A ski field outdoor advertisement monitoring data analysis method and system

Abstract

The present application relates to the technical field of outdoor advertising effect evaluation, and particularly relates to a ski field outdoor advertising monitoring data analysis method and system, which comprises the following steps: collecting multi-source monitoring data of a ski field outdoor advertising site, wherein the multi-source monitoring data comprises advertising site video stream, environmental sensing data and skier behavior trajectory; performing dynamic environmental interference correction on the advertising site video stream based on the environmental sensing data to generate an advertising exposure image sequence; constructing a dynamic space-time association between a skier and the advertising site according to the advertising exposure image sequence and the skier behavior trajectory, and extracting an advertising attention feature set; and inputting the advertising attention feature set into an advertising effect evaluation model to generate an advertising effect quantitative index. Through the present application, the advertising attention quantitative inaccuracy in the high-speed motion scene of a ski field is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of outdoor advertising effectiveness evaluation technology, and in particular to a method and system for analyzing outdoor advertising monitoring data at ski resorts. Background Technology

[0002] Outdoor advertising monitoring data analysis technology aims to evaluate the effectiveness of advertising exposure by collecting video, sensor and user behavior data of advertising locations. In the ski resort scenario, it needs to deal with the unique challenges of skiers' high-speed movement and complex environmental interference. Its core is to transform multi-source data into quantitative indicators of advertising effectiveness to optimize the placement strategy.

[0003] Current mainstream solutions rely on wide-angle camera video analysis or single-positioning technology, with some incorporating environmental sensors. However, these technologies have fundamental flaws: advertising attention measurement is severely inaccurate in high-speed skiing scenarios. Specifically, wide-angle camera solutions produce motion blur when skiers are moving at high speeds, and traditional hardware filters or linear corrections cannot recover effective details after environmental changes cause image distortion, resulting in computational distortion. Single-positioning technology suffers from high signal loss rates in complex terrain, and its inability to continuously track behavior leads to the failure to extract features such as advertising dwell frequency and trajectory deviation. These flaws directly point to the core problem that existing technologies cannot solve—the distortion of advertising effectiveness evaluation caused by the coupling of high-speed movement and environmental interference.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a method and system for analyzing outdoor advertising monitoring data at ski resorts, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for analyzing outdoor advertising monitoring data at ski resorts, the method comprising:

[0008] Collect multi-source monitoring data of outdoor advertising spaces at ski resorts. The multi-source monitoring data includes video streams of advertising spaces, environmental sensor data, and skier behavior trajectories.

[0009] Based on the environmental sensing data, dynamic environmental interference correction is performed on the video stream of the ad slot to generate an ad exposure image sequence;

[0010] Based on the image sequence of the advertisement exposure and the skier's behavioral trajectory, a dynamic spatiotemporal relationship between the skier and the advertisement placement is constructed, and the advertisement attention feature set is extracted.

[0011] The advertising attention feature set is input into the advertising effectiveness evaluation model to generate quantitative indicators of advertising effectiveness.

[0012] Further, a sequence of images for ad exposure is generated, including:

[0013] Acquire the environmental sensing data, including light intensity, snow / fog concentration, and wind speed data;

[0014] A brightness compensation matrix is ​​constructed based on the light intensity to correct the video stream of the advertising space;

[0015] The dynamic blur of the video stream in the ad slot is compensated based on the wind speed data;

[0016] Based on the snow fog concentration, a defogging convolution kernel is generated. The video stream of the ad slot is corrected according to the defogging convolution kernel, and the ad exposure image sequence is generated.

[0017] Furthermore, the generation of the skier's behavioral trajectory includes:

[0018] Obtain the skier's starting point on the slope and generate initial position coordinates;

[0019] Based on the real-time changes of the initial position coordinates, the color of the skier's clothing and the outline of the ski equipment in the corresponding moving area are obtained and merged to generate dynamic visual features;

[0020] The skier's behavioral trajectory is generated by combining the initial position coordinates with the dynamic visual features.

