Intelligent signboard multifunctional object system with video monitoring

Abstract

The present application relates to the technical field of image processing, in particular to a kind of intelligent signboard multifunctional internet of things system with video monitoring, comprising: according to the brightness change condition of pixel point on the boundary of tourist dynamic area, the light and dark boundary characteristic value of tourist dynamic area is obtained;According to tourist crowded information value and light and dark boundary characteristic value, the light influence recognition degree in tourist dynamic area is obtained;According to the degree of behavior confusion and light influence recognition degree, the behavior recognition demand degree of tourist dynamic area is obtained;Based on behavior recognition demand degree, each frame intelligent signboard monitoring image is enhanced, and the enhanced image of intelligent signboard monitoring;Through the enhanced image of intelligent signboard monitoring, tourist is indicated and guided.The present application can provide the best multifunctional internet of things scheme to tourist by real-time adjustment of intelligent signboard.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a smart signboard multifunctional Internet of Things system with video monitoring. BACKGROUND

[0002] In modern smart scenic area management, in order to protect the safety of tourists and improve the touring experience, video monitoring devices are usually deployed at key intersections, narrow passages or popular scenic spot entrances; by analyzing the crowd distribution and flow in the monitoring video, the carrying capacity of the region can be evaluated in real time, and through smart signboards and electronic display screens and other Internet of Things terminals, guide information can be issued to tourists to realize intelligent guidance and distribution of people flow, and avoid safety accidents caused by excessive congestion;

[0003] In the evening scenic area, the emergency relief situation is due to the congestion stagnation or extremely slow movement caused by excessive crowd density; in this state, due to the influence of light, there are various light and dark areas in the monitoring video, causing the human body outline of tourists to appear blurred, and then the smart signboard has errors when identifying the number and behavior of tourists in the monitoring video, resulting in false judgments and thus unable to issue alarm and relief instructions to the formed congestion point which is the most dangerous. SUMMARY

[0004] The present application provides a smart signboard multifunctional Internet of Things system with video monitoring to solve the existing problems.

[0005] The smart signboard multifunctional Internet of Things system with video monitoring of the present application adopts the following technical scheme:

[0006] Comprise the following modules:

[0007] An image acquisition module for acquiring a plurality of frames of smart signboard monitoring images;

[0008] A tourist behavior feature acquisition module for acquiring a tourist dynamic area in the smart signboard monitoring image according to the change of the corner points in the adjacent frames of smart signboard monitoring images; acquiring a tourist congestion information value of the tourist dynamic area according to the gradient performance of the pixel points in the tourist dynamic area; acquiring the tourist information richness in the tourist dynamic area according to the tourist congestion information value; and acquiring the behavior confusion degree in the tourist dynamic area according to the tourist information richness, and the motion and distribution of the corner points in the tourist dynamic area;

[0009] The light influence feature acquisition module is used to obtain the light-dark boundary feature value of the dynamic area of ​​tourists based on the brightness change of pixels on the boundary of the dynamic area of ​​tourists; to obtain the light influence recognition degree of the dynamic area of ​​tourists based on the tourist crowding information value and the light-dark boundary feature value; and to obtain the behavior recognition demand degree of the dynamic area of ​​tourists based on the degree of behavior disorder and the light influence recognition degree.

[0010] The multi-functional smart IoT module is used to enhance each frame of smart signage monitoring images based on behavioral recognition requirements, resulting in enhanced images of smart signage monitoring; these enhanced images are then used to guide tourists.

[0011] Preferably, the specific method for obtaining the dynamic area of ​​tourists in the smart signage monitoring image based on the changes in corner points in adjacent frames of the smart signage monitoring image is as follows:

[0012] Using the FAST corner detection algorithm to obtain the first All corner points in the monitoring image of the smart signage;

[0013] For the The first frame of the smart signage monitoring image The corner point is tracked using the pyramid optical flow algorithm. The corner point at the th The location coordinates in the monitoring image of the smart signage; the first The corner point at the th The location coordinates in the smart signage monitoring image and the first The corner point at the th The normalized value of the Euclidean distance between the location coordinates in the frame of the smart signage monitoring image is denoted as the th frame. Motion factors of each corner point;

[0014] Preset a motion threshold If the first The motion factor of each corner point is greater than or equal to the motion threshold. , will the Each corner point is denoted as a dynamic corner point;

[0015] In the In the monitoring image of the smart sign, the DBSCAN clustering algorithm is used to cluster all dynamic corner points based on their position coordinates to obtain several clusters. The convex hull region formed by each cluster is recorded as the dynamic area of ​​the visitor.

