A scenic spot passenger flow analysis method and system based on multimodal feature recognition
Through multi-modal feature recognition technology, a viewing vehicle diversion plan is formulated, which solves the problems of tourists' mood and diversion efficiency in the scenic area, and improves the operation effect of scenic area and the utilization rate of viewing vehicle.
Patent Information
- Application Number
- CN202411164736.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-23
AI Technical Summary
The existing technology cannot effectively monitor the mood and emotions of tourists in scenic spots, resulting in the inability to divert in time and reduce the tourists' experience. At the same time, the viewing vehicle route is fixed, the utilization rate and benefits are low, and the diversion efficiency is not high.
Multimodal feature recognition technology is adopted to collect tourist data and environmental data, analyze whether the scenic spots have reached the early warning mode, formulate a diversion plan for viewing vehicles, and adjust the diversion plan according to the passenger flow distribution to improve diversion efficiency and tourists' comfort.
The diversity monitoring of the number, mood and environment of tourists in the scenic area has been achieved, the mobility and diversion efficiency of tourists have been improved, the mood of tourists has been maintained, the utilization rate and benefits of viewing vehicles have been increased, and the operation effect of scenic spots has been improved.
Smart Images

Figure CN118690309B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of scenic spot management, and in particular to a scenic spot passenger flow analysis method and system based on multimodal feature recognition. Background Art
[0002] With economic development and the improvement of residents' living standards, the tourism industry has developed rapidly and the number of tourists has increased significantly. Scenic spots are facing many challenges such as congestion during peak passenger flow periods and pressure on environmental maintenance. Multimodal feature recognition can comprehensively utilize multiple information sources such as images and sensors to provide scenic spots with real-time and accurate passenger flow data support, which helps scenic spots make data-driven decisions, such as optimizing routes, thereby improving the mobility of tourists, realizing the diversion of tourists within the scenic spot, and improving management efficiency.
[0003] Existing technologies, such as the invention patent disclosed in the application with publication number CN116503209A, are based on artificial intelligence and data-driven digital twin systems, and involve the field of scenic spot management technology; they include a model management module, a tourist registration module, a tourist number monitoring module, a scenic spot recommendation module and a database; by monitoring and classifying the number of tourists at each scenic spot in the scenic area, analyzing the number of recommended scenic spots by tourists based on the length of time they spend visiting the diverted scenic spots and the number of scenic spots they have visited, and then screening out the recommended scenic spots corresponding to the tourists in the diverted scenic spots from the diversion scenic spots, which solves the problem that the current technology cannot improve the level of intelligence in scenic spot diversion, realizes personalized recommendations for tourist attractions in the scenic area and a rational distribution of the number of tourists, improves the diversion effect of scenic spots with too high a flow of people, and also improves the tourists' travel experience and travel services, while also greatly ensuring the coordinated development of the scenic spot.
[0004] There are at least the following deficiencies in the above scheme: 1. The number of tourists and the environment in the scenic spot affect the tourists' mood for play, and the expressions of tourists also reflect their mood for play. However, the above scheme only monitors the number of tourists in each scenic spot, and does not monitor and analyze the environment and emotions of tourists in each scenic spot. It is impossible to understand the mood of tourists in each scenic spot, and it is impossible to divert tourists in time when tourists have negative emotions, thereby failing to maintain the tourists' mood for play and reducing the tourists' sense of play experience.
[0005] 2. The sightseeing cars in scenic spots can speed up the diversion of tourists. However, in the above scheme, when diverting tourists, only scenic spots that can be diverted are recommended, and the sightseeing cars are not used for diversion, thereby reducing the efficiency of tourist diversion and failing to improve the mobility of tourists. In addition, in traditional technologies, the tour routes of the sightseeing cars in scenic spots are fixed, and the routes of the sightseeing cars cannot be automatically formulated and adjusted according to the trajectories of tourists at various scenic spots in the scenic spot, nor can the sightseeing cars be allocated according to the age distribution of tourists at each scenic spot, thereby failing to improve the utilization rate and revenue of the sightseeing cars and reducing the operation effect of the scenic spot. Summary of the invention
[0006] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a scenic spot passenger flow analysis method and system based on multimodal feature recognition.
[0007] To solve the above technical problems, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a scenic spot passenger flow analysis method based on multimodal feature recognition, which includes the following steps: Step 1: Collect initial tourist data and initial environmental data of each scenic spot in the scenic spot.
[0008] Step 2: Use the initial tourist data and initial environmental data of each scenic spot to analyze whether each scenic spot is in early warning mode, record the scenic spots in early warning mode as marked scenic spots, and formulate a diversion plan for sightseeing vehicles.