[0021] Furthermore, dynamic visual features are generated by merging, including:

[0022] Perform spatial transformation on the color features of the skier's clothing, and extract the hue and saturation components to generate a first feature vector;

[0023] Key point detection is performed on the outline features of the ski equipment to obtain the outline curvature histogram and generate a second feature vector;

[0024] The first feature vector and the second feature vector of each skier are merged to output the dynamic visual features;

[0025] In this process, an attention weight is added to the merging of the first feature vector and the second feature vector. The attention weight is dynamically adjusted based on the real-time distance between the skier and the advertising space.

[0026] Furthermore, the dynamic spatiotemporal correlation includes:

[0027] Using the visible area of ​​the advertising exposure image sequence as a reference plane, the skier's behavioral trajectory is projected onto the reference plane;

[0028] The gaze duration timer is activated when the skier's trajectory enters the reference plane.

[0029] When the angle between the skier's behavior trajectory and the normal of the reference plane and the gaze duration are less than a threshold, the data is recorded as valid gaze data.

[0030] Furthermore, the advertising attention feature set is generated based on the aggregation of the effective gaze data, wherein attention weights are added as accompanying weights to the effective gaze data during the aggregation process.

[0031] Furthermore, generate quantitative metrics for advertising effectiveness, including:

[0032] After frequency domain transformation, the effective gaze data in the ad attention feature set is combined with attention weight and historical behavior data to predict the conversion probability of the behavior guided by the ad, and the quantitative index of the ad effect is output.

[0033] Furthermore, it also includes collecting real-time weather warning data, and when the weather warning data reaches a threshold, adding a confidence correction coefficient to the quantitative indicator of advertising effectiveness.

[0034] A ski resort outdoor advertising monitoring and data analysis system, the system comprising:

[0035] The data acquisition module collects multi-source monitoring data of outdoor advertising spaces in the ski resort. The multi-source monitoring data includes video streams from the advertising spaces, environmental sensor data, and skier behavior trajectories.

[0036] The delayed image module performs dynamic environmental interference correction on the video stream of the ad slot based on environmental sensor data, and generates an ad exposure image sequence;

[0037] The feature generation module constructs a dynamic spatiotemporal relationship between skiers and ad placements based on the ad exposure image sequence and skier behavior trajectory, and extracts ad attention feature sets.

[0038] The effectiveness quantification module inputs the ad attention feature set into the ad effectiveness evaluation model to generate quantitative indicators of ad effectiveness.

[0039] Furthermore, the feature generation module includes:

[0040] The baseline setting unit uses the visible area of ​​the advertisement exposure image sequence as the baseline plane and projects the skier's behavior trajectory onto the baseline plane;

[0041] The gaze timing unit starts timing the gaze duration when the skier's trajectory enters the reference plane.

[0042] The data recording unit records valid gaze data when the angle between the skier's trajectory and the normal to the reference plane and the gaze duration are less than the threshold.

[0043] The technical solution of this invention can achieve the following technical effects:

[0044] By constructing multimodal data fusion in the high-speed motion scenario of a ski resort, accurate quantitative evaluation of advertising attention was achieved, effectively overcoming the problems of inaccurate quantitative results caused by dynamic blur, environmental interference, and trajectory breakage.

[0045] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating a method for analyzing outdoor advertising monitoring data at ski resorts.

[0048] Figure 2 A flowchart illustrating the process of generating an image sequence for ad exposure;

[0049] Figure 3 A flowchart illustrating the process of obtaining a skier's behavioral trajectory;

[0050] Figure 4 This is a conceptual diagram illustrating dynamic spatiotemporal correlation. Detailed Implementation

[0051] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Example 1

[0053] like Figure 1As shown, this application provides a method for analyzing outdoor advertising monitoring data at ski resorts, the method including:

[0054] S10: Collect multi-source monitoring data of outdoor advertising spaces in ski resorts. The multi-source monitoring data includes video streams of advertising spaces, environmental sensor data, and skier behavior trajectories.

[0055] S20: Based on environmental sensor data, perform dynamic environmental interference correction on the video stream of the ad slot to generate an ad exposure image sequence;

[0056] S30: Based on the image sequence of ad exposure and skier behavior trajectory data, construct the dynamic spatiotemporal relationship between skiers and ad placements, and extract the ad attention feature set;

[0057] S40: Input the ad attention feature set into the ad performance evaluation model to generate quantitative indicators of ad performance.