[0016] Preferably, the specific method for obtaining the tourist congestion information value of the tourist dynamic area based on the gradient performance of pixels within the tourist dynamic area includes:

[0017] For the The frame intelligent signboard monitors any one tourist dynamic area in the image, and uses a Sobel operator to obtain a gradient amplitude and a gradient direction of each pixel point in the any one tourist dynamic area;

[0018] A direction parameter is preset , to a direction angle of 360 degrees is uniformly divided into intervals, and a frequency of gradient directions of all pixel points in the any one tourist dynamic area falling into the intervals is counted;

[0019] A difference between a probability of the gradient direction of the pixel point in the any one tourist dynamic area falling into the first interval and probabilities of the gradient direction of the pixel point falling into other intervals is accumulated, and the accumulated sum is recorded as a frequency difference of the first interval; and an inverse proportional normalized value of an accumulated sum of the frequency differences of all the intervals is recorded as a texture direction uniformity in the any one tourist dynamic area.

[0020] A product between a mean value of the gradient amplitudes of all the pixel points in the any one tourist dynamic area and the texture direction uniformity in the any one tourist dynamic area is taken as a tourist congestion information value in the any one tourist dynamic area.

[0021] Preferably, the method for obtaining a tourist information richness in the tourist dynamic area according to the tourist congestion information value comprises the following specific method:

[0022] A window parameter is preset , for any one pixel point in the any one tourist dynamic area, a neighborhood window with a size of is constructed with the any one pixel point as a center, and is recorded as a local neighborhood window of the any one pixel point; a variance of gray values of all the pixel points in the local neighborhood window of the any one pixel point is recorded as a local texture value of the any one pixel point; and a mean value of the local texture values of all the pixel points in the any one tourist dynamic area is recorded as a local texture feature distribution value in the any one tourist dynamic area.

[0023] A normalized value of a product between the local texture feature distribution value in the any one tourist dynamic area and the tourist congestion information value in the any one tourist dynamic area is taken as the tourist information richness in the any one tourist dynamic area.

[0024] Preferably, the method for obtaining a behavior confusion degree in the tourist dynamic area according to the tourist information richness and a motion and distribution of a corner point in the tourist dynamic area comprises the following specific method:

[0025] For the The first frame of the smart signage monitoring image Cluster the dynamic corner points within the dynamic area of ​​each tourist to obtain the first... Several activity clusters within a tourist dynamic area; the several activity clusters include main activity clusters and several non-main activity clusters;

[0026] The first The first tourist dynamic area The distance between the cluster of non-primary activity actions and the cluster center is denoted as the i-th. The action differences of non-subjective activity action clusters; the first The action differences of each non-subjective activity action cluster and the first The ratio between the action differences of the main activity action clusters within a tourist dynamic area is denoted as the i-th The difference factor for each non-subjective activity cluster; the first The sum of the difference factors of each non-subjective activity cluster, the first The number of clusters of all types of activities within a tourist dynamic area, and the number of clusters .... The richness of visitor information in each visitor dynamic area, the normalized value of the product of these three factors, is used as the first... The level of disorder in the dynamic areas for tourists.