[0009] Step 3: After the sightseeing car diversion plan is implemented for a preset time, the passenger flow distribution in the scenic area is monitored. When the passenger flow distribution in the scenic area varies greatly, the sightseeing car diversion plan is adjusted.
[0010] Step 4: When at least one scenic spot is in warning mode or the difference in passenger flow distribution is large, send a warning prompt.
[0011] In a second aspect, the present invention provides a scenic spot passenger flow analysis system based on multimodal feature recognition, including: a data acquisition module for collecting initial tourist data and initial environmental data of each scenic spot in the scenic spot.
[0012] The diversion analysis module is used to use the initial tourist data and initial environmental data of each scenic spot to analyze whether each scenic spot is in early warning mode, and record the scenic spots in early warning mode as marked scenic spots to formulate a diversion plan for sightseeing vehicles.
[0013] The diversion monitoring module is used to monitor the passenger flow distribution in the scenic area after the sightseeing car diversion plan is implemented for a preset time. When the passenger flow distribution in the scenic area varies greatly, the diversion plan for the sightseeing car is adjusted.
[0014] The diversion prompt module is used to send a warning prompt when at least one scenic spot is in early warning mode or the passenger flow distribution difference is large.
[0015] The beneficial effects of the present invention are as follows: the present invention provides a scenic spot passenger flow analysis method and system based on multimodal feature recognition, which monitors the number, emotions and environment of tourists at various scenic spots in the scenic spot. When the early warning mode is reached, a diversion plan for sightseeing vehicles is formulated to implement tourist diversion at the scenic spot. After the diversion, the tourist density at the scenic spot is monitored again. When the tourist distribution is uneven, the diversion plan for the sightseeing vehicle is adjusted, which solves the problems of the singleness of tourist monitoring at the scenic spot and the lack of flexibility in sightseeing vehicle route planning in traditional technologies, improves the diversity of scenic spot data monitoring, improves the mobility and diversion efficiency of tourists in the scenic spot, maintains the tourists' mood and experience, increases the tourists' comfort, and at the same time improves the utilization rate and revenue of the sightseeing vehicle, and also improves the operation effect of the scenic spot. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0017] Figure 1 The present invention is a schematic flow chart of the steps for implementing the method.
[0018] Figure 2 It is a schematic diagram of the system structure connection of the present invention. DETAILED DESCRIPTION
[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0020] See also Figure 1 As shown, a scenic spot passenger flow analysis method based on multimodal feature recognition includes the following steps: Step 1: Collect initial tourist data and initial environmental data of each scenic spot in the scenic spot.
[0021] In a specific embodiment, the initial tourist data and initial environmental data of each scenic spot in the scenic area are collected, and the specific process is as follows: when a tourist enters the scenic area, after the tourist's identity information is confirmed, the smart bracelet is bound to the tourist's identity information, and the tourist wears the smart bracelet to enter the scenic area, where one tourist wears one smart bracelet.
[0022] Among them, the identity information includes name and age, etc. When entering the scenic spot, tourists need to verify the relevant documents of the identity information and obtain the identity information from the documents, which include identity cards and passports, etc.
[0023] When starting to collect initial tourist data, use a smart bracelet to obtain the location data of each tourist, and import the location data of each tourist into the scenic area map to obtain the scenic spots where each tourist is located and count the number of tourists at each scenic spot.
[0024] Among them, the scenic spot map is stored in the scenic spot passenger flow analysis system.
[0025] Cameras are installed at each scenic spot in the scenic area, and panoramic images of each scenic spot are captured by cameras carried by drones. The facial images of each tourist in each scenic spot are identified from the panoramic images, and the facial emotion recognition technology is used to obtain the emotion type of each tourist in each scenic spot; the location and number of each tourist in each scenic spot are used as the initial tourist data.
[0026] It should be noted that facial emotion recognition technology is a technology that uses computer vision and emotional computing to analyze the expressions in facial images to identify and understand people's emotional states. The specific technical principles have been made public and will not be repeated here. Emotion types include positive emotions and negative emotions.
[0027] An environment collection device is arranged in each scenic spot, and when the initial environment data collection starts, the environment collection device is used to collect the initial environment data of each scenic spot.
[0028] It should be noted that the environment collection device includes a temperature sensor, a humidity sensor, a noise sensor and an air quality detector, and the initial environment data includes temperature, humidity, noise value and air quality index.
[0029] It should also be noted that, according to relevant laws, regulations and provisions, in order to protect the personal privacy of tourists, the applicant may need to obtain and process the identity information of tourists in this patent application. Here, the applicant solemnly promises to strictly abide by relevant laws, regulations and provisions, be responsible for the confidentiality of the identity information of tourists, and will not use the identity information of tourists for other purposes. At the same time, the applicant will also take necessary technical and organizational measures to ensure the security and confidentiality of the identity information of tourists. While agreeing to provide identity information, tourists also understand and agree to the above authorization and commitment.