[0058] Specifically, firstly, a multi-source data acquisition system is set up and deployed in multiple key locations within the ski resort. This includes high-definition network cameras facing the advertising spaces, environmental monitoring nodes covering the entire area, and behavior acquisition terminals installed on skiers. The video cameras continuously capture video streams from the advertising spaces to ensure all-weather image data is available. The environmental sensor nodes collect environmental data in real time, providing a basis for subsequent video correction. The skier behavior acquisition terminals record their movement trajectory data within the area, including location, speed, dwell time, and direction of travel. In the data processing stage, the data is first processed based on the environmental sensor data... The video images undergo dynamic environmental interference correction. Specifically, when the sensor detects environmental factors such as strong light reflection, snow and fog obstruction, backlighting, or low nighttime illumination, image processing methods such as image enhancement, region occlusion analysis, and background modeling are used to correct the video images, removing or reducing image distortion caused by these interferences, thereby obtaining a higher-quality advertising exposure image sequence. Next, a dynamic spatiotemporal correlation is established between the advertising exposure image sequence and the skier's trajectory data. To achieve this correlation, the skier's trajectory is time-synchronized and spatially mapped, and its location at a specific moment is cross-referenced with the visible area of ​​the advertising space. If a skier's movement direction is towards the advertising space at a certain moment, and the distance is within visual perception range, the skier is considered to be in a potential attention state. If the skier's head is further detected to be facing the advertising direction in the image, this state is upgraded to an effective attention state. In this way, every effective interaction between the skier and the advertising space can be marked and statistically analyzed. Subsequently, based on the constructed spatiotemporal correlation, an advertising attention feature set is extracted. This feature set includes, but is not limited to, multiple dimensions such as: number of people paying attention per unit time, average attention duration, attention angle distribution, dwell time distribution, and advertising visibility. To improve the effectiveness of the feature set... Effectiveness can also be assessed by classifying and analyzing data based on skier behavior types to select behavioral samples that are meaningful for advertising effectiveness. Finally, the extracted advertising attention feature set is input into a pre-trained advertising effectiveness evaluation model. This model can be built using machine learning methods, such as random forests, support vector machines, or neural networks, and trained under supervision using historical advertising data to output comprehensive evaluation indicators, such as advertising exposure value score, attention level, or advertising effectiveness score. Preferably, the model can be dynamically updated and optimized according to actual operational needs to adapt to the effectiveness evaluation of different advertising strategies.

[0059] Through the technical solution of this invention, multimodal data fusion is constructed in the high-speed sports scene of ski resort, which realizes accurate quantitative evaluation of advertising attention and effectively overcomes the problems of inaccurate quantitative effect caused by dynamic blur, environmental interference and trajectory breakage.

[0060] Furthermore, such as Figure 2As shown, the image sequence for generating ad exposures includes:

[0061] Acquire environmental sensor data, including light intensity, snow and fog concentration, and wind speed data;

[0062] A brightness compensation matrix is ​​constructed based on the light intensity to correct the video stream of the ad space;

[0063] Dynamic blurring of the video stream in the ad placement is compensated based on wind speed data;

[0064] A dehazing convolution kernel is generated based on the snow fog concentration. The video stream of the ad slot is corrected according to the dehazing convolution kernel, and an ad exposure image sequence is generated.

