A method and system for iteratively laying out directional signs based on human spatiotemporal distribution

Abstract

The application relates to a guide sign iterative layout method and system based on human flow space-time distribution, and relates to the technical field of intelligent guide systems. The method comprises the following steps: acquiring a multi-source feature data set of tourists, based on which, a dynamic construction of a crowd analysis graph and a behavior evaluation dimension is carried out, and a crowd analysis graph and a behavior evaluation dimension set are generated; based on this, a collaborative decision of guide sign content form adjustment and intelligent shunting guidance is carried out, and a sign adjustment and shunting strategy set is generated; based on this, tourist behavior feedback data is collected, the generation process of the crowd analysis graph and the behavior evaluation dimension set is iteratively optimized in a closed loop, and an iteratively optimized guide sign knowledge base is generated. The application can realize adaptive adjustment of the guide sign according to real-time tourist group composition and interest preferences, improve the accuracy and acceptance rate of shunting guidance, and continuously optimize the layout effect through a closed-loop feedback mechanism.

Description

Technical Field

[0001] This application relates to the field of intelligent wayfinding system technology, specifically to a method and system for iteratively deploying wayfinding signs based on the spatiotemporal distribution of pedestrian flow. Background Technology

[0002] With the booming development of the tourism industry and the continuous growth in the number of tourists visiting scenic spots, wayfinding signage systems, as key infrastructure for guiding tourist flow and optimizing the visitor experience, are becoming increasingly important. Existing wayfinding signage systems typically rely on pre-set static content or timed update mechanisms based on simple schedules. These systems display general route guidance, attraction introductions, and service facility information through fixed physical signs or electronic screens. During implementation, some advanced systems combine video surveillance or sensors to count real-time pedestrian flow. When the number of people in a certain area exceeds a set threshold, predefined congestion alerts or simple diversion broadcasts are triggered. In addition, some solutions utilize basic tags selected by tourists when purchasing tickets or their browsing history to attempt rough interest classification, so as to push relevant promotional content at specific times. These existing technologies mainly focus on monitoring macroscopic pedestrian density and responding based on fixed rules, aiming to maintain basic order and provide basic information services.

[0003] However, in existing technologies, wayfinding signage systems often struggle to adapt to real-time changes in the composition of tourist groups and individual deep-seated interests and preferences. This results in a lack of targeted diversion and guidance strategies in complex and ever-changing tourism scenarios, failing to effectively alleviate localized congestion or path hesitation caused by the diversity of tourist behaviors, and consequently affecting overall tour efficiency and experience quality. Summary of the Invention

[0004] This application provides a method and system for iteratively deploying wayfinding signs based on the spatiotemporal distribution of pedestrian flow, in order to solve the above-mentioned problems.

[0005] Firstly, this application provides an iterative deployment method for directional signs based on the spatiotemporal distribution of pedestrian flow, the method comprising: A multi-source feature dataset of tourists is acquired. Based on this dataset, a dynamic construction of a crowd analysis map and behavioral assessment dimensions is performed to generate a set of crowd analysis maps and behavioral assessment dimensions. Based on the crowd analysis map and behavioral assessment dimension set, collaborative decision-making is made on adjusting the content and form of directional signs and intelligent diversion guidance to generate a set of sign adjustment and diversion strategies. Based on the set of sign adjustment and diversion strategies, tourist behavior feedback data is collected, and the generation process of the crowd analysis map and behavioral assessment dimension set is iteratively optimized in a closed loop to generate an iteratively updated directional sign knowledge base.

[0006] In an optional embodiment, the process of generating the crowd analysis map and the behavior assessment dimension set includes: the tourist multi-source feature dataset includes: spatiotemporal behavior data, visual behavior feature data, and personal interest profile data; based on the tourist multi-source feature dataset, individual tourist behavior tendencies and group aggregation patterns are identified to construct a crowd analysis map; based on the tourist multi-source feature dataset and the crowd analysis map, interest preferences and behavioral tendencies are fused and analyzed to construct the behavior assessment dimension set.

[0007] In one optional embodiment, the process of constructing the crowd analysis map includes: identifying subsets of tourists in a static gathering state and a fast-moving state by analyzing the dwell time and movement speed in the spatiotemporal behavior data; extracting the group gathering pattern of tourists and the height characteristics of children based on the visual behavior feature data, determining the tourist combination type within each tourist subset, wherein the tourist combination type includes traveling in groups, families with children, and solo travelers; matching the tourist combination type with the preference tags in the personal interest profile data to form several typical group categories, and statistically analyzing the proportion and distribution density of each typical group category in different sub-blocks of the target area, as the crowd analysis map.

[0008] In an optional embodiment, the process of constructing the behavioral assessment dimension set includes: based on the spatiotemporal behavioral data, extracting the degree of deviation of the tourist's trajectory path from the main road, and combining the tourist's facial orientation and handheld device posture in the visual behavioral feature data to assign an exploration dispersion tendency score to each individual tourist; by analyzing the matching degree between the preference tags in the personal interest profile data and the type of attraction at the current location, and combining the dwell time, assigning an interest concentration tendency score to each individual tourist; based on the spatiotemporal behavioral data, extracting the number of times tourists turn back and forth at intersections or signs and the duration of their hesitation, assigning a path hesitation score to each individual tourist; and statistically summarizing all the scores of all tourists according to the group categories in the crowd analysis map to generate the behavioral assessment dimension set.

[0009] In an optional embodiment, the process of generating the identification adjustment and diversion strategy set includes: identifying the typical group categories and their corresponding interest hotspot distributions for each sub-block within the target area based on the crowd analysis map; detecting the congestion risk status and tourist behavior contradictions for each sub-block based on the behavior evaluation dimension set; and making collaborative decisions on adjusting the content form of directional identification signs and intelligent diversion guidance by combining the interest hotspot distribution, the congestion risk status, and the tourist behavior contradictions, thereby generating the identification adjustment and diversion strategy set containing identification update instructions and diversion push information.

[0010] In one optional embodiment, the adjustment of the content format of the directional signage includes: when the crowd analysis map shows that the typical group category of a certain sub-block has changed, and the corresponding behavioral tendency index of the behavioral assessment dimension set exceeds a set threshold, the guidance focus of the current area signage is switched from the existing content to interest hotspots or service facility information that match the changed typical group category, and the visual presentation of the signage is adjusted simultaneously to adapt to the group characteristics; when the interest aggregation tendency score indicates that tourists in this area have a strong interest in a certain interest hotspot, dynamic guidance information pointing to such interest hotspots is added to the digital signage screen and tourists' personal mobile terminals.

[0011] In one optional embodiment, the intelligent diversion guidance includes: when the congestion risk status indicates that the waiting time or crowd density of a certain interest hotspot area exceeds the safe crowd threshold, diversion suggestions are simultaneously pushed to the digital signage screen at the entrance of this area and to the tourist's personal mobile terminal; the content of the diversion suggestions is customized according to the preference tags of the typical group category to which the tourist belongs in the crowd analysis map, that is, explicitly recommending other sub-blocks with similar interest attributes to the current hotspot and with lower current crowd density, and attaching the walking distance and the expected queuing time saved.

[0012] In an optional embodiment, the process of generating the wayfinding sign knowledge base includes: extracting diversion feedback indicators such as the acceptance rate of diversion suggestions, the change in regional flow speed, and the decrease in path repetition rate based on tourist behavior feedback data collected after the implementation of the sign adjustment and diversion strategy set; determining the positive or negative effect of the strategy based on the diversion feedback indicators, and generating the wayfinding sign knowledge base; if the determination is positive, storing the sub-block context and strategy parameters corresponding to the current strategy into the wayfinding sign knowledge base, and increasing the trigger weight of this strategy in similar scenarios; if the determination is negative, decreasing the trigger weight of this strategy, and adjusting the threshold parameters of the corresponding dimensions in the behavior evaluation dimension set.

[0013] In an optional embodiment, the collection and processing of the tourist behavior feedback data includes: comparing the spatiotemporal behavior data of each sub-block before and after the implementation of the identifier adjustment and diversion strategy set; calculating the increase in the average tourist movement speed and the decrease in the path repetition rate, respectively, as the change in the regional flow speed and the decrease in the path repetition rate; tracking whether the trajectory path of tourists who receive the diversion suggestion turns to the recommended alternative sub-block, and calculating the proportion of the number of people who turn to the total number of people who receive the suggestion, as the acceptance rate of the diversion suggestion.

[0014] This application provides an iterative deployment method for directional signs based on the spatiotemporal distribution of pedestrian flow. This method acquires a multi-source feature dataset of tourists, utilizing spatiotemporal behavioral data, visual behavioral feature data, and personal interest profile data. First, it identifies individual tourist behavioral tendencies and group aggregation patterns to construct a crowd analysis map. Then, it integrates interest preferences and behavioral tendencies to construct a behavioral assessment dimension set, thereby achieving a refined characterization of tourist group structure and a quantitative assessment of behavioral psychology. Based on this, the system identifies typical group categories and interest hotspot distributions according to the crowd analysis map, and combines the behavioral assessment dimension set to detect congestion risks and behavioral contradictions, collaboratively generating signage that includes… The system integrates updated instructions and diversion push information with a set of signage adjustments and diversion strategies. This allows wayfinding signs to dynamically adjust their content and visual presentation based on real-time group characteristics, and to push customized diversion suggestions that align with tourist preferences when congestion occurs. Subsequently, the system collects tourist behavior feedback data after strategy implementation, calculating indicators such as diversion suggestion acceptance rate, changes in regional flow speed, and the decrease in path repetition rate. Based on this, the system assesses the effectiveness of the strategies, storing successful experiences in the wayfinding signage knowledge base and increasing their trigger weights, or reducing the weights of failed strategies and adjusting threshold parameters. This completes a closed-loop iterative optimization process for generating crowd analysis maps and behavioral evaluation dimension sets. This effectively solves the problems of traditional wayfinding signage systems being unable to adapt to dynamic tourist composition and personalized needs, having limited guidance strategies, and lacking self-evolution capabilities. Therefore, it avoids localized congestion exacerbation and tourist path hesitation caused by mismatched guidance information, improving the responsiveness, diversion strategy acceptance rate, and overall operational reliability and adaptability of the wayfinding signage system. It represents a systematic upgrade from static, passive display to dynamic, proactive perception, decision-making, feedback, and continuous optimization.