[0030] Step 2: Use the initial tourist data and initial environmental data of each scenic spot to analyze whether each scenic spot is in early warning mode, record the scenic spots in early warning mode as marked scenic spots, and formulate a diversion plan for sightseeing vehicles.
[0031] In a specific embodiment, the analysis of whether each scenic spot is in early warning mode is carried out as follows: the threshold value of the number of tourists and the comfortable environment data interval of each scenic spot are obtained from the scenic spot control center, and the emotional type of each tourist in each scenic spot in the initial tourist data of each scenic spot is extracted, and the number of tourists with negative emotions in each scenic spot is counted, and the number of tourists in each scenic spot, the number of tourists with negative emotions, the initial environmental data, the threshold value of the number of tourists and the comfortable environment data interval of the scenic spot are input into the scenic spot pattern analysis model, and the passenger flow pattern results of each scenic spot are output.
[0032] The passenger flow pattern result includes values of 1 and -1. If the passenger flow pattern result is 1, it indicates that the passenger flow pattern is in a normal mode. If the passenger flow pattern result is -1, it indicates that the passenger flow pattern is in a warning mode.
[0033] In the above, the expression of the scenic spot pattern analysis model is:
[0034] In the formula, α j represents the passenger flow pattern result of the jth scenic spot, P j , P j ′、P1 j , H j They represent the number of tourists, the threshold of the number of tourists, the number of tourists with negative emotions, and the initial environmental data of the jth scenic spot, respectively. H represents the comfortable environmental data interval of the scenic spot, and ε 1 , ε 2 are the weight coefficients of the preset tourist excess rate and the weight coefficient of the proportion of tourists with negative emotions, respectively. j represents the number of each scenic spot, j is a positive integer, and χ represents the preset threshold of the crowd flow evaluation coefficient.
[0035] Among them, the number of tourists in each scenic spot exceeded the rate The proportion of tourists with negative emotions at each attraction Select the maximum tourist number excess rate, the minimum tourist number excess rate, the maximum tourist number proportion and the minimum tourist number proportion as μ max , μ min , and The fluctuation value of the tourist number excess rate is calculated from this and the fluctuation of the proportion of tourists and
[0036] The threshold of the crowd flow evaluation coefficient represents the critical value of the upper limit of the tourist carrying state of the scenic spot, which is set by the scenic spot management personnel. When the quantitative result of the tourist carrying state of the scenic spot is less than the threshold of the crowd flow evaluation coefficient, it indicates that more tourists can be added to the scenic spot. Otherwise, it indicates that the tourist load in the scenic spot is overloaded and needs to be diverted. The expression of the scenic spot pattern analysis model is The result represents the quantitative result of the tourist carrying capacity of each scenic spot. For example, when the quantitative result of the tourist carrying capacity of the scenic spot is 8, and the threshold of the crowd flow assessment coefficient is 12, 8<12, indicating that the tourist load in the scenic spot is not overloaded and more tourists can be added.
[0037] In another specific embodiment, the diversion plan for the sightseeing car is formulated, and the specific process is as follows: S2-1, obtaining the identity information of each tourist in each marked scenic spot, extracting the age of each tourist in each marked scenic spot, obtaining the age ratio data of tourists in each marked scenic spot, and using trajectory prediction technology to predict the movement trajectories of tourists in each marked scenic spot and the flow of people on each movement trajectory, and importing them into the scenic area map, taking the end point of each movement trajectory of tourists in each marked scenic spot as the point, and drawing a circle with a preset radius, the area inside the circle is the target diversion area corresponding to each movement trajectory of each marked scenic spot.
[0038] Among them, trajectory prediction technology is a technology that uses data and models to predict the future movement path of an object. It is widely used in the fields of autonomous driving, intelligent transportation, robot navigation, motion analysis, etc. Machine learning or deep learning models are often used to predict future trajectories. The specific technical principles have been made public and will not be repeated here.
[0039] It should be noted that the preset radius is set by the scenic area management personnel and can be 5 meters or 8 meters, etc.
[0040] It should also be noted that the age ratio data of tourists at each marked scenic spot is the ratio of each age group in each marked scenic spot, which can be obtained by counting the number of tourists of each age group in each marked scenic spot and then dividing it by the number of tourists of each marked scenic spot. Among them, the age groups can be divided into 0-15 years old, 16-35 years old, 36-50 years old and over 50 years old.
[0041] S2-2. Obtain the status coefficient of each marked scenic spot, and use the age ratio data of tourists at each marked scenic spot to analyze the demand value of the sightseeing cars at each marked scenic spot, and obtain the number of vacant sightseeing cars in the scenic area, allocate the number of sightseeing cars to each marked scenic spot according to the demand ratio of each marked scenic spot, and allocate the number of sightseeing cars to each moving track of each marked scenic spot according to the passenger flow ratio of each moving track.