[0065] As a preferred embodiment of the above, firstly, light intensity data for the current time period is collected from an environmental sensor at regular intervals, and a brightness compensation matrix is ​​constructed based on this data to perform brightness correction processing on the collected video stream of the advertising space. The construction process of the brightness compensation matrix comprehensively considers the overall brightness level of the video image, the distribution characteristics of local overexposed or underexposed areas, and the periodic pattern of light changes over time, so as to achieve dynamic balance correction of the image brightness. Especially in the scene of strong light reflection or shadow occlusion common in ski resorts, this brightness compensation mechanism can effectively improve the visual clarity of the advertising area and ensure the stability of the image quality on which subsequent analysis depends. Secondly, wind speed, as an important external factor for the generation of dynamic blur, is considered in this embodiment. In this method, dynamic blur compensation is applied to the video stream of the advertising space. By acquiring real-time wind speed data from the video capture area, combined with camera capture parameters and motion blur characteristics of target edges in the video image, it identifies slight shaking, image trailing, or target ghosting caused by wind speed. To compensate, an adaptive deconvolution method combined with an image sharpening strategy is used. A blur estimation model restores the image edges and texture structure, enhancing the boundary contrast and detail restoration of the advertising area. Preferably, multiple wind speed threshold ranges can be set, with different ranges corresponding to different degrees of blur models, thereby improving the targeting and accuracy of image restoration. Furthermore, considering the significant impact of snow and fog concentration on image quality... To address the impact of snow and fog, this implementation introduces a dehazing convolutional kernel based on snow and fog concentration for dehazing video images. Visibility sensors deployed near the advertising location or the camera's built-in image analysis function are used to calculate the snow and fog concentration in the current frame in real time. When dense fog or significant snowfall obscures the image, a dehazing convolutional kernel matching the current fog level is automatically loaded. This kernel, designed based on image dehazing principles, specifically suppresses brightness diffusion and detail loss caused by snow and fog during image processing. To enhance the dehazing effect, an image contrast enhancement mechanism and color reconstruction algorithm are also introduced. Based on spatial domain filtering, the clarity and color saturation of the advertising image are further restored. After the above processing steps, the corrected image frames are combined in chronological order to generate a stable, clear image sequence of advertisement exposure with minimal environmental interference. This image sequence serves as an important foundational dataset for subsequent skier behavior analysis and advertisement attention extraction. It not only effectively solves the adverse effects of complex weather conditions on video quality but also provides high-quality image-level support for evaluating the effectiveness of outdoor advertising at ski resorts. For example, in a period of heavy snowfall, by accurately identifying snow fog concentration and constructing defogging convolution kernels in real time, the image clarity was significantly improved without changing the hardware. This enabled the subsequent advertisement recognition module to accurately determine the advertisement location and content, thereby ensuring the continuity and effectiveness of data analysis.

[0066] Furthermore, such as Figure 3As shown, the generation of skier behavior trajectory data includes:

[0067] Obtain the skier's starting point on the slope and generate initial position coordinates;

[0068] Based on the real-time changes in the initial position coordinates, the color of the skier's clothing and the outline of the ski equipment in the corresponding moving area are obtained and merged to generate dynamic visual features.

[0069] Skier behavior trajectory data is generated by combining initial position coordinates with dynamic visual features.

[0070] As a preferred embodiment of the above, firstly, the initial position coordinates of the skier are obtained by a positioning sensor placed near the starting point of the ski slope. When the skier is about to start, the sensor captures the geographical coordinates of the skier at the start and records them as the initial coordinates. These initial position coordinates will serve as the basis for subsequent trajectory calculation and dynamic behavior tracking. After obtaining the initial position coordinates, the skier's real-time movement status and surrounding environment are further monitored, and the skier's movement area is analyzed based on the changes in the initial coordinates. Specifically, combined with real-time camera image acquisition data, the skier's clothing color and ski equipment outline are identified through target detection and image segmentation algorithms. The skier's clothing color and ski equipment shape outline are important visual features for identifying the skier's identity and dynamic behavior. Using image processing techniques, such as color space conversion and outline extraction, the skier's clothing color and ski equipment within a specific time period are identified. These visual features, together with the initial position coordinates, constitute the skier's dynamic visual features during the movement process. By merging the skier's initial position coordinates with their dynamic visual features, the skier's behavioral trajectory can be accurately tracked. This process involves the collaborative work of multiple modules: First, based on the real-time changes in the initial position, the skier's... The system updates the direction, speed, and position of the skier on the slope. Then, it identifies and locates the skier in each frame based on visual features, ensuring that the trajectory data at each moment matches the corresponding visual information. The introduction of dynamic visual features enables the differentiation of different skiers, even if they are in the same or similar positions, and allows for independent behavioral analysis based on clothing color and ski equipment outline. For example, in a typical ski resort application scenario, suppose skier A starts skiing from the starting point of the slope. The coordinates of the starting point are recorded using the initial position coordinates. As skier A moves, the camera in front captures the skier's trajectory in real time, and combined with the visual features of the skier's red jacket and yellow skis, the skier's trajectory is accurately marked. Even in complex skiing environments, the system can eliminate background noise or interference from other skiers based on the skier's clothing color and ski equipment outline features, thereby generating accurate behavioral trajectory data for skier A. This behavioral trajectory data includes not only the skier's position information but also dynamic parameters such as speed, acceleration, and movement path at different time points. This data can be used for subsequent advertising attention analysis and effect evaluation, providing important support for optimizing advertising strategies.