[0027] Preferably, the first The first frame of the smart signage monitoring image Cluster the dynamic corner points within the dynamic area of ​​each tourist to obtain the first... Clustering of several activity actions within a dynamic area of ​​a tourist, including the following specific methods:

[0028] The first The first frame of the smart signage monitoring image The first visitor dynamic area The position coordinates of the dynamic corner point and the first The first frame of the smart signage monitoring image The first visitor dynamic area The Euclidean distance between the position coordinates of the first dynamic corner point and the direction of the straight line constitute the first... Motion feature vectors of dynamic corner points;

[0029] The inverse proportional value of the cosine similarity between the motion feature vectors of dynamic corner points is denoted as the clustering distance; based on the first... The clustering distance between all dynamic corner points in a tourist dynamic area is used to cluster all dynamic corner points using the DBSCAN clustering algorithm to obtain several activity action clusters;

[0030] In the Among all activity clusters within a visitor's dynamic area, the activity cluster with the largest number of dynamic corner points is denoted as the i-th cluster. The main activity clusters within a visitor's dynamic area; all other activity clusters besides the main activity clusters are denoted as non-main activity clusters.

[0031] Preferably, the specific method for obtaining the light-dark boundary feature value of the tourist dynamic area based on the brightness change of pixels on the boundary of the tourist dynamic area includes:

[0032] For the The first frame of the smart signage monitoring image The first visitor dynamic area, obtain the first The brightness feature value of each pixel in the edge brightness feature sequence of a dynamic tourist area;

[0033] Three preset feature parameters , and The The brightness feature value is and The sequentially arranged segments are denoted as the brightness drastic change characteristic segments; the first segment is... The edge brightness feature sequence of the dynamic tourist area is the first The absolute value of the difference between the brightness value corresponding to the first brightness feature value and the brightness value corresponding to the second brightness feature value within a brightness variation feature segment is denoted as the first brightness feature value. The difference in brightness between segments with significant brightness variations; the first The sum of the brightness differences of all brightness variation segments in the edge brightness feature sequence of a dynamic tourist area is denoted as the i-th. The difference in brightness between the boundaries of each tourist dynamic area;

[0034] The first The number of all brightness feature values ​​within all brightness change feature segments in the edge brightness feature sequence of the dynamic tourist area is related to the number of brightness feature values ​​in the first... The ratio between the total number of pixels in the edge brightness feature sequence of each dynamic tourist area is denoted as the ratio of the total number of pixels in the sequence. The characteristic value of alternating light and dark in the dynamic area of ​​a tourist;

[0035] The first The normalized value of the product between the alternating light and dark characteristic values ​​of each dynamic tourist area and the light and dark difference value at the boundary is used as the first... The characteristic value of the light and dark boundary of the dynamic area of ​​the tourist.

[0036] Preferably, the acquisition of the first The brightness feature value of each pixel in the edge brightness feature sequence of the dynamic tourist area includes the following specific methods:

[0037] The first The smart signage monitoring image is converted into the YUV color space to obtain the brightness value of each pixel; in the first frame... On the boundary of each dynamic tourist area, starting from the pixel with the highest brightness value, rotate clockwise for one full cycle to obtain the [number of pixels]. All pixels on the boundary of a dynamic tourist area, denoted as the i-th Edge brightness sequence of a dynamic tourist area;

[0038] Using the least squares method to apply the first Curve fitting was performed on the brightness values ​​of all pixels in the edge brightness sequence of the dynamic tourist area to obtain the first... Edge brightness fitting curve of a dynamic tourist area;

[0039] For the For any data point on the edge brightness fitting curve of a dynamic tourist area, if the slope of that data point is negative, then the brightness feature value of the corresponding pixel is recorded as... If the slope of any data point is 0, then the brightness feature value of the pixel corresponding to that data point is denoted as... If the slope of any data point is positive, then the brightness feature value of the pixel corresponding to that data point is denoted as... ; will the first The sequence consisting of the brightness feature values ​​of all pixels in the edge brightness sequence of the dynamic tourist area is used as the first... The edge brightness sequence of a visitor's dynamic area.

