, and a variety of object recognition results can be obtained. The object recognition result of clothing is often accompanied by a price. There are many recognition results and many prices. The prices are classified into intervals, the types of prices are simplified, and the price interval with the most corresponding recognition results is obtained, which is called the mode price interval, and can be used as a personnel feature.

[0105] The above scheme actually provides personnel characteristics other than age and gender. It relies on existing object recognition technology to obtain the average price of the user's clothing and predict their consumption capacity. In most cases, the price of clothing is positively correlated with the consumption level (service demand). The present invention does not require alignment, and it only needs to meet the positive correlation conditions.

[0106] Figure 3 This is a second sub-flow diagram of the intelligent service method, wherein the step of determining the service demand table according to the crowd portrait includes:

[0107] Step S201: Read all the mode price ranges in the crowd portrait;

[0108] Step S202: determining the consumer characteristics of the population based on the mode price range;

[0109] Step S203: Input the consumption characteristics of the crowd into the trained neural network model to obtain the service type and service amount, and construct a service demand table;

[0110] The above content limits the application process of crowd portraits, in which the characteristics of people in the crowd portraits are limited to the mode price range. All the mode price ranges in the crowd portraits are analyzed to extract the crowd consumption characteristics. The crowd consumption characteristics are the comprehensive values determined by the mode price ranges corresponding to multiple people. They are input into the trained neural network model, and the obtained service types and service amounts can be used to construct a service demand table.

[0111] Among them, the training process of the neural network model is:

[0112] Select service units within a preset time and space range to obtain their population portraits and real consumption records;

[0113] Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges;

[0114] Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts;

[0115] Train the neural network model based on the sample set;

[0116] The above content explains the training process of the neural network model. The focus of the training process of the neural network model is its sample construction process. First, the management party sets a time and space range, which represents the time range and space range. The time range represents the service unit within which time period, and the space range represents the service units at which locations. Service units are selected within the preset time and space range, and their crowd portraits and real consumption records are obtained (both are known data in the technical solution of the present invention). All the mode price ranges in the crowd portraits are read, and the crowd consumption characteristics are determined according to the mode price range. The crowd consumption characteristics are used as input, and the real consumption records are used as output to construct a sample set. The neural network model can be trained based on the sample set; wherein, the real consumption records include service types and service amounts.

[0117] In the technical solution of the present invention, there is a conversion relationship between the mode price range and the consumption characteristics of the population. One feasible solution is as follows:

[0118] The process of determining the consumption characteristics of a population is as follows:

[0119]

[0120] In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α i represents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N},|S|=k represents the set {A1,A2,…,A N} Randomly select a subset of k mode value intervals, ∩ i∈s (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N},|S|=k} means to find the mean of the midpoint values of the intersection.

[0121] The above content provides a specific solution for determining the consumption characteristics of a population. First, the mode price range corresponding to all persons is queried. Then, the number is selected in turn. For example, there are 10 persons in total, and the number is an integer from 1 to 10. The mode price range is selected based on the number, and the intersection is calculated and the midpoint value of the intersection is selected. Since the selected persons may be different, for example, 9 persons are selected from 10 persons, there are a total of 10 selection methods, and there are 9 corresponding intersection results and 9 midpoint values. The mean of the 9 midpoint values is calculated as the characteristic value of the number 9. Finally, the mean of all characteristic values is calculated to obtain a single value as the consumption characteristic of the population.

[0122] Figure 4 This is a third sub-flow diagram of the intelligent service method. The steps of receiving a service request sent by a service unit in real time, generating a service instruction, analyzing the service request, and updating the service demand table include:

[0123] Step S301: receiving the service type and service quota sent by the service unit in real time, uploading the service unit and its service type and service quota to the master control terminal, and generating a service instruction;

[0124] Step S302: Count the service types and service amounts of the service units as actual consumption records and update the sample set.

[0125] In actual application, after receiving the service type and service amount sent by the service unit, it is necessary to immediately arrange staff to provide services, that is, to provide service resources; the service type indicates what kind of service the customer in the service unit needs, and the service amount indicates how much is needed; at the same time, the service types and service amounts sent by all service units are counted to generate historical records, which are called real consumption records. At this time, the number of recent samples in the sample set increases. The more samples in the recent sample set, the smaller the loss function of the neural network model and the higher the accuracy.

[0126] Figure 5 This is a fourth sub-flow chart of the intelligent service method, wherein the step of pre-scheduling service resources according to the service demand tables of all service units comprises:

[0127] Step S401: every preset period, obtaining the service demand table of all service units in the previous period;

[0128] Step S402: combining the service units according to their positional relationships and merging corresponding service demand tables;

[0129] Step S403: pre-schedule service resources according to the merged service demand table.