[0071] Furthermore, the merging and generation of dynamic visual features includes:

[0072] Perform spatial transformation on the color features of skier's clothing, and extract hue and saturation components to generate the first feature vector;

[0073] Key point detection is performed on the contour features of the ski equipment to obtain the contour curvature histogram and generate a second feature vector;

[0074] Merge the first and second feature vectors of each skier to output dynamic visual features;

[0075] In this process, an attention weight is added to the merging of the first and second feature vectors. The attention weight is dynamically adjusted based on the real-time distance between the skier and the advertisement.

[0076] As a preferred embodiment of the above, firstly, the color features of the skier's clothing are spatially transformed using image processing technology, specifically employing a color space transformation algorithm. During this process, the hue and saturation components of the clothing are extracted. Hue represents the type of color, while saturation describes the purity or intensity of the color. By extracting these two components, the color features of the skier's clothing can be obtained. These features are crucial for tracking skiers in complex background environments. The extracted hue and saturation components are normalized to generate a first feature vector, representing the color features of the skier's clothing. Secondly, keypoint detection is performed on the outline features of the skier's ski equipment. To accurately capture the appearance features of the skier's ski equipment, image segmentation and outline detection algorithms are used to extract the edge outlines of the ski equipment. Subsequently, a curvature histogram of the ski equipment is calculated based on the outline features. This is a statistical feature that describes changes in the shape of the outline. The curvature histogram reflects the degree and direction of the curvature of the outline, thus helping the system identify differences between different ski equipment shapes. Through these operations, a second feature vector is generated, representing the morphological features of the skier's ski equipment. Next, the extracted... The first and second feature vectors are merged to generate the skier's dynamic visual features. To ensure the priority of visual features under different time and space conditions, an attention weighting mechanism is introduced to dynamically adjust the merging weights of the first and second feature vectors. These attention weights are determined by the real-time distance between the skier and the advertising space. Specifically, when the skier is close to the advertising space, the weight of clothing color features is increased to highlight the skier's visual recognizability; while when the skier is far from the advertising space, the weight of ski equipment outline features is increased to improve the recognition accuracy of the ski equipment. This process ensures that it can flexibly respond to changes in the relative position of the skier and the advertising space in different scenarios and optimize the performance of visual features. Finally, the merged dynamic visual features are output and used as input data for subsequent modules such as advertising exposure analysis and skier behavior recognition. This feature synthesis and weighting strategy effectively improves the recognition accuracy of skiers, especially in complex ski resort environments, and can stably track the skier's movement trajectory and behavioral characteristics, maintaining efficient and accurate data analysis at both long and short distances.

[0077] Furthermore, such as Figure 4 As shown, dynamic spatiotemporal correlation includes:

[0078] Using the visible area of ​​the advertising exposure image sequence as a reference plane, the skier's behavioral trajectory is projected onto the reference plane;

[0079] The gaze duration timer starts when the skier's trajectory enters the reference plane;

[0080] When the angle between the skier's trajectory and the reference plane normal, and the gaze duration are both less than the threshold, the data is recorded as valid gaze data.