[0040] Preferably, the specific method for obtaining the light impact recognition degree within the dynamic area of ​​tourists based on tourist crowding information value and light-dark boundary feature value is as follows:

[0041] Through the first The brightness values ​​of all pixels in the edge brightness sequence of the dynamic tourist area constitute the first... The edge brightness histogram of the dynamic tourist area; the first... The root mean square error of the edge brightness histogram of each dynamic tourist area is denoted as the i-th Edge distribution characteristic values ​​of a dynamic tourist area;

[0042] This will be recorded as the number The edge distribution characteristic value of the dynamic tourist area, the first Visitor congestion information values ​​for each visitor dynamic area, and the first visitor congestion information value for the second visitor dynamic area. The characteristic values ​​of the light and dark boundary of each dynamic tourist area, and the normalized value of the product of these three values ​​are used as the first... Light within a visitor's dynamic area affects visibility.

[0043] Preferably, the specific method for obtaining the behavioral recognition requirement of the tourist dynamic area based on the degree of behavioral disorder and the influence of light on recognition is as follows:

[0044] The first The influence of light on the visibility of the first tourist dynamic area is related to the first The average level of behavioral disorder in each tourist dynamic area is denoted as the first mean; the average level of behavioral disorder in the second tourist dynamic area is denoted as the third mean. The normalized value of the product of the mean brightness values ​​of all pixels within the dynamic area of ​​a tourist and the first mean is used as the first... The need for behavioral recognition in dynamic tourist areas.

[0045] The beneficial effects of the technical solution of this invention are as follows: This invention obtains the light-dark boundary feature value of the dynamic tourist area based on the brightness change of pixels on the boundary of the dynamic tourist area; obtains the light influence recognition degree of the dynamic tourist area based on the tourist crowding information value and the light-dark boundary feature value; obtains the behavioral recognition demand degree of the dynamic tourist area based on the degree of behavioral chaos and the light influence recognition degree; enhances each frame of the smart sign monitoring image based on the behavioral recognition demand degree, resulting in an enhanced image of the smart sign monitoring; guides tourists through the enhanced image of the smart sign monitoring; thereby, it can accurately locate abnormal tourist areas with high chaos or high density in crowded and poorly lit environments, and the smart sign can immediately pop up a customized warning for the area; and can adjust in real time to show tourists the best multi-functional IoT solution. 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 of 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 structural block diagram of a smart signage multi-functional IoT system with video surveillance according to the present invention;

[0048] Figure 2A feature relationship flowchart of the intelligent signboard multifunctional object system with video monitoring. DETAILED DESCRIPTION

[0049] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object of the application, the specific embodiments, structure, features and effects of the intelligent signboard multifunctional object system with video monitoring according to the present application are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0050] 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 the present application belongs.

[0051] The specific scheme of the intelligent signboard multifunctional object system with video monitoring provided by the present application is described in detail below with reference to the accompanying drawings.

[0052] Please refer to Figure 1 which shows the structure block diagram of the intelligent signboard multifunctional object system with video monitoring provided by one embodiment of the present application, which includes the following modules:

[0053] The image acquisition module is used to acquire a plurality of frames of intelligent signboard monitoring images.

[0054] Specifically, first, a plurality of frames of intelligent signboard monitoring images need to be acquired, and the specific process is as follows:

[0055] The monitoring video of the intelligent signboard is acquired through the monitoring camera of the intelligent signboard in the scenic area; the monitoring video of the intelligent signboard is processed by frame using FFmpeg technology to acquire a plurality of frames of intelligent signboard monitoring images; each frame of intelligent signboard monitoring image is processed by Mask R-CNN neural network and using cross-entropy loss function as the loss function to remove the background influence of the non-tourist activity area.

[0056] Among them, FFmpeg technology and Mask R-CNN neural network are prior art, and this embodiment will not be described in detail here.

[0057] At this point, a plurality of frames of intelligent signboard monitoring images are obtained by the above method.

[0058] The tourist behavior feature acquisition module is used to acquire the dynamic area of ​​tourists in the smart signage monitoring image based on the changes in corner points in adjacent frames of the smart signage monitoring image; acquire the tourist congestion information value of the dynamic area based on the gradient performance of pixels in the dynamic area; acquire the tourist information richness in the dynamic area based on the tourist congestion information value; and acquire the degree of behavioral disorder in the dynamic area based on the tourist information richness and the movement and distribution of corner points in the dynamic area.