[0130] Steps S401 to S403 are the focus of the technical solution of the present invention. A pre-scheduling process is executed every period of time (preset period). Starting from the current moment, the process is reversed for a period range to obtain the service demand tables of all service units in the previous period. These service demand tables are predicted service demand tables (the actual demand will be provided once received in step S300). According to the location relationship of the service units, similar service units are classified into one category, and their corresponding service demand tables are merged. This process is equivalent to partitioning; finally, according to the service demand table of each zone, service resources are prepared in advance. This is the meaning of pre-scheduling.

[0131] Figure 6 : This is a structural block diagram of an intelligent service system. In an embodiment of the present invention, an intelligent service system 10 includes:

[0132] A crowd portrait building module 11 is used to acquire real-time personnel information of each service unit based on information collectors installed in public areas and build a crowd portrait of the service unit;

[0133] A service demand determination module 12 is configured to determine a service demand table based on the crowd portrait; the service demand table includes service types and service amounts;

[0134] A service instruction generation module 13 is used to receive service requests sent by service units in real time, generate service instructions, analyze the service requests, and update the generation process of the service demand table;

[0135] The resource pre-scheduling module 14 is used to pre-scheduling service resources according to the service demand tables of all service units at regular intervals;

[0136] The crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

[0137] Furthermore, the crowd portrait construction module 11 includes:

[0138] A usage information query unit is used to obtain usage information of any service unit according to the information input port; the usage information includes the number of users and the usage time period;

[0139] A personnel feature detection unit, configured to obtain the personnel features of persons entering and exiting the service unit based on cameras installed in public areas;

[0140] The personnel feature statistics unit is used to count the personnel features of the service unit during the use period according to the entry and exit conditions corresponding to each personnel feature, and obtain a crowd portrait as the crowd portrait.

[0141] Specifically, the personnel feature detection unit includes:

[0142] The personnel tracking subunit is used to track people and locate personnel areas based on cameras installed in public areas;

[0143] A contour recognition subunit, configured to perform contour recognition on the personnel area and determine an upper garment area and a lower garment area according to the relative position of each contour in the personnel area;

[0144] The item recognition subunit is used to recognize items in the top and bottom areas, query the mode price range of the recognition results, and use the mode price range as a person feature;

[0145] The mode price range refers to the mode price range of all the recognition results.

[0146] Furthermore, the service demand determination module 12 includes:

[0147] Price range reading unit, used to read all the majority price ranges in the crowd portrait;

[0148] a consumption characteristics determination unit, configured to determine the consumption characteristics of a group of people based on the mode price range;

[0149] Build an execution unit to input the consumption characteristics of the crowd into the trained neural network model, obtain the service type and service amount, and build a service demand table;

[0150] Among them, the training process of the neural network model is:

[0151] Select service units within a preset time and space range to obtain their population portraits and real consumption records;

[0152] Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges;

[0153] Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts;

[0154] Train the neural network model based on the sample set;

[0155] The process of determining the consumption characteristics of a population is as follows:

[0156]

[0157] In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α irepresents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N},|S|=k represents the set {A1,A2,…,A N} Randomly select a subset of k mode value intervals, ∩ i∈S (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N},|S|=k} means to find the mean of the midpoint values of the intersection.

[0158] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent service method, characterized in that: The method comprises: Based on the information collectors installed in public areas, real-time personnel information of each service unit is obtained to build a crowd portrait of the service unit; Determine a service demand table based on the crowd portrait; the service demand table includes service type and service amount; The process of receiving service requests sent by service units in real time, generating service instructions, analyzing the service requests, and updating the service demand table; Regularly pre-schedule service resources according to the service demand tables of all service units; The crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

2. The intelligent service method according to claim 1, characterized in that: The step of acquiring real-time personnel information of each service unit based on information collectors installed in public areas and constructing a crowd portrait of the service unit includes: For any service unit, the usage information of the service unit is obtained according to the information input port; the usage information includes the number of users and the usage time period; Obtaining personal characteristics of people entering and leaving the service unit based on cameras installed in public areas; According to the entry and exit situations corresponding to each personnel characteristic, the personnel characteristics of the service unit during the use period are counted to obtain a crowd portrait.

3. The intelligent service method according to claim 2, characterized in that: The step of obtaining the personal characteristics of the persons entering and leaving the service unit according to the cameras installed in the public area comprises: Track people and locate areas with people based on cameras installed in public areas; Performing contour recognition on the personnel area, and determining an upper garment area and a lower garment area according to the relative position of each contour in the personnel area; Identify items in the tops and bottoms areas, query the mode price range of the identification results, and use the mode price range as a person feature; The mode price range refers to the mode price range of all the recognition results.