[0081] As a preferred embodiment of the above, firstly, the visible area of ​​the advertisement is extracted from the advertisement exposure image sequence and defined as a planar reference plane. This reference plane is typically a two-dimensional projection area of ​​the viewpoint of the camera device facing the advertisement position. Its spatial position and orientation are determined by the camera parameters and the advertisement installation angle. By utilizing the ski resort's 3D environment modeling system and the spatial annotation information of the advertisement positions, this reference plane can be accurately mapped to the entire ski resort coordinate system. Next, based on the skier's behavioral trajectory data, the skier's movement path in three-dimensional space is projected onto the aforementioned reference plane. This projection process considers the skier's real-time position information and viewing direction, thereby ensuring the accuracy of the trajectory mapping on the reference plane. Secondly, when the skier's trajectory first enters the projection area of ​​the reference plane, the gaze duration timing module is automatically activated. Specifically, when the skier's projected trajectory intersects or overlaps with the advertisement reference plane, time recording begins, and the position and orientation changes of the skier relative to the advertisement area are continuously tracked to determine whether the skier is in a state of attention. The core indicators of the status include two elements: gaze duration and observation angle. The gaze duration is obtained through timestamp accumulation, recording the total time the skier maintains a gaze within the projection area. The observation angle is calculated based on the angle between the skier's line of sight and the normal vector of the advertising reference plane. The smaller the angle, the closer the skier's line of sight is to the advertising direction, and the higher the probability of visual contact with the advertisement. To determine whether the skier's gaze is valid, two thresholds are set: a minimum gaze duration threshold to exclude misjudgments caused by unintentional brief glances; and a maximum angle threshold to ensure that the line of sight is sufficiently close to the advertisement position. When the skier's behavior trajectory enters the advertising reference plane, if the gaze duration exceeds the set threshold and the angle between the skier and the advertising plane normal is lower than the angle threshold, the data segment is recorded as valid gaze data. This data is not only used for subsequent advertising exposure analysis, but can also be further combined with skier identity data, ski route, dwell time, and other factors to form a complete advertising influence evaluation model.In a preferred embodiment of the present invention, to achieve dynamic spatiotemporal correlation between advertising spaces and skier behavior trajectories, a spatially visible area is first established as a reference plane based on the field of view of each advertising space in the advertising exposure image sequence. This reference plane can be accurately determined using multi-camera calibration technology combined with known advertising installation location information and mapped onto the ski resort's unified three-dimensional coordinate system. The advertising reference plane not only includes two-dimensional position but also carries normal vector direction information for subsequent determination of the skier's observation angle. During the process of identifying the skier's real-time three-dimensional trajectory, the skier's spatial position is obtained using depth cameras, millimeter-wave radar, or multi-camera image fusion, and their current line of sight is estimated by combining their movement orientation. During the skier's movement, it is continuously determined whether their trajectory enters the aforementioned advertising reference plane. The visible area represents the baseline area. When the trajectory projection first enters this baseline, the gaze duration recording mechanism is immediately activated, and the relative position changes between the skier and the advertising space, as well as the angle between the skier's gaze direction and the advertising normal, are tracked. This angle can be estimated by analyzing the angle between the skier's facial direction and the advertising surface, typically achieved through a visual posture estimation model or head orientation detection technology. If the skier's observation angle remains within the set attention range (i.e., the angle with the advertising normal is less than a set threshold), and the skier's gaze duration in this state exceeds another threshold, then this behavioral data is marked as a valid gaze behavior, and a valid gaze data record is generated. The threshold settings here can be flexibly configured according to the actual needs of the ski resort's advertising campaign to balance recognition accuracy and applicability.

[0082] Furthermore, the advertising attention feature set is generated based on aggregated effective gaze data, with attention weights added as accompanying weights to the effective gaze data during the aggregation process.

[0083] As a preferred embodiment of the above, effective gaze data of skiers is first obtained from the aforementioned steps. Effective gaze data includes information such as the duration of the skier's gaze in front of the advertising space, the gaze angle, and the distance between the skier and the advertising space. After initial acquisition, this data undergoes data cleaning and preprocessing to remove noise and outliers to ensure data accuracy. In this embodiment, the attention weight is mainly adjusted based on the real-time distance between the skier and the advertising space. Specifically, when the skier is closer to the advertising space, the attention weight of the gaze is higher, and vice versa. The adjustment of the weight can consider other factors, such as the skier's movement state and the potential impact of environmental factors on attention. When generating the advertising attention feature set, an initial weight is first assigned to each piece of effective gaze data. At this time, the weight is mainly set based on the gaze duration and gaze angle, for example, as... If a skier's gaze duration is longer and the angle is more perpendicular to the ad placement, the weight is higher. Subsequently, this weight is dynamically adjusted based on the distance between the skier and the ad placement. For example, when the skier is far from the ad placement, the weight of that gaze data is reduced, and vice versa. The aggregation process combines the attention weight of each valid gaze data with other features, and then performs a weighted average or other suitable aggregation method. In this way, the generated ad attention feature set can accurately reflect the skier's attention to the ad placement and provide a basis for subsequent ad performance evaluation. Finally, the weighted aggregated ad attention feature set will be used as output and enter the subsequent ad performance analysis and optimization process. This feature set can help advertisers accurately evaluate the effectiveness of ad placement, thereby making reasonable adjustments to ad content, ad placement, and placement strategies.