[0015] In summary, this application constructs a dual analysis system through multi-source data fusion, realizing contextualized adaptive adjustment of signage content and form, personalized customization of traffic guidance, and establishing a closed-loop iterative mechanism based on measured feedback data. This ensures the logical integrity and systematic advantages of the technical solution, enabling the deployment of wayfinding signs to have the ability to continuously learn and optimize.

[0016] Secondly, this application provides an iterative deployment system for wayfinding signs based on the spatiotemporal distribution of pedestrian flow, the system comprising: The crowd profiling and behavior module is used to acquire a multi-source feature dataset of tourists. Based on the multi-source feature dataset, it dynamically constructs a crowd analysis map and behavior assessment dimensions, generating a set of crowd analysis maps and behavior assessment dimensions. The adjustment and diversion strategy module is used to make collaborative decisions on adjusting the content form of directional signs and intelligent diversion guidance based on the crowd analysis map and the behavior assessment dimension set, generating a set of sign adjustment and diversion strategies. The sign knowledge base module is used to collect tourist behavior feedback data based on the sign adjustment and diversion strategy set, and to perform closed-loop iterative optimization of the generation process of the crowd analysis map and the behavior assessment dimension set, generating an iterated directional sign knowledge base. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of an application scenario provided in an embodiment of this application; Figure 2 A flowchart illustrating an iterative deployment method for directional signs based on the spatiotemporal distribution of pedestrian flow, provided as an embodiment of this application; Figure 3 This is a flowchart of an iterative deployment system for directional signs based on the spatiotemporal distribution of pedestrian flow, provided as an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0020] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0021] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0022] Figure 1 This application provides an illustration of an application scenario. In the iterative deployment of directional signs, the method provided in this application can be used to achieve adaptive adjustment of directional signs based on the real-time composition and interests of tourist groups, thereby improving the accuracy and acceptance rate of diversion guidance and continuously optimizing the deployment effect through a closed-loop feedback mechanism.

[0023] Specifically, the method provided in this application can be applied to any server, where the server interacts with the scenic area's multi-dimensional monitoring system and the tourist's mobile client to obtain a multi-source feature dataset of tourists jointly provided by the scenic area's multi-dimensional monitoring system and the tourist's mobile client, and finally updates the wayfinding sign knowledge base to the wayfinding sign system.

[0024] The specific implementation method can be referred to in the following embodiments, wherein the data mentioned in the embodiments are only for reference and examples, so that relevant personnel can better understand them.

[0025] Figure 2 This is a flowchart illustrating an iterative deployment method for directional signs based on the spatiotemporal distribution of pedestrian flow, provided as an embodiment of this application. The method of this embodiment can be applied to servers in the above scenarios. For example... Figure 2 As shown, the method includes: Example

[0026] Existing wayfinding signage systems in tourist attractions mostly rely on static or periodically updated methods, failing to capture real-time information about the composition, individual interests, and dynamic behavioral patterns of the current visitor group. When the types of visitors within a scenic area are diverse, the uniform signage content is insufficient to meet the guidance needs of various groups. This is especially true at key locations such as popular photo spots and rest areas, where the lack of differentiated guidance often leads to localized congestion or visitors turning back, negatively impacting the visitor experience. Furthermore, traditional methods typically rely solely on crowd density or time distribution for simple guidance, failing to utilize the interest and preference information proactively provided by visitors during ticket purchases and QR code scanning, and lacking analysis of fine-grained characteristics such as visitor visual behavior. This results in untargeted flow guidance suggestions and the absence of a closed-loop iterative mechanism based on real-time behavioral feedback, hindering the continuous optimization of the signage system.

[0027] To address the aforementioned problems, this application provides an iterative deployment method for directional signs based on the spatiotemporal distribution of pedestrian flow. The method includes: Step 1: Obtain a multi-source feature dataset of tourists. Based on the multi-source feature dataset of tourists, dynamically construct the crowd analysis map and behavior assessment dimensions to generate a set of crowd analysis maps and behavior assessment dimensions. The multi-source tourist feature dataset refers to a comprehensive data set collected in real time by the system within the target area, encompassing tourists' spatiotemporal trajectories, visual behaviors, and personal preferences. This dataset specifically includes spatiotemporal behavioral data, visual behavioral feature data, and personal interest profile data. Spatiotemporal behavioral data is obtained through positioning beacons, wireless probes, or visual tracking devices, including the location coordinates, dwell time, movement speed, and trajectory path of each tourist, used to quantify the spatial distribution and flow of tourists. Visual behavioral feature data is extracted through image acquisition equipment, including tourists' facial orientation, handheld device posture, group aggregation patterns (such as spacing between groups), and children's height characteristics, used to identify tourists' immediate behavioral intentions and group structure. Personal interest profile data comes from tourists' preference tags selected during ticket purchase, browsing history on the official mini-program after scanning the QR code to enter the park, and publicly available interest information obtained with authorization from big data platforms, used to construct personalized interest models for tourists.

[0028] The crowd analysis map is a dynamic map constructed based on a multi-source tourist feature dataset, integrating the clustering distribution of tourists' age, appearance, height ratio, group composition, and personal interest tags. Its function is to classify tourists in the current area into several typical group categories (such as groups of tourists, families with children, and solo tourists), and dynamically record the proportion, spatial distribution density, and changing trends of each group. The behavior assessment dimension set is an assessment set generated based on the multi-source tourist feature dataset and the crowd analysis map, through comprehensive analysis of indicators such as tourist dwell time, movement speed, visual focus direction, path exploration (whether they deviate from the main path), and repeated return rate. Its function is to quantify various key behavioral tendencies in the current scenario, including interest concentration tendency, exploration dispersion tendency, fatigue and rest needs, and path hesitation.

[0029] For example, at the entrance of a large natural scenic area, the system obtained data from the ticketing platform showing that 40% of visitors that day were tagged as taking photos and checking in. Simultaneously, visual tracking devices detected a large number of visitors lingering at the entrance area, checking their mobile maps and frequently glancing at surrounding signs. The system fused this spatiotemporal behavioral data with visual behavioral feature data to identify a high proportion of interest-based clusters in the area, assigning them high interest clustering tendency scores. This generated a crowd analysis map containing the distribution hotspots of this group and a corresponding set of behavioral assessment dimensions. Through this fusion and dynamic construction of multi-source data, the system significantly improved its accuracy in perceiving crowd structure and behavioral intentions in complex tourism scenarios, providing a solid data foundation for subsequent differentiated guidance.

[0030] Step 2: Based on the crowd analysis map and behavior assessment dimension set, make collaborative decisions on adjusting the content form of directional signs and intelligent diversion guidance, and generate a set of sign adjustment and diversion strategies; The adjustment of directional signage content and form refers to the process of dynamically changing the presentation content and physical form of directional signage based on the dominant group categories and their behavioral tendencies identified in the crowd analysis map. Specifically, this includes: when the dominant group in a certain area shows a specific interest clustering tendency, adding guidance information pointing to this interest hotspot area and downplaying irrelevant information in the signage; when the group shows fatigue and a need for rest, increasing the prominence of signs for rest areas and service facilities; when the group has specific age characteristics (such as a large proportion of elderly people), adjusting the font size, contrast, and graphic complexity of the signage, and even using auxiliary interactive methods such as voice prompts or ground projection lighting. Intelligent diversion guidance refers to the process of simultaneously pushing customized diversion suggestions to the physical signage at relevant entrances and to tourists' personal mobile terminals when the behavioral assessment dimension set detects that the flow density or waiting time in a specific interest hotspot area exceeds a safety threshold. The content of the diversion suggestions is generated based on the preference tags of the typical group categories to which tourists belong in the crowd analysis map, explicitly recommending other sub-blocks with similar interest attributes but currently lower flow density, along with walking distances and expected savings in queuing time.

[0031] The identifier adjustment and diversion strategy set is the output of the aforementioned collaborative decision-making. It includes specific identifier update instructions (such as switching screen content, adjusting light color, and changing font style) and diversion push information (such as push notifications of alternative routes to specific user groups). The generation of this strategy set relies on the group context provided by the crowd analysis graph and the risk warnings provided by the behavioral assessment dimension set. The two work together to enable the system not only to perceive the real-time composition of the crowd in front of it, but also to combine individual interest data and dynamic behavioral tendencies to provide a comprehensive guidance solution.

[0032] For example, when the system detects congestion at a photo spot, and crowd analysis shows that the area is mainly frequented by people taking photos, the collaborative decision-making module generates a strategy: firstly, it adjusts the digital signage at the entrance of the spot to a yellow warning light and highlights the real-time lower crowd levels at a similar spot to the west on the screen; secondly, it pushes a message to tourists' mobile phones in the area, guiding them to discover a new hidden gem, emphasizing the similar experience and shorter queue times at the western spot. This synergy between content format adjustments and intelligent crowd control effectively avoids the guidance failure caused by static signage, significantly improving the accuracy of crowd control and tourist acceptance.

[0033] Step 3: Based on the set of signage adjustment and diversion strategies, collect tourist behavior feedback data, and perform closed-loop iterative optimization on the generation process of crowd analysis map and behavior assessment dimension set to generate an iterated wayfinding signage knowledge base.