[0042] Among them, allocating the number of sightseeing vehicles to each marked scenic spot according to the demand ratio of each marked scenic spot means: dividing the demand value of each marked scenic spot by the sum of the demand values of all marked scenic spots, and then multiplying it by the number of free sightseeing vehicles in the scenic area to obtain the number of sightseeing vehicles to each marked scenic spot.
[0043] Similarly, allocating the number of sightseeing cars for each moving track of each marked scenic spot according to the proportion of the flow of people on each moving track means: dividing the flow of people on each moving track of each marked scenic spot by the sum of the flow of people on each moving track, and then multiplying it by the number of sightseeing cars for each marked scenic spot, to obtain the number of sightseeing cars for each moving track of each marked scenic spot.
[0044] In the above, the analysis of the demand value of the sightseeing car of each marked scenic spot is specifically carried out as follows: the initial environmental data of each marked scenic spot is compared with the comfortable environmental data interval of the scenic spot to obtain the environmental data difference of each marked scenic spot, which is recorded as H b , the calculation formula of the state coefficient is: In the formula, δ b , P b , P b ′、P1 b They represent the state coefficient, number of tourists, threshold of number of tourists, and number of tourists with negative emotions of the b-th marked attraction respectively;
[0045] Among them, when the initial environmental data is less than the lower limit of the scenic area's comfortable environmental data interval, the environmental data difference is the lower limit minus the initial environmental data; when the initial environmental data is greater than the upper limit of the scenic area's comfortable environmental data interval, the environmental data difference is the upper limit minus the initial environmental data; when the initial environmental data is within the scenic area's comfortable environmental data interval, the environmental data difference output is 1.
[0046] The proportion of each age group is obtained from the age ratio data of tourists at each marked scenic spot, and the intention weight factor of each age group is obtained based on the preset age group intention weight table. The proportion of each age group in each marked scenic spot is multiplied by the intention weight factor of the corresponding age group, and then added to obtain the intention value of each marked scenic spot, which is recorded as
[0047] Among them, the process of formulating the age group intention weight table is as follows: the ages of historical tourists who have taken the sightseeing bus are obtained from the scenic spot passenger flow analysis system, the number of historical tourists in each age group is counted, and then divided by the total number of historical tourists to obtain the intention weight factor of each age group.
[0048] It should be noted that when tourists take a sightseeing bus, they can generate a QR code in their smart bracelet. The sightseeing bus driver uses a scanning device to scan the QR code to complete payment and confirm identity information, and transmit the scanning record to the scenic spot passenger flow analysis system for storage, realizing real-name vehicle riding and convenient payment methods, while improving the use scenarios of smart bracelets, unifying payment methods within the scenic area, and optimizing the management of the scenic area.
[0049] Using analytical formulas Get the demand value φ of the sightseeing car for the bth marked scenic spot b , where γ 1 , γ2 They are respectively the weight coefficient of the preset state coefficient and the weight coefficient of the intention value.
[0050] It should be noted that the state coefficient of each marked scenic spot is calculated by difference, the maximum state coefficient difference and the minimum state coefficient difference are selected, and the maximum state coefficient difference is subtracted from the minimum state coefficient difference, and the result is divided by the minimum state coefficient difference to obtain the state coefficient fluctuation value; similarly, the intention value of each marked scenic spot is calculated by difference, the maximum intention difference and the minimum intention difference are selected, and the maximum intention difference is subtracted from the minimum intention difference, and the result is divided by the minimum intention difference to obtain the intention value fluctuation value; the weight coefficient of the state coefficient is equal to the state coefficient fluctuation value divided by the sum of the state coefficient fluctuation value and the intention value fluctuation value, and the weight coefficient of the intention value is equal to the intention value fluctuation value divided by the sum of the state coefficient fluctuation value and the intention value fluctuation value.
[0051] S2-3, the location of each marked scenic spot, each moving track and the target diversion area of each moving track are recorded as z b , g by and q by , where b represents the number of each marked scenic spot, y represents the number of each moving track, and both b and y are positive integers; according to the method of analyzing whether each scenic spot is in warning mode, analyze q by Is it in warning mode? 11 For early warning mode, select 11 The target diversion area with the closest distance and normal passenger flow pattern replaces q 11 , and change g 11 The end point forms a new g 11 , and the new g 11 Cover 11 , so as to change the movement track of each marked scenic spot.
[0052] Among them, q 11 Indicates the target diversion area of the first marked scenic spot corresponding to the first moving trajectory, g 11 Indicates the first movement trajectory of the first marked scenic spot.