[0084] Furthermore, generating quantitative metrics for advertising effectiveness includes:

[0085] After frequency domain transformation, effective gaze data from the ad attention feature set is combined with attention weight and historical behavior data to predict the conversion probability of ad-guided behavior and output quantitative indicators of ad effectiveness.

[0086] As a preferred embodiment of the above, firstly, the advertising attention feature set generated in the aforementioned steps is obtained, including effective gaze data and corresponding attention weights. These effective gaze data represent the skier's level of attention to the advertisement, while the attention weights reflect the skier's intensity of interest in the advertisement. In this step, frequency domain transformation (such as Fourier transform) is used to convert these time-domain data into frequency-domain data. The purpose of frequency domain transformation is to analyze the skier's attention patterns to the advertisement from different frequency levels, especially to identify which frequencies of gaze behavior are more strongly correlated with the advertising effect through periodic patterns. The gaze duration and gaze angle data are decomposed, and their frequency domain features are extracted respectively. Fourier transform is used to obtain the low-frequency components representing continuous attention and the high-frequency components representing intermittent or short-term attention in the frequency domain. Frequency domain transformation helps identify different types of attentional behavior. After frequency domain transformation, the next step is to combine skiers' attention weights and historical behavioral data to predict the conversion probability of ad-guided behavior. Historical behavioral data includes skiers' past ad viewing records, ski routes, and interactive behaviors. Based on this data, machine learning algorithms can be used to predict ad effectiveness. Through the combination of the above frequency domain transformation and behavioral conversion probability prediction, a quantitative indicator of ad effectiveness is finally generated. This indicator can reflect the actual influence of the ad, including but not limited to the ad's attractiveness, its ability to guide skiers' behavior, and its performance in different environments or conditions. This indicator can also be combined with other monitoring data from the ski resort to further improve the accuracy of ad effectiveness evaluation.

[0087] Furthermore, it also includes collecting real-time weather warning data, and when the weather warning data reaches a threshold, a confidence correction coefficient is added to the quantitative indicators of advertising effectiveness. Example 2

[0088] Based on the same inventive concept as the ski resort outdoor advertising monitoring data analysis method in the foregoing embodiments, the present invention also provides a ski resort outdoor advertising monitoring data analysis system, the system comprising:

[0089] The data acquisition module collects multi-source monitoring data of outdoor advertising spaces in the ski resort. The multi-source monitoring data includes video streams from the advertising spaces, environmental sensor data, and skier behavior trajectories.

[0090] The delayed image module performs dynamic environmental interference correction on the video stream of the ad slot based on environmental sensor data, and generates an ad exposure image sequence;

[0091] The feature generation module constructs a dynamic spatiotemporal relationship between skiers and ad placements based on the ad exposure image sequence and skier behavior trajectory, and extracts ad attention feature sets.

[0092] The effectiveness quantification module inputs the ad attention feature set into the ad effectiveness evaluation model to generate quantitative indicators of ad effectiveness.

[0093] The adjustment system described above in this invention can effectively realize a method for monitoring and analyzing outdoor advertising data in ski resorts. The technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0094] Furthermore, the feature generation module includes:

[0095] The baseline setting unit uses the visible area of ​​the advertisement exposure image sequence as the baseline plane and projects the skier's behavior trajectory onto the baseline plane;

[0096] The gaze timing unit starts timing the gaze duration when the skier's trajectory enters the reference plane.

[0097] The data recording unit records valid gaze data when the angle between the skier's trajectory and the normal to the reference plane and the gaze duration are less than the threshold.