[0034] The tourist behavior feedback data refers to the data set collected in real time by the system after implementing the aforementioned signage adjustment and diversion strategy set, reflecting the effectiveness of the strategy. Specifically, it includes the diversion suggestion acceptance rate (obtained by tracking whether the trajectory of tourists who received the suggestion turned to the recommended alternative sub-block), the change in regional flow speed (comparing the increase in the average movement speed of tourists before and after the strategy implementation), and the decrease in path repetition rate (comparing the decrease in the rate of tourists getting lost and turning back before and after the strategy implementation). The closed-loop iterative optimization of the generation process of the crowd analysis map and behavior evaluation dimension set can refer to using the aforementioned feedback data to determine whether the effectiveness of this strategy is positive or negative, and adjusting the initial modeling parameters and strategy weights accordingly. If the determination is positive (i.e., high acceptance rate and relief of congestion), the sub-block scenario context and strategy parameters corresponding to the current strategy are stored in the guidance signage knowledge base, and the trigger weight of this strategy under similar scenarios is increased; if the determination is negative, the trigger weight of this strategy is decreased, and the threshold parameters of the corresponding dimensions in the behavior evaluation dimension set are adjusted.

[0035] The guidance signage knowledge base is a database that stores historically optimal strategies and their applicable contexts. Its purpose is to enable the system to continuously learn and adapt. By storing the context of each adjustment (such as date, time period, weather, and passenger flow level) and the final effect in the knowledge base, when the same scenario occurs in the future, the system can directly call the historically optimal strategy for rapid configuration, achieving real-time closed-loop iteration.

[0036] For example, if, after implementing a gamified guidance strategy targeting families with children, data shows a 20% decrease in path repetition and an 85% acceptance rate for diversion suggestions, the system considers this a positive effect. It automatically binds the strategy's parameters (such as cartoon icon style and 50-meter interval directional points) to the scene characteristics and stores them in the knowledge base, prioritizing the reuse of this strategy during similar peak family-oriented periods. Conversely, if an adjustment leads to a decrease in visitor satisfaction, the system automatically reduces the strategy's weight and recalibrates relevant thresholds. Thus, the system constructs a complete closed-loop mechanism of data collection, analysis, adjustment, feedback, and iteration, continuously optimizing the deployment effect without manual intervention. Furthermore, the experience gained from each adjustment can be accumulated and reused, achieving self-evolution of the wayfinding signage system.

[0037] This application achieves refined perception of tourist group structure and behavioral intentions by acquiring multi-source feature datasets of tourists and dynamically constructing crowd analysis maps and behavioral assessment dimension sets. Based on this, it makes collaborative decisions on adjusting the content and form of directional signage and intelligent diversion guidance, enabling guidance strategies to adaptively match the personalized needs of different groups. Furthermore, it performs closed-loop iterative optimization of the analysis process based on tourist behavior feedback data, generating an iteratively updated directional signage knowledge base, ensuring that the system can continuously accumulate experience and improve decision-making accuracy over time. Through the synergistic effect of the above technical features, this application not only solves the problem that traditional directional signage systems cannot adaptively adjust according to real-time pedestrian flow composition and behavioral characteristics, but also significantly improves the accuracy of guidance, tourist acceptance rate, and overall system response efficiency by integrating individual interest profiles into group analysis and closed-loop feedback mechanisms.

[0038] Example 2: In one possible implementation, the process for generating a population analysis map and a behavioral assessment dimension set provided in this application embodiment includes: Step 1: The tourist multi-source feature dataset includes: spatiotemporal behavior data, visual behavior feature data, and personal interest profile data; The multi-source tourist feature dataset refers to the set of basic input information relied upon by the system during the iterative deployment of wayfinding signs. Its sources include physical sensing devices, network interaction records, and third-party data platforms. Specifically, spatiotemporal behavioral data is collected in real-time by positioning beacons, wireless probes, or visual tracking devices deployed within the scenic area, describing tourists' movement trajectories, dwell time, and instantaneous speed in physical space. Visual behavioral feature data is obtained through non-contact capture and extraction of tourists using image acquisition devices, including fine-grained visual information such as facial orientation, handheld device posture, group aggregation patterns, and children's height characteristics. Personal interest profile data is constructed by fusing tourists' actively selected preference tags during ticket purchase, their browsing history in official mini-programs or applications after scanning the QR code to enter the park, and publicly available interest information obtained from authorized big data platforms. These three types of data work together: spatiotemporal behavioral data provides objective facts about where they are and how they are moving; visual behavioral feature data supplements the on-site state of what they are looking at and who they are with; and personal interest profile data reveals the psychological motivations behind their preferences. For example, spatiotemporal behavioral data can record that a tourist stayed at attraction A for 15 minutes; visual behavioral feature data can identify that the tourist is an adult male with a child who is 1.1 meters tall; and personal interest profile data shows that the tourist prefers parent-child interaction and nature education. Through this combination of multi-source data, the system can establish a comprehensive characteristic description of the tourist, providing a solid data foundation for subsequent analysis. This step aims to complete the aggregation and standardization of multi-dimensional heterogeneous data, ensuring that subsequent analysis models can simultaneously utilize objective behavioral trajectories and subjective interest preferences, thereby avoiding judgment bias caused by a single data source.

[0039] Step 2: Based on the multi-source feature dataset of tourists, identify individual tourist behavior tendencies and group aggregation patterns, and construct a crowd analysis map; Crowd analysis mapping refers to a dynamically generated logical mapping relationship reflecting the structural distribution and spatial aggregation status of tourist groups within a target area. The construction of this mapping is based on the aforementioned spatiotemporal behavioral data, visual behavioral feature data, and personal interest profile data, achieved through clustering analysis and pattern recognition algorithms. Specifically, the system first utilizes movement speed and dwell time from the spatiotemporal behavioral data, combined with group aggregation patterns (such as multiple people walking side-by-side, following each other) and children's height characteristics from the visual behavioral feature data, to identify subsets of tourists in a static aggregation state or a fast-moving state, and determines the tourist combination type within each subset, such as traveling in groups, families with children, or solo travelers. Subsequently, the system matches the identified tourist combination types with preference tags from the personal interest profile data to form several typical group categories, and statistically analyzes the proportion and distribution density of each typical group category in different sub-blocks of the target area. Through this construction method, the crowd analysis mapping not only describes how many people there are, but also clarifies what kind of people they are and where they are, achieving an abstraction from discrete individual data to macroscopic group characteristics. The output of this step provides a clear target basis for subsequent adjustments to guidance signage, enabling the system to adopt differentiated guidance strategies for groups with different characteristics.

[0040] Step 3: Based on the multi-source feature dataset of tourists and the population analysis map, conduct a fusion analysis of interest preferences and behavioral tendencies to construct a set of behavioral assessment dimensions.

[0041] The behavioral assessment dimension set refers to a set of indicators used to quantitatively assess tourists' psychological state and behavioral trends in the current scenario. Its construction relies on deep coupling analysis of multi-source tourist feature datasets and a pre-constructed crowd analysis map. Specifically, the system compares preference labels in personal interest profile data with actual trajectories in spatiotemporal behavioral data to analyze the matching degree between interest preferences and the type of attractions at the current location, thereby inferring tourists' interest concentration tendencies. Simultaneously, it assesses tourists' exploration dispersion tendencies by combining facial orientation and handheld device posture in visual behavioral feature data, as well as path deviation in spatiotemporal behavioral data. Furthermore, it quantifies tourists' path hesitation by statistically analyzing the number of times they turn back and forth at intersections or signs and the duration of their hesitation. These dimensions are not isolated but are hierarchically statistically summarized based on group categories in the crowd analysis map. For example, for families with children identified in the crowd analysis map, the system focuses on analyzing their dwell time near rest areas and fatigue-related behaviors, assigning a higher fatigue and rest need score; while for the youth-adventure group, it focuses on the frequency of their deviation from the main path, assigning a higher exploration dispersion tendency score. Through this fusion analysis, the behavioral assessment dimension set can accurately reflect the real needs and potential risks of different groups in specific scenarios, such as congestion risk, getting lost risk, or service deficiency risk. This step aims to transform static group classification into dynamic behavioral assessment indicators, providing direct decision parameters for generating targeted guidance signage adjustments and diversion strategies. This ensures that guidance solutions not only meet the interests and preferences of groups but also effectively alleviate potential behavioral conflicts.

[0042] This application establishes a complete data-driven analysis system through the collaborative construction of multi-source tourist feature datasets, crowd analysis maps, and behavioral assessment dimension sets in the aforementioned steps. By complementing and integrating spatiotemporal behavioral data, visual behavioral feature data, and individual interest profile data, the system can not only accurately identify the group composition of tourists (i.e., crowd analysis maps) but also gain in-depth insights into the psychological motivations behind their behavior (i.e., behavioral assessment dimension sets). This two-layer analysis mechanism allows the placement of wayfinding signage to move beyond fixed rules or simple crowd counting, relying instead on a deep understanding of who is doing what, where, and why. Based on this, the crowd analysis map provides a grouping benchmark for behavioral assessment, enabling the differentiated quantification of behavioral tendencies among different groups; while the behavioral assessment dimension sets, in turn, verify and enrich the dynamic meaning of the crowd analysis map. Together, these two elements constitute the core basis for the intelligent iteration of wayfinding signage, significantly improving the accuracy of traffic guidance and tourist acceptance, effectively solving the technical problem that traditional static signage cannot adapt to the complex and ever-changing tourist composition.

[0043] Example 3: In another optional embodiment, a population analysis map construction process provided for embodiments of this application includes the following steps: Step 1: By analyzing the dwell time and movement speed in the spatiotemporal behavior data, identify the subsets of tourists who are in a static gathering state and a fast-moving state; The system defines a "static gathering" state as a movement pattern where tourists move at a speed below a preset low-speed threshold and stay for a cumulative duration exceeding a preset time threshold within a specific geographic boundary. A "fast movement" state refers to a movement pattern where tourists consistently move at a speed above a preset high-speed threshold and exhibit significant displacement per unit time. The system acquires the coordinate sequence of each tourist in real time using positioning beacons or visual tracking devices, calculating their average speed and dwell time within a sliding time window. For example, setting a low-speed threshold of 0.2 m / s and a time threshold of 3 minutes, if a group's average movement speed is 0.1 m / s and they stay in the same area for more than 5 minutes, the system classifies it as a static gathering state. This typically corresponds to tourists taking photos at a viewing platform, resting in a rest area, or waiting in line. Conversely, if tourists move at a speed greater than 1.5 m / s and their trajectories are continuous, it is identified as a fast movement state, corresponding to tourists traveling along a main road or heading to the next attraction. This initial screening based on spatiotemporal behavioral data allows for the division of a large tourist group into subsets with different dynamic characteristics, providing fundamental data support for subsequent refined group type determination and effectively filtering out invalid background noise data.