[0053] S2-4, if z 1 With z 2 The position spacing is less than the preset spacing threshold, g 11 With g 21 The overlap area is greater than the first overlap area threshold and q 11 With q 21 If the overlap area is greater than the second overlap area threshold, select g 11 With g 21 The longer moving trajectory is taken as the coincident trajectory, i.e. g 11 With g 21The driving routes of all the sightseeing cars are the overlapping trajectories; if a moving trajectory of a marked scenic spot does not have an overlapping trajectory, then the moving trajectory is used as the driving route of all the sightseeing cars in the moving trajectory, thereby obtaining each driving route and the number of sightseeing cars in each driving route, and serving as a diversion plan for the sightseeing cars.
[0054] Among them, g 21 represents the first moving trajectory of the second marked scenic spot, q 21 Indicates that the second marked scenic spot corresponds to the target diversion area of the first moving trajectory, z 1 Indicates the first marked scenic spot, z 2 Indicates the second marked spot.
[0055] It should be noted that when the diversion plan for the sightseeing car is completed, it is sent to the scenic area manager and the manager is prompted to execute the diversion plan.
[0056] Step 3: After the sightseeing car diversion plan is implemented for a preset time, the passenger flow distribution in the scenic area is monitored. When the passenger flow distribution in the scenic area varies greatly, the sightseeing car diversion plan is adjusted.
[0057] In a specific embodiment, the passenger flow distribution in the scenic area is monitored, and the specific process is as follows: infrared sensors are deployed in each scenic spot, the passenger flow density in each scenic spot is collected, the passenger flow density in each scenic spot is calculated, and the maximum passenger flow density difference is selected as the passenger flow density difference of the scenic spot, and the minimum passenger flow density is selected from the passenger flow density in each scenic spot as the minimum passenger flow density of the scenic spot, and the passenger flow density difference of the scenic spot is divided by the minimum passenger flow density to obtain the passenger flow difference rate of the scenic spot.
[0058] The passenger flow difference rate of the scenic spot is compared with the preset passenger flow difference rate threshold. If the passenger flow difference rate of the scenic spot is greater than the preset passenger flow difference rate threshold, it indicates that the passenger flow distribution difference in the scenic spot is small, otherwise it indicates that the passenger flow distribution difference in the scenic spot is large.
[0059] In another specific embodiment, the diversion scheme of the sightseeing vehicles is adjusted, and the specific process is as follows: the passenger flow density in each scenic spot is averaged to obtain the average passenger flow density, and each scenic spot with a passenger flow density greater than the average passenger flow density is used as a target scenic spot; each target scenic spot is compared with each marked scenic spot. If a marked scenic spot is different from all target scenic spots, all the sightseeing vehicles of the marked scenic spot are recovered and used as idle sightseeing vehicles. If a target scenic spot is the same as all marked scenic spots, the target scenic spot is used as a second-class scenic spot, and each second-class scenic spot is obtained, and the number of secondary sightseeing vehicles is allocated according to the proportion of the passenger flow density in each target scenic spot.
[0060] The meaning of allocating the number of secondary sightseeing vehicles according to the proportion of passenger flow density in each target scenic spot is: dividing the passenger flow density in each target scenic spot by the sum of the passenger flow density in each target scenic spot, and then multiplying by the number of idle sightseeing vehicles to obtain the number of secondary sightseeing vehicles in each target scenic spot.
[0061] Using trajectory prediction technology, the movement trajectories of tourists in each target scenic spot and the passenger flow density of each movement trajectory are predicted, and the number of secondary viewing vehicles for each movement trajectory is allocated according to the proportion of the passenger flow density of each movement trajectory; if the number of secondary viewing vehicles for a certain movement trajectory of a Class II scenic spot is less than the number of viewing vehicles, some viewing vehicles are recovered; if the number of secondary viewing vehicles for a certain movement trajectory of a Class II scenic spot is greater than the number of viewing vehicles, some viewing vehicles are supplemented, so as to allocate the number of secondary viewing vehicles for each movement trajectory of each target scenic spot.
[0062] In the above, some of the sightseeing cars are recovered, wherein the number recovered is equal to the number of sightseeing cars minus the number of secondary sightseeing cars; similarly, some of the sightseeing cars are replenished, wherein the number replenished is equal to the number of secondary sightseeing cars minus the number of sightseeing cars.
[0063] According to the analysis method of steps S2-3 and S2-4, the adjusted driving routes and the number of sightseeing vehicles in each driving route are analyzed to adjust the diversion plan of the sightseeing vehicles.
[0064] It should be noted that after the diversion plan for the sightseeing car is adjusted, it will be sent to the scenic area management personnel, and the management personnel will be prompted to adjust the diversion plan.