4. The intelligent service method according to claim 1, characterized in that: The step of determining a service demand table based on the crowd portrait includes: Read all the mode price ranges in the crowd portrait; Determining the consumption characteristics of the population based on the mode price range; Input the consumption characteristics of the crowd into the trained neural network model to obtain the service type and service amount, and construct a service demand table; Among them, the training process of the neural network model is: Select service units within a preset time and space range to obtain their population portraits and real consumption records; Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges; Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts; Train the neural network model based on the sample set; The process of determining the consumption characteristics of a population is as follows: In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α i represents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N },|S|=k represents the set {A1,A2,…,A N } Randomly select a subset of k mode value intervals, ∩ i∈S (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N },|S|=k} means to find the mean of the midpoint values of the intersection.

5. The intelligent service method according to claim 1, characterized in that: The steps of receiving the service request sent by the service unit in real time, generating a service instruction, analyzing the service request, and updating the service requirement table include: Receive the service type and service quota sent by the service unit in real time, upload the service unit and its service type and service quota to the main control terminal, and generate service instructions; The service type and service amount of the service unit are counted as real consumption records and the sample set is updated.

6. The intelligent service method according to claim 1, characterized in that: The step of pre-scheduling service resources according to the service demand tables of all service units at regular intervals includes: Every preset period, obtain the service demand table of all service units in the previous period; Combining the service units according to their positional relationships and merging the corresponding service demand tables; Pre-schedule service resources according to the merged service demand table.

7. An intelligent service system, characterized in that: The system comprises: The crowd portrait construction module is used to obtain real-time personnel information of each service unit based on information collectors installed in public areas and construct crowd portraits of service units; A service demand determination module is used to determine a service demand table based on the crowd portrait; the service demand table includes service types and service amounts; A service instruction generation module is used to receive service requests sent by service units in real time, generate service instructions, analyze the service requests, and update the generation process of the service demand table; Resource pre-scheduling module, used to pre-scheduling service resources according to the service demand tables of all service units at regular intervals; The crowd portrait is a collection of character portraits, and the character portrait is a numerical value group corresponding to each person, and each element in the numerical value group corresponds to a preset feature.

8. The intelligent service system according to claim 7, characterized in that: The crowd portrait construction module includes: A usage information query unit is used to obtain usage information of any service unit according to the information input port; the usage information includes the number of users and the usage time period; A personnel feature detection unit, configured to obtain the personnel features of persons entering and exiting the service unit based on cameras installed in public areas; The personnel feature statistics unit is used to count the personnel features of the service unit during the use period according to the entry and exit conditions corresponding to each personnel feature, and obtain a crowd portrait as the crowd portrait.

9. The intelligent service system according to claim 7, characterized in that: The personnel feature detection unit includes: The personnel tracking subunit is used to track people and locate personnel areas based on cameras installed in public areas; A contour recognition subunit, configured to perform contour recognition on the personnel area and determine an upper garment area and a lower garment area according to the relative position of each contour in the personnel area; The item recognition subunit is used to recognize items in the top and bottom areas, query the mode price range of the recognition results, and use the mode price range as a person feature; The mode price range refers to the mode price range of all the recognition results.

10. The intelligent service system according to claim 7, characterized in that: The service demand determination module includes: Price range reading unit, used to read all the majority price ranges in the crowd portrait; a consumption characteristics determination unit, configured to determine the consumption characteristics of a group of people based on the mode price range; Build an execution unit to input the consumption characteristics of the crowd into the trained neural network model, obtain the service type and service amount, and build a service demand table; Among them, the training process of the neural network model is: Select service units within a preset time and space range to obtain their population portraits and real consumption records; Read all the mode price ranges in the crowd portrait, and determine the crowd consumption characteristics based on the mode price ranges; Establishing a sample set of consumer characteristics and real consumption records; the real consumption records include service types and service amounts; Train the neural network model based on the sample set; The process of determining the consumption characteristics of a population is as follows: In the formula, S represents the consumption characteristics of the crowd, N represents the total number of people in the crowd portrait, and Z i represents the i-th eigenvalue, α i represents the coefficient corresponding to the ith eigenvalue; β is the preset correction coefficient; S∈{A1,A2,…,A N },|S|=k represents the set {A1,A2,…,A N } Randomly select a subset of k mode value intervals, ∩ i∈S (A j ) represents the intersection of the mode value intervals in the set S, A j represents the i-th mode price interval in S; j∈[1,k], k∈[1,N]; E{∩ i∈S (A j ),S∈{A1,A2,…,A N },|S|=k} means to find the mean of the midpoint values of the intersection.

Images