[0044] Step 2: Based on visual behavior feature data, extract the group aggregation pattern of tourists and the height characteristics of children to determine the tourist combination type within each tourist subset. Tourist combination types include group tours, families with children, and solo tours. The group aggregation pattern refers to the spatial relative positions, spacing density, and synchronicity of movement among tourists captured by image acquisition equipment; the children's height characteristics refer to the individual height values ​​detected by computer vision algorithms or the height ratio differences with surrounding adults. Based on the above-identified tourist subsets, the system calls on camera video streams for in-depth analysis. Specifically, for a subset in a static aggregation state, if multiple individuals are detected to be less than 1 meter apart and have the same facial orientation or frequent interactive postures, it is determined to be a group tour; if at least one individual is detected to be less than 1.4 meters tall (or shorter than the shoulder height of surrounding adults), and this shorter individual maintains a close following relationship with a specific adult, it is determined to be a family with children; for a subset in a fast-moving state, if only a single independent individual is detected without any accompanying persons, it is determined to be a solo tour. For example, at the entrance of a family amusement area, the system identified a group of three, two of whom were of adult height and one was significantly shorter and walking hand in hand. Combined with their close spatial aggregation pattern, the system accurately determined them to be a family with children, rather than a typical group of three. By incorporating the morphology and height dimensions from visual behavioral features, the system can penetrate simple location data and deeply analyze the social relationship structure of tourists, thereby avoiding misclassifying temporary gatherings of strangers as family units and significantly improving the granularity and accuracy of group classification.

[0045] Step 3: Match tourist combination types with preference tags in personal interest profile data to form several typical group categories, and count the proportion and distribution density of each typical group category in different sub-blocks of the target area to form a population analysis map.

[0046] The typical group category is a segmented user profile formed by the cross-integration of tourist combination types and personal interest preference tags; the quantity proportion can refer to the proportion of a certain typical group in the total number of people in a specific sub-block; and the distribution density can refer to the number of such groups per unit area. The system correlates and matches the tourist combination types (such as families with children) determined above with the personal interest profile data (such as tags for preferring natural scenery, science education, or thrill rides) provided by the tourist during the ticket purchase or registration process. For example, when a group identified as a family with children has members whose interest tags are mainly focused on science education, the system classifies them as a typical family-oriented science education group; if another group is also a family with children but their interest tags are for thrill rides, they are classified as a typical family-oriented adventure group. Subsequently, the system divides the target scenic area into several grid-like sub-blocks, counts the number of each type of typical group in each sub-block in real time, calculates their proportion of the total flow of people in that area, and generates a distribution density heat map based on the area of ​​the sub-blocks. Through this multi-dimensional matching mechanism, the resulting crowd analysis map not only shows where people are concentrated, but also clearly reveals what kind of people are in those areas and what they like. For example, in a museum exhibition area, the map shows that the distribution density of parent-child science-learning groups is as high as 80%, while in the roller coaster area, young adventure groups, whether alone or in groups, dominate. This map serves as the core basis for subsequent adjustments to wayfinding signage, enabling the system to push highly matched guidance information based on the dominant group characteristics of different sub-areas, thus achieving a shift from extensive crowd monitoring to refined crowd management.

[0047] This application constructs a three-tiered, progressive crowd analysis mechanism—initial screening of behavioral states, confirmation of visual structure, and refinement of interest attributes—through the synergistic effect of the aforementioned steps. First, by utilizing dwell time and movement speed data from spatiotemporal behavioral data, it quickly identifies subsets of tourists with potential clustering or traffic characteristics, significantly reducing the processing load of the entire dataset. Building upon this, by leveraging group clustering patterns and children's height characteristics from visual behavioral feature data, it accurately identifies social relationship structures such as families, groups, and individuals, solving the problem that traditional methods cannot distinguish group nature based solely on location. Finally, it introduces preference tags from individual interest profile data for deep matching, elevating the physical-level group classification to a psychological and demand-level classification of typical categories. The combined use of these three elements allows the generated crowd analysis map to simultaneously reflect tourists' dynamic behavioral patterns, social composition structure, and intrinsic interest needs. Thus, the system can accurately identify specific scenarios such as highly fatigued parent-child families or highly exploratory single young people, providing quantitative and precise decision-making input for subsequent differentiated content presentation of directional signage and intelligent diversion strategies, effectively avoiding mismatched guidance information caused by group identification bias.

[0048] Example 4: In one possible implementation, the method for constructing a set of behavioral assessment dimensions provided in this application embodiment further includes quantitative scoring of individual tourist behavioral tendencies and group aggregation steps.

[0049] Step 1: Based on spatiotemporal behavioral data, extract the degree of deviation of tourists' trajectory paths from the main road, and combine the tourists' facial orientation and handheld device posture in the visual behavioral feature data to assign an exploration dispersion tendency score to each individual tourist. The exploration dispersion tendency score is a quantitative indicator used to characterize the likelihood of tourists exploring off-track, searching for new attractions, or deviating from the planned tour route within the current area. This score is calculated based on multi-source data fusion. Specifically, the system first analyzes the tourist's real-time trajectory from spatiotemporal behavioral data and calculates the geometric deviation distance and angle between this path and the scenic area's preset main road or recommended route. Simultaneously, it uses visual behavioral feature data acquired by image acquisition equipment to identify whether the tourist's facial orientation frequently scans the surrounding environment instead of focusing on the path ahead, and whether the handheld device posture indicates frequent viewing of electronic maps or searching for information. These multi-dimensional features work together in the scoring model. When a tourist's trajectory significantly deviates from the main road and is accompanied by frequent environmental scanning or map querying behavior, the system will assign a high exploration dispersion tendency score. For example, in a historical and cultural district, if a tourist is detected stopping multiple times on the main road and looking left and right, while their mobile phone screen continuously displays a navigation interface, and their actual walking trajectory deviates from the main road by more than 5 meters in a jagged pattern, the system will determine that the tourist has a strong exploration dispersion tendency and assign a high score of 0.85 (out of 1.0). By combining trajectory deviation and visual attention direction into a joint judgment mechanism, it is possible to accurately distinguish between aimless wandering and purposeful exploration behavior, thereby providing data support for subsequent push of guidance information on discovering niche attractions for such tourists.

[0050] Step 2: By analyzing the match between preference tags in personal interest profile data and the types of attractions in the current location, and combining the length of stay, assign an interest concentration tendency score to each individual tourist. The interest-gathering tendency score is a quantitative indicator used to represent a tourist's strong interest in the current area or specific attractions and their tendency to linger and engage in in-depth experiences. This score is derived from cross-analysis of personal interest profile data and real-time location data, specifically including two core dimensions: first, semantic matching, which compares the degree of overlap between the preference tags (such as ancient architecture, natural scenery, parent-child interaction) selected by the tourist when purchasing tickets or registering and the attribute tags of attractions around the current location; second, behavioral duration, which counts the continuous time the tourist stays at the current coordinate point. When the preference tags are highly matched and the dwell time exceeds a set threshold, the system will assign a higher interest-gathering tendency score. For example, for a tourist tagged with ceramic art, if the system detects that they linger in front of a certain exhibit for more than 3 minutes after entering the ceramics museum exhibition area, and the exhibit type perfectly matches their tag, the system will assign them an interest-gathering tendency score of 0.9. Conversely, if the tourist lingers but the tag does not match (such as being forced to wait for companions), the score will be lower. This scoring mechanism can effectively identify potential congestion risk points, because a high interest clustering tendency often indicates that the area is about to experience a traffic jam, requiring the system to intervene in advance to divert and guide the flow of people or to increase guidance from relevant service facilities.

[0051] Step 3: Based on spatiotemporal behavioral data, extract the number of times tourists turn back and forth at intersections or signs and the duration of their hesitation, and assign a path hesitation score to each individual tourist. The path hesitation score is a quantitative indicator used to characterize tourists' hesitation and repeated confirmation behaviors at path decision points due to missing information, unclear signage, or directional confusion. This score is obtained by mining specific movement patterns in spatiotemporal behavioral data, focusing on extracting the movement trajectory characteristics of tourists at intersections, T-junctions, or in front of large directional signs. Specifically, it includes the number of back-and-forth movements per unit time (i.e., the number of times moving back and forth within a short distance) and the duration of ineffective lingering within the decision-making area. When a tourist is detected repeatedly changing direction at an intersection or lingering in place for more than a preset time (e.g., 30 seconds), the system will assign a high path hesitation score. For example, at a three-way intersection in a scenic area, if a tourist is detected moving back and forth three times within one minute and lingering in the center of the intersection for 45 seconds without making a clear turn, the system will determine that the tourist has serious path hesitation and assign a high score of 0.8. This score directly reflects the insufficient guidance effectiveness of the current directional signs at that location. A high score cluster indicates that the signage layout urgently needs optimization or dynamic enhancement of guidance information to eliminate tourist confusion and improve traffic efficiency.

[0052] Step 4: Statistically summarize all the ratings of all tourists according to the group categories in the crowd analysis map to generate a set of behavioral assessment dimensions.