[0065] Step 4: When at least one scenic spot is in warning mode or the difference in passenger flow distribution is large, send a warning prompt.
[0066] In a specific embodiment, the sending of warning prompts is specifically carried out as follows: a prompt indicating that traffic at the scenic spot is high is sent to the smart bracelets of all tourists in the early warning mode, and each driving route is pushed to remind tourists to take a sightseeing bus for diversion.
[0067] When there are large differences in the distribution of passenger flow in a scenic area, a reminder of high traffic flow at each target scenic spot will be sent to the smart bracelets of all tourists at the target scenic spot, and the adjusted driving routes will be pushed to remind tourists to take the sightseeing bus for diversion.
[0068] See also Figure 2 As shown, a scenic spot passenger flow analysis system based on multimodal feature recognition includes: a data acquisition module, a diversion analysis module, a diversion monitoring module and a diversion prompt module.
[0069] The data collection module is used to collect the initial tourist data and initial environmental data of each scenic spot in the scenic area.
[0070] The diversion analysis module is used to use the initial tourist data and initial environmental data of each scenic spot to analyze whether each scenic spot is in early warning mode, and record the scenic spots in early warning mode as marked scenic spots to formulate a diversion plan for sightseeing vehicles.
[0071] The diversion monitoring module is used to monitor the passenger flow distribution in the scenic area after the sightseeing car diversion plan is implemented for a preset time. When the passenger flow distribution in the scenic area varies greatly, the diversion plan for the sightseeing car is adjusted.
[0072] The diversion prompt module is used to send a warning prompt when at least one scenic spot is in early warning mode or the passenger flow distribution difference is large.
[0073] The embodiment of the present invention monitors the number, mood and environment of tourists at various scenic spots in the scenic area. When the early warning mode is reached, a diversion plan for sightseeing vehicles is formulated to implement tourist diversion at the scenic spot. After the diversion, the tourist density at the scenic spot is monitored again. When the tourists are unevenly distributed, the diversion plan for the sightseeing vehicles is adjusted, thereby solving the problems of single-minded tourist monitoring at scenic spots and insufficient flexibility in sightseeing vehicle route planning in traditional technologies, improving the diversity of scenic spot data monitoring, improving the mobility and diversion efficiency of tourists in the scenic area, maintaining the tourists' mood and experience, increasing the tourists' comfort, and at the same time improving the utilization rate and revenue of sightseeing vehicles, and also improving the operation effect of the scenic area.
[0074] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they shall all fall within the protection scope of the present invention.
Claims
1. A tourist flow analysis method for scenic spots based on multimodal feature recognition, It is characterized in that The steps include: Step 1: Collect initial tourist data and initial environmental data of each scenic spot in the scenic area; Step 2: Using the initial tourist data and initial environmental data of each scenic spot, analyze whether each scenic spot is in early warning mode, record each scenic spot in early warning mode as a marked scenic spot, and formulate a diversion plan for sightseeing vehicles; The specific process of formulating the diversion plan for the sightseeing car is as follows: S2-1, obtain the identity information of each tourist in each marked scenic spot, extract the age of each tourist in each marked scenic spot, obtain the age ratio data of tourists in each marked scenic spot, and use trajectory prediction technology to predict the movement trajectory of tourists in each marked scenic spot and the flow of people in each movement trajectory, and import them into the scenic area map, take the end point of each movement trajectory of tourists in each marked scenic spot as the circle point, and draw a circle with a preset radius. The area inside the circle is the target diversion area of each marked scenic spot corresponding to each movement trajectory; S2-2, obtaining the state coefficient of each marked scenic spot, and using the age ratio data of tourists at each marked scenic spot, analyzing the demand value of the sightseeing car at each marked scenic spot, and obtaining the number of free sightseeing cars in the scenic area, allocating the number of sightseeing cars at each marked scenic spot according to the demand ratio of each marked scenic spot, and allocating the number of sightseeing cars at each moving track of each marked scenic spot according to the flow ratio of each moving track; S2-3, the location of each marked scenic spot, each moving track and the target diversion area of each moving track are recorded as z b , g by and q by , where b represents the number of each marked scenic spot, y represents the number of each moving track, and both b and y are positive integers; according to the method of analyzing whether each scenic spot is in warning mode, analyze q by Is it in warning mode? 11 For early warning mode, select 11 The target diversion area with the closest distance and normal passenger flow pattern replaces q 11 , and change g 11 The end point forms a new g 11 , and the new g 11 Cover 11 , so as to change the movement track of each marked scenic spot, q 11 Indicates the target diversion area of the first marked scenic spot corresponding to the first moving trajectory, g 11 Indicates the first moving track of the first marked scenic spot; S2-4, if z 1 With z 2 The position spacing is less than the preset spacing threshold, g 11 