[0098] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0099] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for analyzing outdoor advertising monitoring data at ski resorts, characterized in that, The method includes: Collect multi-source monitoring data of outdoor advertising spaces at ski resorts. The multi-source monitoring data includes video streams of advertising spaces, environmental sensor data, and skier behavior trajectories. Based on the environmental sensing data, dynamic environmental interference correction is performed on the video stream of the ad slot to generate an ad exposure image sequence; Based on the image sequence of the advertisement exposure and the skier's behavioral trajectory, a dynamic spatiotemporal relationship between the skier and the advertisement placement is constructed, and the advertisement attention feature set is extracted. Input the advertising attention feature set into the advertising effectiveness evaluation model to generate quantitative indicators of advertising effectiveness. Generate an image sequence for ad exposure, including: Acquire the environmental sensing data, including light intensity, snow / fog concentration, and wind speed data; A brightness compensation matrix is ​​constructed based on the light intensity to correct the video stream of the advertising space; The dynamic blur of the video stream in the ad slot is compensated based on the wind speed data; Based on the snow fog concentration, a defogging convolution kernel is generated, and the video stream of the ad slot is corrected according to the defogging convolution kernel to generate the ad exposure image sequence; The generation of skier behavioral trajectories includes: Obtain the skier's starting point on the slope and generate initial position coordinates; Based on the real-time changes of the initial position coordinates, the color of the skier's clothing and the outline of the ski equipment in the corresponding moving area are obtained and merged to generate dynamic visual features; The skier's behavioral trajectory is generated by combining the initial position coordinates with the dynamic visual features.

2. The method for analyzing outdoor advertising monitoring data at ski resorts according to claim 1, characterized in that, Merging to generate dynamic visual features, including: Perform spatial transformation on the color features of the skier's clothing, and extract the hue and saturation components to generate a first feature vector; Key point detection is performed on the outline features of the ski equipment to obtain the outline curvature histogram and generate a second feature vector; The first feature vector and the second feature vector of each skier are merged to output the dynamic visual features; In this process, an attention weight is added to the merging of the first feature vector and the second feature vector. The attention weight is dynamically adjusted based on the real-time distance between the skier and the advertising space.

3. The method for analyzing outdoor advertising monitoring data at ski resorts according to claim 1, characterized in that, The dynamic spatiotemporal correlation includes: Using the visible area of ​​the advertising exposure image sequence as a reference plane, the skier's behavioral trajectory is projected onto the reference plane; The gaze duration timer is activated when the skier's trajectory enters the reference plane. When the angle between the skier's behavior trajectory and the normal of the reference plane and the gaze duration are less than a threshold, the data is recorded as valid gaze data.

4. The method for analyzing outdoor advertising monitoring data at ski resorts according to claim 3, characterized in that, The advertising attention feature set is generated based on the aggregation of the effective gaze data, wherein attention weights are added as accompanying weights to the effective gaze data during the aggregation process.

5. The method for analyzing outdoor advertising monitoring data at ski resorts according to claim 1, characterized in that, Generate quantitative metrics for advertising effectiveness, including: After frequency domain transformation, the effective gaze data in the ad attention feature set is combined with attention weight and historical behavior data to predict the conversion probability of the behavior guided by the ad, and the quantitative index of the ad effect is output.

6. The method for analyzing outdoor advertising monitoring data at ski resorts according to claim 1, characterized in that, It also includes collecting real-time weather warning data, and when the weather warning data reaches a threshold, adding a confidence correction coefficient to the quantitative indicator of advertising effectiveness.

7. The analysis system for the ski resort outdoor advertising monitoring data analysis method according to claim 1, characterized in that, The system includes: The data acquisition module collects multi-source monitoring data of outdoor advertising spaces in the ski resort. The multi-source monitoring data includes video streams from the advertising spaces, environmental sensor data, and skier behavior trajectories. The delayed image module performs dynamic environmental interference correction on the video stream of the ad slot based on environmental sensor data, and generates an ad exposure image sequence; The feature generation module constructs a dynamic spatiotemporal relationship between skiers and ad placements based on the ad exposure image sequence and skier behavior trajectory, and extracts ad attention feature sets. The effectiveness quantification module inputs the ad attention feature set into the ad effectiveness evaluation model to generate quantitative indicators of ad effectiveness.

8. The system according to claim 7, characterized in that, The feature generation module includes: The baseline setting unit uses the visible area of ​​the advertisement exposure image sequence as the baseline plane and projects the skier's behavior trajectory onto the baseline plane; The gaze timing unit starts timing the gaze duration when the skier's trajectory enters the reference plane. The data recording unit records valid gaze data when the angle between the skier's trajectory and the normal to the reference plane and the gaze duration are less than the threshold.

Images