[0053] The behavioral assessment dimension set refers to aggregating individual-level quantitative scores into a macro-level behavioral trend dataset at the group level, used to guide the generation of regional wayfinding signage strategies. The generation of this dataset relies on calculating three scores for each visitor (exploration dispersion tendency, interest concentration tendency, and path hesitation) in the aforementioned steps. Based on the population analysis map constructed in the above embodiments, visitors are divided into different typical group categories (such as young people traveling in groups, families with children, and elderly people traveling alone). The system then performs statistical processing on all individual scores within each category, calculating the mean, variance, or distribution density to form dimension vectors for different group characteristics. For example, for families with children, the system statistically found that their mean path hesitation score in the amusement area was 0.7, while their mean interest concentration tendency score was 0.4, indicating that this group generally had difficulty finding their way and had only a moderate interest in the current facilities. In contrast, for photography enthusiasts, the mean interest concentration tendency score in the same area was as high as 0.85. Through this classification and aggregation, the generated behavioral assessment dimension set not only retains the quantitative characteristics of the behavior, but also endows it with group attributes, so that subsequent adjustments to directional signs can be targeted. For example, intersection signs can be optimized specifically for family groups, while in-depth introductions of scenic spots can be added for photography groups, thus achieving a leap from individual perception to group decision-making.

[0054] This application, through the aforementioned steps, transforms abstract tourist behavior patterns into calculable scores for exploration dispersion tendency, interest aggregation tendency, and path hesitation, and further aggregates these scores by group category to form a set of behavioral assessment dimensions. Through the synergy of spatiotemporal behavioral data and visual behavioral feature data, the system can accurately capture tourists' micro-movements (such as facial orientation and turning back), and, combined with macro-level preferences from individual interest profiles, achieve in-depth quantification of tourists' psychological states and behavioral intentions. This quantification mechanism works closely with the group classification function of the crowd analysis map, enabling the system not only to identify the confusion or interests of individual tourists, but also to gain insight into the common behavioral trends of a certain group in a specific area. Thus, the wayfinding signage system can shift from passive information display to proactive behavioral intervention. For example, when a group's path hesitation score is detected to be generally high, the system automatically triggers an enhanced signage mode for that area; or when the interest aggregation tendency score is too high, a diversion plan is activated in advance. This effectively solves the technical problem that traditional signage systems cannot perceive tourists' real-time psychological states and group differences, significantly improving the efficiency of wayfinding in scenic areas and the tourist experience.

[0055] Example 5: In another optional embodiment, a process for generating an identifier adjustment and traffic splitting strategy set provided in this application includes the following steps: Step 1: Based on the crowd analysis map, identify the typical group categories and their corresponding interest hotspot distribution areas for each sub-block within the target area; The crowd analysis map is a dynamic data structure constructed by integrating spatiotemporal behavioral data, visual behavioral feature data, and personal interest profile data. It stores the classification statistics of tourist groups within each sub-block after the target area is divided. Typical group categories can refer to standardized classifications formed by matching tourist combination types (such as group travel, families with children, and solo travelers) with personal interest preference tags, such as parent-child recreation preference type or youth photo-taking type. The distribution of interest hotspot areas can refer to the geofence of high attention exhibited by each typical group in space, which is calculated based on historical dwell time density and real-time aggregation degree.

[0056] Specifically, the system divides the target scenic area into several logical sub-blocks, traverses the data records of each sub-block in the crowd analysis map, and extracts the most prevalent typical group category as the dominant group of that block. Simultaneously, it retrieves the trajectory heatmap of this dominant group within the current time and a preset time window to determine the specific physical areas where their interests are focused, i.e., interest hotspots. For example, when the dominant group of a sub-block is identified as having a parent-child / entertainment preference, the system automatically associates the coordinates of frequently visited children's playgrounds or interactive experience areas with this group, marking them as the interest hotspot distribution of that sub-block. Through this identification mechanism, the system can accurately locate the core service targets and their focus of attention in different areas, providing a spatial and target basis for subsequent differentiated guidance.

[0057] Step 2: Based on the behavioral assessment dimension set, detect the contradictions between the congestion risk status and tourist behavior in each sub-block; The behavioral assessment dimension set is a multi-dimensional set of indicators including exploration dispersion tendency score, interest aggregation tendency score, and path hesitation score. Congestion risk status refers to the regional carrying capacity pressure level determined based on current pedestrian density, movement speed decay rate, and predicted waiting time, typically categorized into low-risk, medium-risk (warning), and high-risk congestion levels. Tourist behavior contradictions refer to locations or phenomena where there is a significant deviation between a tourist's actual behavioral trajectory and the expected guided path or interest demand, such as abnormal backtracking under a high interest aggregation tendency score, or high path hesitation in a low-congestion area.

[0058] Specifically, the system reads the scoring data of each sub-block in the behavioral assessment dimension set in real time and sets dynamic thresholds for status determination. If the interest aggregation tendency score of a sub-block exceeds the set threshold and the average movement speed is lower than the critical value, the area is determined to be in a state of congestion risk. At the same time, by analyzing the spatial distribution of path hesitation scores, areas where tourists frequently linger at intersections or in front of signs are identified and marked as tourist behavior contradiction points. For example, at the entrance of a cultural and creative store, if a large number of tourists are detected to have extremely high interest aggregation tendency scores, but are also accompanied by a high proportion of turning back and lingering time, the system determines that there is a high intention-low passage behavioral contradiction point, indicating that the existing guidance has failed to effectively guide high-intention groups. This step aims to quantitatively assess the operational health and guidance effectiveness of each area and promptly identify potential risks of congestion and guidance failure points.

[0059] Step 3: Based on the distribution of interest hotspots, congestion risk status, and conflicting tourist behaviors, make collaborative decisions on adjusting the content and form of directional signs and intelligent diversion guidance, and generate a set of sign adjustment and diversion strategies that include sign update instructions and diversion push information; Collaborative decision-making refers to the process of generating optimal guidance solutions by coupling spatial distribution characteristics (hotspots of interest), temporal status characteristics (congestion risk), and behavioral and psychological characteristics (points of behavioral conflict) in a multi-dimensional manner. Signage update instructions are control signals used to change the displayed content, font size, color contrast, or arrow direction of physical electronic signage screens or projection devices. Traffic diversion push information is text or multimedia data containing alternative routes, estimated time savings, and reasons for recommendation sent to tourists' personal mobile terminals.

[0060] Specifically, the system establishes a decision matrix to match the three types of input information mentioned above. When the distribution of interest hotspots in a sub-block points to a specific attraction, and a congestion risk is detected at the entrance of that attraction, the decision logic triggers intelligent diversion guidance, generating diversion push information and suggesting that some tourists go to alternative sub-blocks with similar interest attributes but currently lower visitor flow. At the same time, if a point of conflict in tourist behavior (such as high hesitation) is detected in the area, a signage update instruction is generated simultaneously to adjust the visual presentation of on-site signs (such as increasing arrow size, adding dynamic flashing effects) or supplementing temporary guidance content to eliminate tourist confusion. For example, in a scenario where families with children with a preference for amusement park activities gather and there is congestion near a children's playground, the system generates a set of strategies including: pushing a message to families with children that there is no need to queue in the western adventure area, and instructing the on-site signage screens to switch to cartoon-style dynamic guidance, clearly indicating the westward path. Through the three-way collaboration of interest hotspots, congestion status, and behavioral conflict points, this step realizes the transformation from passive response to proactive intervention, ensuring that the generated strategy set not only meets the needs of tourist interests but also effectively alleviates traffic pressure.

[0061] This application, through the aforementioned steps, achieves the organic synergy between adjusting the content and form of directional signage and intelligent traffic diversion guidance. Based on the distribution of typical group categories and interest hotspots identified through crowd analysis mapping, the targeted nature of the guidance content is ensured, solving the problem of generic signage. Combining congestion risk status and conflicting tourist behaviors detected by behavioral assessment dimension sets allows the system to intervene promptly before congestion occurs or when tourists experience confusion, improving the timeliness of intervention. The distribution of interest hotspots provides directional guidance for traffic diversion, congestion risk status provides triggering opportunities, and conflicting tourist behaviors optimize the specific form of guidance. The combined effect of these three elements allows the generated signage adjustment and traffic diversion strategy set to dynamically adapt to the complex and ever-changing tourist flow environment in scenic areas. This collaborative decision-making mechanism not only improves the acceptance rate of traffic diversion suggestions but also constructs a three-dimensional guidance network through the complementarity of physical signage and mobile terminals. This effectively avoids the exacerbation of localized congestion caused by information conflicts or lack of guidance, significantly improving the overall traffic efficiency of the scenic area and the tourist experience.

[0062] Example 6: In another optional embodiment, a method for adjusting the content format of a directional sign provided in this application includes the following steps: Step 1: When the crowd analysis map shows that the typical group category of a certain sub-block has changed, and the corresponding behavioral tendency index of the behavior assessment dimension exceeds the set threshold, the guidance focus of the current area label will be switched from the existing content to interest hotspots or service facility information that match the typical group category after the switch, and the visual presentation of the label will be adjusted simultaneously to adapt to the group characteristics. This step refers to the system triggering a dual adaptive adjustment mechanism for both the content and form of the identifier when it detects a substantial change in the dominant tourist group type within the target area, accompanied by an excess of specific behavioral tendencies. A change in typical group category could mean a shift in the most prevalent tourist combination type within the current sub-block, such as a change from predominantly young couples to families with children, or from primarily solo travelers to primarily elderly groups. The corresponding behavioral tendency indicators within the behavioral assessment dimensions can be scoring data strongly correlated with the changed group characteristics, such as fatigue and rest needs scores, route hesitation scores, or interest concentration tendency scores. The threshold setting is a pre-set critical value based on historical data statistics or on-site management needs, used to filter out occasional fluctuations and ensure the accuracy of the adjustment.

[0063] Shifting the focus of signage guidance from existing content to information on areas of interest or service facilities that align with the new target demographic can mean dynamically prioritizing displayed information based on the new dominant group profile. Specifically, if the dominant group is identified as elderly tourists with fatigue and rest needs exceeding a threshold, the system replaces the previously displayed scenic descriptions or directions to distant attractions with location information for nearby rest seats, restrooms, accessible pathways, and medical points. If the dominant group is identified as families with children and a high level of route hesitation, the guidance focus shifts to children's play facilities, mother-and-baby rooms, and safety warnings. Simultaneously adjusting the visual presentation of signage to suit the group's characteristics can involve changing the physical or digital display parameters to improve readability and user-friendliness. For example, for the elderly, the system automatically increases font size, enhances background and text color contrast, simplifies graphic elements, and reduces redundant text; for families, it introduces cartoon characters, vibrant colors, and interactive animation elements.