With g 21 The overlap area is greater than the first overlap area threshold and q 11 With q 21 If the overlap area is greater than the second overlap area threshold, select g 11 With g 21 The longer moving trajectory is taken as the coincident trajectory, i.e. g 11 With g 21 The driving routes of all the sightseeing cars are the overlapping trajectories; if a moving trajectory of a marked scenic spot does not have an overlapping trajectory, then the moving trajectory is used as the driving route of all the sightseeing cars in the moving trajectory, thereby obtaining each driving route and the number of sightseeing cars in each driving route, and serving as the diversion plan for the sightseeing cars, g 21 represents the first moving trajectory of the second marked scenic spot, q 21 Indicates that the second marked scenic spot corresponds to the target diversion area of the first moving trajectory, z 1 Indicates the first marked scenic spot, z 2 Indicates the second marked scenic spot; The specific process of analyzing the demand value of the sightseeing car for each marked scenic spot is as follows: Compare the initial environmental data of each marked scenic spot with the comfortable environmental data interval of the scenic spot to obtain the environmental data difference of each marked scenic spot, which is recorded as H b , the calculation formula of the state coefficient is: In the formula, δ b , P b , P b ′、P1 b They represent the state coefficient, number of tourists, threshold of number of tourists, number of tourists with negative emotions of the b-th marked attraction, ε 1 , ε 2 They are the weight coefficient of the preset tourist number excess rate and the weight coefficient of the proportion of tourists with negative emotions; The proportion of each age group is obtained from the age ratio data of tourists at each marked scenic spot, and the intention weight factor of each age group is obtained based on the preset age group intention weight table. The proportion of each age group in each marked scenic spot is multiplied by the intention weight factor of the corresponding age group, and then added to obtain the intention value of each marked scenic spot, which is recorded as Using analytical formulas Get the demand value φ of the sightseeing car for the bth marked scenic spot b , where γ 1 , γ 2 They are respectively the weight coefficient of the preset state coefficient and the weight coefficient of the intention value; Step 3: After the sightseeing car diversion plan is executed for a preset time, the passenger flow distribution in the scenic area is monitored. When the passenger flow distribution in the scenic area is quite different, the sightseeing car diversion plan is adjusted; The specific process of monitoring the distribution of passenger flow in the scenic area is as follows: Infrared sensors are deployed in each scenic spot to collect the passenger flow density in each scenic spot, and the passenger flow density in each scenic spot is calculated by difference. The maximum passenger flow density difference is selected as the passenger flow density difference of the scenic spot, and the minimum passenger flow density is selected from the passenger flow density in each scenic spot as the minimum passenger flow density of the scenic spot. The passenger flow density difference of the scenic spot is divided by the minimum passenger flow density to obtain the passenger flow difference rate of the scenic spot. The passenger flow difference rate of the scenic spot is compared with the preset passenger flow difference rate threshold. If the passenger flow difference rate of the scenic spot is greater than the preset passenger flow difference rate threshold, it indicates that the passenger flow distribution difference in the scenic spot is small, otherwise it indicates that the passenger flow distribution difference in the scenic spot is large; The diversion plan for the sightseeing car is adjusted, and the specific process is as follows: The passenger flow density in each scenic spot is averaged to obtain the average passenger flow density, and each scenic spot with a passenger flow density greater than the average passenger flow density is taken as each target scenic spot; each target scenic spot is compared with each marked scenic spot. If a marked scenic spot is different from each target scenic spot, all the sightseeing vehicles of the marked scenic spot are recovered and used as idle sightseeing vehicles. If a target scenic spot is the same as each marked scenic spot, the target scenic spot is taken as a second-class scenic spot, and each second-class scenic spot is obtained, and the number of secondary sightseeing vehicles is allocated according to the proportion of the passenger flow density in each target scenic spot; Using trajectory prediction technology, the movement trajectories of tourists in each target scenic spot and the passenger flow density of each movement trajectory are predicted, and the number of secondary sightseeing vehicles for each movement trajectory is allocated according to the proportion of the passenger flow density of each movement trajectory; If the number of secondary viewing vehicles in a certain moving track of a certain Class II scenic spot is less than the number of viewing vehicles, some viewing vehicles will be recovered; if the number of secondary viewing vehicles in a certain moving track of a certain Class II scenic spot is greater than the number of viewing vehicles, some viewing vehicles will be added, so as to allocate the number of secondary viewing vehicles in each moving track of each target scenic spot; According to the analysis method of steps S2-3 and S2-4, the adjusted driving routes and the number of sightseeing vehicles in each driving route are analyzed to adjust the diversion plan of the sightseeing vehicles; Step 4: When at least one scenic spot is in warning mode or the difference in passenger flow distribution is large, send a warning prompt.