[0064] This adjustment method, based on both group switching and behavioral indicators, ensures that signage information always matches the most pressing needs on-site, avoiding guidance failures caused by information misalignment. This step aims to address the problem of static signage failing to adapt to dynamic changes in pedestrian flow. By sensing group characteristics in real time and responding immediately, it significantly improves the efficiency of information access and the sense of security for special groups in complex environments.

[0065] Step 2: When the interest clustering tendency score indicates that tourists in this area have a strong interest in a certain hotspot, add dynamic guidance information pointing to such hotspots on digital signage screens and tourists' personal mobile devices.

[0066] This step can refer to proactively enhancing the prominence and coverage of guidance information when the system detects that tourists are showing a high level of interest in a specific area or facility. The interest concentration tendency score is a quantitative indicator calculated based on a combination of factors including tourist dwell time, facial orientation concentration, frequency of handheld device photography, and trajectory convergence, used to characterize the level of tourist attention to the area. Strong attention can be defined as the score reaching or exceeding a preset high-heat threshold, indicating that the area has become a current traffic hotspot or a potential congestion point.

[0067] Adding dynamic guidance information pointing to such interest hotspots on digital signage screens and tourists' personal mobile devices can refer to building a three-dimensional guidance network combining fixed and mobile terminals. Specifically, on digital signage screens at sub-block entrances or key intersections, the system automatically overlays highlighted dynamic arrows, flashing light strips, or flowing particle effects to intuitively indicate the path to the interest hotspot, and even marks the hotspot's location on the map with a breathing light effect. At the same time, through Bluetooth beacons or wireless networks, pop-up notifications or mini-program cards containing the same directional information are pushed to tourists' personal mobile devices (such as smartphones and tablets) within the area, achieving precise information delivery. For example, when the system detects a sharp increase in tourists' interest in a certain popular photo spot, it not only projects a striking green flowing arrow on the intersection's large screen to guide the way, but also sends a dynamic prompt to nearby tourists' mobile phones that the crowd at the photo spot ahead is moderate and they can click to navigate there.

[0068] By combining the wide-area visual impact of digital screens with personalized and precise push notifications from mobile devices, high-density crowds can be effectively managed, reducing disorderly wandering caused by searching for destinations. This result provides a proactive intervention method for subsequent congestion risk prevention, thereby significantly improving traffic efficiency in popular areas and enhancing the smoothness of the visitor experience when searching for points of interest.

[0069] This application, through the synergistic effect of the aforementioned technical features, achieves a leap from passive display to proactive adaptation in the wayfinding signage system. By capturing real-time shifts in typical group categories through crowd analysis mapping and combining this with quantitative indicators focused on behavioral assessment dimensions for dual verification, the timing of signage content adjustments is ensured to be accurate and well-founded. Furthermore, the simultaneous implementation of content focus switching and visual presentation optimization ensures that information delivery meets both the functional needs of the group (e.g., finding restrooms, finding attractions) and their physiological and psychological characteristics (e.g., large fonts, cartoons), significantly reducing cognitive load. Moreover, leveraging interest-based clustering tendency scoring to keenly perceive high-traffic areas, the system adds dynamic guidance information through dual-terminal linkage between digital signage and mobile terminals. This not only enhances the penetration of guidance signals but also creates spatial three-dimensional coverage, effectively dispersing local attention bottlenecks. This multi-dimensional collaborative mechanism enables wayfinding signage to act like an intelligent assistant, dynamically reconstructing its expression and distribution strategies based on the composition and behavioral intentions of the crowd in different scenarios, thereby maintaining highly efficient guidance capabilities in complex and ever-changing tourism environments.

[0070] Example 7: In another optional embodiment, a smart traffic redirection method provided in this application includes the following steps: Step 1: When the congestion risk status indicates that the waiting time or crowd density in a certain interest hotspot exceeds the safe crowd threshold, diversion suggestions are simultaneously pushed to the digital signage screen at the entrance of this area and to tourists' personal mobile terminals; The congestion risk status is determined by combining real-time pedestrian density data and predicted waiting time data, which are statistically analyzed across the aforementioned behavioral assessment dimensions. The safe pedestrian flow threshold is a dynamic value preset based on the target area's spatial carrying capacity, fire evacuation requirements, and visitor comfort standards. For example, at a narrow mountain path entrance, the safe pedestrian flow threshold can be set at 2 people per square meter, while in an open plaza area, it can be set at 4 people per square meter. When the system detects that the real-time data for a popular area of ​​interest (such as a central observation deck or a popular ride) exceeds this threshold, a diversion mechanism is triggered.

[0071] Digital signage screens can refer to electronic display devices deployed at key decision-making points (such as intersections and entrance gates), while personal mobile terminals for tourists can refer to smartphones, tablets, or smart navigation devices rented by the scenic area. Synchronous push notifications mean that the system simultaneously sends consistent routing instructions to public displays in physical spaces and user terminals in private spaces within milliseconds, forming a three-dimensional information coverage network to ensure that tourists with different attention levels receive alert information.

[0072] For example, when the system detects that the waiting time at the glass walkway attraction exceeds 45 minutes and the entrance density reaches the warning line, a red warning box immediately pops up on the standing advertising screen at the entrance of the attraction, and a pop-up notification is sent to the mobile phones of all tourists within range who have opened the scenic area's mini-program. This dual-end synchronization mechanism effectively avoids information loss from a single channel and significantly improves the reach rate of warning information.

[0073] This step aims to intervene quickly in the early stages of congestion through multi-channel instant communication, leveraging visual impact and the strong reminder attributes of personal devices to break tourists' inherent path dependence and buy time for subsequent route replanning.

[0074] Step 2: The diversion suggestions are customized based on the preference tags of the typical group categories of tourists in the crowd analysis map. That is, it clearly recommends other sub-blocks with similar interest attributes to the current hotspot and with lower current passenger density, and attaches the walking distance and the expected saving of queuing time.

[0075] The crowd analysis map, a dynamic model built upon a multi-source tourist feature dataset, records the spatial distribution of typical group categories (such as families with children, photography enthusiasts, and history and culture explorers) and their corresponding preference labels within the current area. Customized descriptions can mean that the system no longer outputs generic, mechanical prompts like "Congestion ahead, please detour," but instead generates attractive alternatives based on tourist interest profiles.

[0076] Similar interest attributes can refer to the high degree of isomorphism between the recommended target sub-blocks and current congested hotspots in terms of content theme, landscape type, or experience. For example, for tourists identified as photography enthusiasts, if the currently popular sunrise viewing platform is crowded, the system will not recommend a nearby children's playground, but will instead recommend another cloud-top boardwalk with a wide view, suitable lighting, and less crowd. This recommendation logic based on interest matching leverages tourists' inherent psychological preferences, transforming forced diversion into proactive discovery, thereby significantly increasing the acceptance of suggestions.

[0077] Other sub-blocks can refer to areas within the scenic area that possess equal tourism value, excluding the current congestion points. Walking distance and expected queuing time savings are quantitative indicators calculated based on real-time traffic network data and pedestrian flow velocity models, used to visually demonstrate the actual benefits of choosing alternative solutions.

[0078] For example, if the system identifies a large number of families tagged as "family-friendly" visitors gathered in a sub-block, and the carousel in that area is severely congested, the generated diversion suggestion would be: "The carousel ahead has a 50-minute queue. We recommend the 'Dream Forest' interactive area 300 meters away (a similar family-friendly experience), which currently has no queue and is expected to save 45 minutes. There are rest stops for mothers and babies along the way." This description not only points out an alternative but also emphasizes the two core values ​​of a similar experience and time-saving, and thoughtfully provides the additional information of mother and baby rest stops, which is tailored to the characteristics of families.

[0079] Through this deeply customized diversion guidance, the system transforms impersonal traffic control into personalized service recommendations, effectively alleviating tourists' resistance to changes in their routes and achieving a balanced distribution of people in space.

[0080] This application constructs an efficient intelligent diversion and guidance mechanism through the synergistic effect of the aforementioned technical features. By monitoring congestion risk status in real time and comparing it with safe crowd flow thresholds, the system can keenly detect local overload signals. Based on this, it utilizes preference tags from crowd analysis maps to customize and reconstruct diversion suggestions, ensuring that recommended content accurately matches tourists' interests and needs, thus solving the problem of low acceptance rates caused by the lack of targeting in traditional diversion strategies. Furthermore, synchronous push notifications on digital signage screens and personal mobile terminals ensure full coverage and immediacy of information delivery; while the accompanying quantitative data on walking distance and estimated time savings provide tourists with clear decision-making basis, enhancing the credibility and persuasiveness of diversion suggestions. This combined strategy of precise matching + dual-terminal reach + quantified benefits not only effectively alleviates crowd pressure in hotspot areas but also improves tourists' visitor experience and satisfaction, achieving a dual improvement in scenic area operational efficiency and service quality.

[0081] Example 8: In one possible implementation, the generation process of a guidance identifier knowledge base provided in this application embodiment further includes a closed-loop iteration step based on feedback data after the implementation of identifier adjustment and diversion strategy set.

[0082] Step 1: Based on the tourist behavior feedback data collected after the implementation of the signage adjustment and diversion strategy set, extract diversion feedback indicators such as the acceptance rate of diversion suggestions, the change in regional flow speed, and the decrease in path repetition rate. The visitor behavior feedback data can be a set of real-time dynamic data recaptured by the system through location beacons, visual tracking devices, and mobile terminal interaction logs after the adjustment of the content and form of directional signage or the completion of intelligent diversion guidance strategies. This data comes from continuous monitoring within two time windows before and after the strategy implementation and is used to quantitatively evaluate the actual intervention effect of the strategy. The diversion feedback indicators include three core dimensions: diversion suggestion acceptance rate, which is defined as the proportion of visitors who received diversion push information and whose actual trajectory changed and entered the recommended alternative sub-block out of the total number of recipients. This indicator directly reflects the accuracy of the diversion strategy and the compliance of visitors; the change in regional flow speed, which is obtained by calculating the difference between the average movement speed of visitors in the target sub-block after the strategy implementation and the average movement speed before implementation. This is used to characterize the degree of congestion relief and the improvement of traffic efficiency; and the decrease in path repetition rate, which can refer to the decrease in the proportion of visitors turning back, lingering, or repeatedly passing through the same node in a specific area before and after the strategy implementation. This is used to measure the effectiveness of directional signage in eliminating visitor confusion. For example, after implementing a gamified guidance strategy in a family-friendly play area, the system statistics showed that out of 500 families who received the push notification, 320 families changed their original routes to the recommended low-density area, resulting in a 64% acceptance rate of the diversion suggestion. Simultaneously, the average movement speed of visitors in this area increased from 0.8 m / s to 1.2 m / s, a change of 0.4 m / s; and the path repetition rate decreased from 15% to 6%, a reduction of 9%. This multi-dimensional indicator extraction allows for a comprehensive and objective quantification of the immediate effects of the strategy, avoiding the bias of single-indicator evaluations. This step aims to transform abstract visitor behavior into calculable structured data, providing a solid data foundation for subsequent assessment of the strategy's effectiveness.

[0083] Step 2: Based on the diversion feedback indicators, determine the positive or negative effect of this strategy and generate a guidance identifier knowledge base; The positive or negative determination refers to the system's pre-set multi-dimensional threshold logic, which compares the extracted diversion feedback indicators with preset standard thresholds to determine whether the current strategy successfully solved the technical problem or failed to achieve the expected results. Specifically, when the diversion suggestion acceptance rate is higher than the first preset threshold (e.g., 50%), the change in regional flow speed is greater than zero, and the decrease in path repetition rate exceeds the second preset threshold (e.g., 5%), the strategy is considered to have a positive effect. Conversely, if any of the above key indicators fails to meet the standards or even deteriorates (e.g., a decrease in flow speed or an increase in repetition rate), it is considered negative. The guidance signage knowledge base is a dynamically updated structured database used to store the optimal strategy parameters, context, and corresponding weight coefficients in historical scenarios. Its purpose is to provide decision-making basis for similar future scenarios. This determination process works closely with the knowledge base generation, achieving the transformation from a single experiment to experience accumulation through automated logical judgment. For example, the system is set to mark a positive outcome when the acceptance rate is >60% and the flow rate increases by >0.2 m / s. If a flow diversion strategy targeting people taking photos and checking in meets this condition, the system immediately generates a positive record containing the strategy parameters and stores it in the knowledge base. If another adjustment to the large font labels targeting the elderly causes increased crowd gathering (decreased flow rate), a negative record is generated. Through this binary judgment mechanism, the system can quickly filter out effective strategies and eliminate invalid attempts, ensuring that the knowledge base stores only high-value information that has been verified in practice.

[0084] Step 3: If the result is positive, store the sub-block context and strategy parameters corresponding to the current strategy in the guidance identifier knowledge base, and increase the trigger weight of this strategy in similar scenarios. The sub-block context can refer to the set of environmental characteristics in which the strategy is executed, including but not limited to time period (e.g., afternoon on a holiday), weather conditions, passenger flow level, dominant group category (e.g., families with children, senior citizen groups), and the congestion risk level at that time. Strategy parameters can refer to the details of the specific adjustment instructions, such as the font size and color scheme of the displayed identifier, the text template of the pushed content, and the alternative area number for traffic diversion recommendations. Increasing the trigger weight can refer to increasing the probability coefficient of the strategy being prioritized when matching similar scenarios in the knowledge base retrieval algorithm. This mechanism ensures that successful experiences can be solidified and reused. When encountering the same or similar context again in the future, the system does not need to retry and can directly call the historically best strategy with high weight for a quick response. For example, if the system achieves a positive effect using a cartoon character guidance + ground light strip strategy in a scenario of 10 am on a weekend, a sunny day, and 70% families with children, the system will bind the scenario tag with the strategy parameters and store it in the knowledge base, increasing the trigger weight of the strategy from the default 1.0 to 1.5. Subsequently, once the sensors detect a similar combination of spatiotemporal and crowd characteristics, the system will prioritize deploying this high-weight strategy. Through deep binding of context and strategy parameters and dynamic adjustment of weights, the wayfinding signage system achieves adaptive learning and intelligent evolution, significantly improving guidance efficiency in complex and ever-changing tourism scenarios.

[0085] Step 4: If the result is negative, reduce the trigger weight of this strategy and adjust the threshold parameters of the corresponding dimension in the behavior evaluation dimension set.

[0086] Reducing the trigger weight can mean decreasing the likelihood of the failed strategy being selected in subsequent decisions, preventing the system from repeatedly executing invalid operations. Adjusting the threshold parameters of the corresponding dimensions in the behavior evaluation dimension set can mean revising the pre-judgment criteria used to generate strategies to optimize the accuracy of the next decision. The behavior evaluation dimension set includes key indicators such as interest aggregation tendency score, exploration dispersion tendency score, and path hesitation score. Its threshold parameters determine when and what type of guidance strategy is triggered. When a strategy is judged negatively, it indicates that the current threshold setting may be too sensitive or too sluggish, causing the strategy to mismatch with the actual situation. Therefore, these parameters need to be corrected in reverse. For example, if the system triggers strong diversion guidance because the path hesitation score slightly exceeds the threshold, resulting in visitor confusion (judged negatively), the system will not only reduce the weight of the diversion strategy but also automatically increase the trigger threshold for path hesitation, requiring more obvious wandering behavior in the future before initiating the same type of guidance. Similarly, if a push notification targeting a certain type of interest hotspot is not accepted, the system will adjust the calculation weight or threshold of the interest aggregation tendency score to make future interest identification more accurate. Through this negative feedback adjustment mechanism, the system can continuously correct its perception sensitivity and decision-making logic, avoid misjudgments caused by fixed thresholds, and ensure robustness and adaptability in the continuous iteration process.

[0087] This application constructs a complete closed-loop iterative mechanism through the aforementioned steps, enabling the wayfinding signage system to continuously learn and self-evolve. By using the acceptance rate of diversion suggestions, changes in regional flow speed, and the decrease in path repetition rate as core feedback indicators, the system can accurately quantify the actual effectiveness of each strategy execution. Based on this, positive judgments solidify successful contextual scenarios and strategy parameters into the wayfinding signage knowledge base and increase their weight, enabling rapid reuse of high-quality experiences. Simultaneously, negative judgments reduce the weight of ineffective strategies and dynamically adjust the threshold parameters of the behavior evaluation dimension set, enabling immediate correction of errors and model optimization. This cyclical process of execution-feedback-learning-optimization allows the system to evolve continuously based on real-time data, rather than relying on static rules or human experience. This ensures that the wayfinding signage deployment method can adapt to dynamic needs under different time periods, different population compositions, and different environmental conditions, ultimately achieving continuous improvement in the visitor experience and maximizing management efficiency.

[0088] Example 9: One embodiment of this application provides a process for collecting and processing tourist behavior feedback data, the method comprising: Step 1: Compare the spatiotemporal behavior data of each sub-block before and after the implementation of the signage adjustment and diversion strategy set, calculate the increase in average tourist movement speed and the decrease in path repetition rate, and use them as the change in regional flow speed and the decrease in path repetition rate, respectively. This step aims to objectively evaluate the actual improvement effect of directional signage adjustments on traffic efficiency by quantifying and comparing key motion indicators before and after the strategy implementation. Spatiotemporal behavioral data can refer to the sequence of tourist location coordinates, timestamps, and movement trajectory information continuously collected through location beacons, wireless probes, or visual tracking devices. The change in area flow speed is calculated by subtracting the average movement speed of all tourists in the target sub-block within a specific time period after strategy implementation from the average movement speed within the same time period before implementation; this value characterizes the improvement in traffic flow smoothness. The decrease in path repetition rate is calculated by statistically analyzing the number of trips or path overlaps of tourists in non-destination areas per unit time, and the reduction percentage before and after implementation; this value reflects the degree of relief from disorientation or hesitation. For example, in a main passageway sub-block of a scenic area, before the strategy was implemented, tourists frequently turned back due to unclear signage, with an average movement speed of 0.8 m / s and a path repetition rate of 15%. After implementing the dynamic diversion and signage optimization strategy, the average movement speed measured within the same time period increased to 1.2 m / s, and the path repetition rate decreased to 5%. Therefore, the change in regional flow speed was +0.4 m / s, and the path repetition rate decreased by 10%. Through this quantitative calculation based on objective motion data, subjective perception interference can be eliminated, and the specific contribution of signage adjustment to eliminating congestion nodes and reducing ineffective wandering can be accurately identified, providing a reliable efficiency dimension basis for subsequent strategy weight adjustments.

[0089] Step 2: Track whether the trajectory paths of tourists who received the diversion suggestion turned to the recommended alternative sub-block, and count the proportion of the number of people who turned to the total number of people who received the suggestion as the diversion suggestion acceptance rate.