2. According to the method for analyzing tourist flow in a scenic spot based on multimodal feature recognition according to claim 1, It is characterized in that The specific process of collecting the initial tourist data and initial environment data of each scenic spot in the scenic area is as follows: When a tourist enters a scenic spot, after the tourist's identity information is confirmed, the smart bracelet is bound to the tourist's identity information, and the tourist wears the smart bracelet to enter the scenic spot, wherein one tourist wears one smart bracelet; When collecting initial tourist data, use a smart bracelet to obtain the location data of each tourist, and import the location data of each tourist into the scenic area map to obtain the scenic spots where each tourist is located and count the number of tourists at each scenic spot; Cameras are installed at each scenic spot in the scenic area, and panoramic images of each scenic spot are taken by cameras carried by drones. Facial images of each tourist in each scenic spot are identified from the panoramic images of each scenic spot. Facial emotion recognition technology is used to obtain the emotion type of each tourist in each scenic spot. The location and number of each tourist in each scenic spot are used as initial tourist data. An environment collection device is arranged in each scenic spot, and when the initial environment data collection starts, the environment collection device is used to collect the initial environment data of each scenic spot.
3. According to the method for analyzing tourist flow in a scenic spot based on multimodal feature recognition according to claim 1, It is characterized in that The specific process of analyzing whether each scenic spot is in early warning mode is as follows: The threshold of the number of tourists and the comfortable environment data interval of each scenic spot are obtained from the scenic spot control center, and the emotional type of each tourist in each scenic spot is extracted from the initial tourist data of each scenic spot, and the number of tourists with negative emotions in each scenic spot is counted. The number of tourists in each scenic spot, the number of tourists with negative emotions, the initial environmental data, the threshold of the number of tourists and the comfortable environment data interval of the scenic spot are input into the scenic spot pattern analysis model, and the passenger flow pattern results of each scenic spot are output; The passenger flow pattern result includes values of 1 and -1. If the passenger flow pattern result is 1, it indicates that the passenger flow pattern is in a normal mode. If the passenger flow pattern result is -1, it indicates that the passenger flow pattern is in a warning mode.
4. According to claim 3, a method for analyzing tourist flow in a scenic spot based on multimodal feature recognition, It is characterized in that The expression of the scenic spot pattern analysis model is: In the formula, α j represents the passenger flow pattern result of the jth scenic spot, P j , P j ′、P1 j , H j They represent the number of tourists, the threshold of the number of tourists, the number of tourists with negative emotions, and the initial environmental data of the jth scenic spot, respectively. H represents the comfortable environmental data interval of the scenic spot, and ε 1 , ε 2 are the weight coefficients of the preset tourist excess rate and the weight coefficient of the proportion of tourists with negative emotions, respectively. j represents the number of each scenic spot, j is a positive integer, and χ represents the preset threshold of the crowd flow evaluation coefficient.
5. According to claim 1, a method for analyzing tourist flow in a scenic spot based on multimodal feature recognition, It is characterized in that The specific process of sending the warning prompt is as follows: Send a reminder of high traffic volume at each scenic spot to the smart bracelets of all tourists in the early warning mode, and push each driving route to remind tourists to take the sightseeing bus for diversion; When there are large differences in the distribution of passenger flow in a scenic area, a reminder of high traffic flow at each target scenic spot will be sent to the smart bracelets of all tourists at the target scenic spot, and the adjusted driving routes will be pushed to remind tourists to take the sightseeing bus for diversion.
6. A scenic spot passenger flow analysis system using the scenic spot passenger flow analysis method based on multimodal feature recognition as described in any one of claims 1 to 5, It is characterized in that include: The data collection module is used to collect the initial tourist data and initial environmental data of each scenic spot in the scenic area; The diversion analysis module is used to use the initial tourist data and initial environmental data of each scenic spot to analyze whether each scenic spot is in the early warning mode, and record each scenic spot in the early warning mode as a marked scenic spot to formulate a diversion plan for the sightseeing car; The diversion monitoring module is used to monitor the passenger flow distribution in the scenic area after the sightseeing car diversion plan has been implemented for a preset period of time. When the passenger flow distribution in the scenic area varies greatly, the diversion plan for the sightseeing car will be adjusted; the diversion prompt module is used to send a warning prompt when at least one scenic spot is in early warning mode or the passenger flow distribution varies greatly.
Citation Information
Patent Citations
Digital twin system based on artificial intelligence and data driving
CN116503209A
Space-time regulation and emergency guidance system and method for scenic spot tourists
CN104036352A
A scenic spot data management system and method
CN109710660A
Scenic area real-time load monitoring method and system based on space-time big data
CN111083728A
Scenic spot passenger flow early warning system
CN112365638A