[0090] This step focuses on evaluating the actual responsiveness and effectiveness of the intelligent diversion and guidance strategy on the tourist side. Diversion suggestions refer to guidance instructions generated by the system based on crowd analysis maps and behavioral assessment dimensions, pushed to digital signage screens or tourists' personal mobile terminals, containing information on alternative routes, estimated time savings, and points of interest. The acceptance rate of diversion suggestions is calculated by matching the tourist's device ID that received the push instruction with their subsequent real-time trajectory data to determine whether they entered the recommended alternative sub-block within a preset time window (e.g., within 10 minutes), and dividing the number of tourists meeting this condition by the total number of recipients. For example, if the system pushes a suggestion to 500 tourists located at the entrance of a popular photo spot to go to a less popular but similar viewpoint on the west side, and trajectory tracking reveals that 320 tourists deviated from their walking routes and entered the west viewpoint area, then the acceptance rate of this diversion suggestion is 64%. This metric directly reflects the attractiveness and credibility of the diversion strategy; a high acceptance rate means that the recommended content is highly matched with the tourist's interest profile and is clearly expressed. By combining the acceptance rate with the aforementioned efficiency indicators, the system can not only determine whether the flow of people has been diverted, but also verify whether the diversion methods meet the wishes of tourists. This avoids the problem of decreased experience that may be caused by forced diversion, and ensures that the direction of closed-loop iterative optimization is both in line with management efficiency and takes into account tourist satisfaction.

[0091] This application constructs a multi-dimensional strategy effectiveness evaluation system by leveraging the synergistic effects of changes in regional flow speed, the decrease in path repetition rate, and the acceptance rate of diversion suggestions in the aforementioned steps. Changes in regional flow speed and the decrease in path repetition rate reflect improvements in traffic efficiency and reductions in congestion at a macroscopic physical level, while the acceptance rate of diversion suggestions verifies the accuracy of the guidance strategy and user engagement at a microscopic individual behavior level. The combination of these three factors allows the system to distinguish between the effects of passive congestion and active guidance: if speed increases but the acceptance rate is low, it may mean that tourists are forced to change routes due to congestion rather than accepting the guidance; if the acceptance rate is high but speed does not significantly increase, it suggests that the carrying capacity of alternative areas or path planning still needs optimization. With this comprehensive data feedback mechanism, the system can accurately determine whether each sign adjustment and diversion strategy is positive or negative, thereby dynamically updating the strategy trigger weights and threshold parameters in the directional sign knowledge base. This enables continuous adaptive evolution of the directional sign deployment scheme, effectively solving the technical problem in existing technologies where the lack of quantitative feedback prevents the system from self-optimizing.

[0092] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0093] Figure 3 This application provides a schematic diagram of the structure of an iterative deployment system for wayfinding signs based on the spatiotemporal distribution of pedestrian flow, as shown in one embodiment. Figure 3 As shown, the data transmission device 300 in this embodiment includes: a crowd profiling and behavior module 301, an adjustment and diversion strategy module 302, and an identification knowledge base module 303.

[0094] The crowd profiling and behavior module 301 is used to acquire a multi-source feature dataset of tourists, and based on the multi-source feature dataset of tourists, dynamically construct a crowd analysis map and behavior evaluation dimensions to generate a set of crowd analysis maps and behavior evaluation dimensions. The adjustment and diversion strategy module 302 is used to make collaborative decisions on adjusting the form of the guiding signage content and intelligent diversion guidance based on the population analysis map and the behavior assessment dimension set, and to generate a set of signage adjustment and diversion strategies. The identification knowledge base module 303 is used to collect tourist behavior feedback data based on the identification adjustment and diversion strategy set, and to perform closed-loop iterative optimization on the generation process of the crowd analysis map and the behavior evaluation dimension set to generate an iterated guidance identification knowledge base.

[0095] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A method for iteratively deploying directional signs based on the spatiotemporal distribution of pedestrian flow, characterized in that, include: Obtain a multi-source feature dataset of tourists, and based on the multi-source feature dataset of tourists, dynamically construct a crowd analysis map and behavioral assessment dimensions to generate a set of crowd analysis maps and behavioral assessment dimensions; Based on the aforementioned population analysis map and the aforementioned behavior assessment dimension set, collaborative decision-making is made on adjusting the content form of directional signage and intelligent traffic diversion guidance, generating a set of signage adjustment and traffic diversion strategies; Based on the aforementioned signage adjustment and diversion strategy set, tourist behavior feedback data is collected, and the generation process of the crowd analysis map and the behavior evaluation dimension set is optimized in a closed loop to generate an iteratively updated directional signage knowledge base.

2. The method according to claim 1, characterized in that, The generation process of the population analysis map and the behavioral assessment dimension set includes: The multi-source feature dataset of tourists includes: spatiotemporal behavior data, visual behavior feature data, and personal interest profile data; Based on the aforementioned multi-source feature dataset of tourists, we can identify individual tourist behavior tendencies and group aggregation patterns, and construct a crowd analysis map. Based on the tourist multi-source feature dataset and the population analysis map, an interest preference and behavioral tendency fusion analysis is performed to construct the behavioral evaluation dimension set.

3. The method according to claim 2, characterized in that, The process of constructing the population analysis map includes: By analyzing the dwell time and movement speed in the spatiotemporal behavior data, subsets of tourists in static and fast-moving states are identified. Based on the visual behavior feature data, the group aggregation pattern of tourists and the height characteristics of children are extracted to determine the tourist combination type within each tourist subset. The tourist combination type includes group tours, families with children, and solo tours. The tourist combination types are matched with the preference tags in the personal interest profile data to form several typical group categories. The proportion and distribution density of each typical group category in different sub-blocks of the target area are counted to form the population analysis map.

4. The method according to claim 3, characterized in that, The process of constructing the behavioral assessment dimension set includes: Based on the spatiotemporal behavioral data, the degree of deviation of the tourist's trajectory path from the main road is extracted, and combined with the tourist's facial orientation and handheld device posture in the visual behavioral feature data, each tourist is assigned an exploration dispersion tendency score. By analyzing the matching degree between the preference tags in the personal interest profile data and the type of attractions in the current location, and combining the length of stay, each tourist individual is assigned an interest clustering tendency score. Based on the spatiotemporal behavioral data, the number of times tourists turned back and the duration of their hesitation at intersections or signs were extracted, and each individual tourist was assigned a score for the degree of path hesitation. The scores of all tourists are statistically summarized according to the group categories in the aforementioned population analysis map to generate the behavioral assessment dimension set.

5. The method according to claim 4, characterized in that, The process of generating the identifier adjustment and traffic splitting strategy set includes: Based on the population analysis map, identify the typical group categories and their corresponding interest hotspot distribution in each sub-block within the target area; Based on the aforementioned behavioral assessment dimension set, the congestion risk status and tourist behavior inconsistencies in each sub-block are detected. By combining the distribution of interest hotspots, the congestion risk status, and the contradictions in tourist behavior, a collaborative decision is made on adjusting the content and form of directional signs and providing intelligent diversion guidance, generating a set of sign adjustment and diversion strategies that includes sign update instructions and diversion push information.

6. The method according to claim 5, characterized in that, The adjustment of the content format of the directional signage includes: When the population analysis map shows that the typical group category of a certain sub-block has changed, and the corresponding behavioral tendency index of the behavior evaluation dimension set exceeds the set threshold, the guidance focus of the current area identifier will be switched from the existing content to interest hotspots or service facility information that match the typical group category after the switch, and the visual presentation of the identifier will be adjusted simultaneously to adapt to the group characteristics. When the interest clustering tendency score indicates that tourists in this area have a strong interest in a certain hotspot, dynamic guidance information pointing to such hotspots will be added to the digital signage screens and tourists' personal mobile terminals.

7. The method according to claim 5, characterized in that, The intelligent traffic redirection guidance includes: When the congestion risk status indicates that the waiting time or crowd density of a certain interest hotspot area exceeds the safe crowd threshold, diversion suggestions are simultaneously pushed to the digital signage screen at the entrance of this area and to tourists' personal mobile terminals. The diversion suggestions are customized based on the preference tags of the typical group categories to which tourists belong in the crowd analysis map. Specifically, they explicitly recommend other sub-blocks with similar interest attributes to the current hotspot and with lower current passenger density, along with walking distance and expected savings in queuing time.

8. The method according to claim 5, characterized in that, The generation process of the navigation signage knowledge base includes: Based on visitor behavior feedback data collected after the implementation of the aforementioned signage adjustment and diversion strategy set, diversion feedback indicators are extracted, including diversion suggestion acceptance rate, change in regional flow speed, and decrease in path repetition rate. Based on these diversion feedback indicators, the effectiveness of the strategy is determined as positive or negative, and the aforementioned directional signage knowledge base is generated. If the result is positive, the sub-block context and strategy parameters corresponding to the current strategy are stored in the guidance identifier knowledge base, and the trigger weight of this strategy in similar scenarios is increased. If the result is negative, the trigger weight of this strategy is reduced, and the threshold parameters of the corresponding dimension in the behavior evaluation dimension set are adjusted.

9. The method according to claim 8, characterized in that, The collection and processing of tourist behavior feedback data includes: By comparing the spatiotemporal behavior data of each sub-block before and after the implementation of the identification adjustment and diversion strategy set, the increase in the average movement speed of tourists and the decrease in the path repetition rate are calculated, which are respectively used as the change in the regional flow speed and the decrease in the path repetition rate. Track whether the trajectory path of tourists who receive the diversion suggestion turns to the recommended alternative sub-block, and count the proportion of the number of people who turn to the total number of people who receive the suggestion as the acceptance rate of the diversion suggestion.

10. A wayfinding signage iterative deployment system based on the spatiotemporal distribution of pedestrian flow, characterized in that, The method applied to any one of claims 1-9 includes: The crowd profiling and behavior module is used to acquire multi-source feature datasets of tourists, and based on the multi-source feature datasets of tourists, dynamically construct crowd analysis maps and behavior evaluation dimensions to generate crowd analysis maps and behavior evaluation dimension sets. The adjustment and diversion strategy module is used to make collaborative decisions on adjusting the form of the guiding signage content and intelligent diversion guidance based on the population analysis map and the behavior assessment dimension set, and to generate a set of signage adjustment and diversion strategies. The identification knowledge base module is used to collect tourist behavior feedback data based on the identification adjustment and diversion strategy set, and to perform closed-loop iterative optimization on the generation process of the crowd analysis map and the behavior evaluation dimension set to generate the iterated guidance identification knowledge base.

Images