[0059] It should be noted that while the static areas within the smart signage monitoring images are mostly tourist areas for leisure and recreation, the dynamic areas are often densely populated and frequently active, posing significant safety hazards. Therefore, it is necessary to focus on the dynamic areas of tourists. This can be achieved by adaptively acquiring the dynamic areas of tourists in the smart signage monitoring images based on the dynamic changes of corner points in adjacent frames. Since different dynamic areas of tourists have different activity frequencies, the richness of tourist activity information varies. The richer the tourist activity information, the higher the degree of behavioral chaos should be.

[0060] Preferably, in some embodiments of the present invention, the specific method for obtaining the dynamic area of ​​tourists in the smart signage monitoring image based on the changes in corner points in adjacent frames of smart signage monitoring images is as follows:

[0061] Each frame of the smart sign monitoring image is grayscaled to [0, 255]. A Cartesian coordinate system is constructed with the vertex of the lower left corner of each frame of the smart sign monitoring image as the origin, the horizontal direction to the right as the positive direction of the horizontal axis, and the vertical direction to the up as the positive direction of the vertical axis. The position coordinates of each pixel in each frame of the smart sign monitoring image are obtained on the Cartesian coordinate system.

[0062] Using the FAST corner detection algorithm to obtain the first All corner points in the monitoring image of the smart signage;

[0063] Because the corner points in static regions change very little in consecutive frames, while the corner points in dynamic regions undergo significant displacement; for the first... The first frame of the smart signage monitoring image The corner point is tracked using the pyramid optical flow algorithm. The corner point at the th The location coordinates in the monitoring image of the smart signage; the first The corner point at the The location coordinates in the smart signage monitoring image and the first The corner point at the th The normalized value of the Euclidean distance between the location coordinates in the frame of the smart signage monitoring image is denoted as the th frame. Motion factors of each corner point;

[0064] Preset a motion threshold In this embodiment, This example is used for illustration; no specific limitations are set in this embodiment. It depends on the specific implementation situation;

[0065] If the first The motion factor of each corner point is greater than or equal to the motion threshold. , will the The first corner point is denoted as the dynamic corner point; the second corner point is denoted as the dynamic corner point. In the monitoring image of the smart sign, the DBSCAN clustering algorithm is used to cluster all dynamic corner points based on the position coordinates of the dynamic corner points to obtain several clusters. The convex hull region formed by each cluster is recorded as the dynamic area of ​​the visitor.

[0066] For corners that cannot be tracked, no further analysis is performed. Since the position coordinates between dynamic corners are far apart, there is no overlap between convex hull regions. The convex hull detection algorithm is used to obtain the convex hull region. The DBSCAN clustering algorithm, pyramid optical flow algorithm, FAST corner detection algorithm and gray value processing are all existing technologies, and will not be described in detail here.

[0067] It should be noted that when the dynamic area of ​​tourists in a scenic spot contains a large number of crowded people and frequent tourist activities, the dense activities, diverse behaviors, and rapid changes of the area make the texture of the dynamic area of ​​tourists complex and varied, with large local variance. However, when there are only a few tourists in the dynamic area of ​​tourists, and their activities are relatively infrequent, such as leisure, sightseeing, and experience, that is, when the activities of tourists are relatively static and infrequent, the texture changes in the dynamic area of ​​tourists are limited, the visual features of the area are relatively stable, resulting in low texture richness and small local variance.

[0068] Preferably, in some embodiments of the present invention, when the dynamic area of ​​a tourist contains a large number of edges and details, such as the outline of a crowd of tourists and the folds of clothing, the gradient amplitude of the pixels within the dynamic area of ​​the tourist is generally high; conversely, if the content of the dynamic area of ​​the tourist is leisure, viewing, or experiential activities with low activity frequency and few tourists, the gradient amplitude of the pixels within the dynamic area of ​​the tourist is generally low. Therefore, the specific method for obtaining the tourist crowding information value of the dynamic area of ​​the tourist based on the gradient performance of the pixels within the dynamic area of ​​the tourist is as follows:

[0069] For the In any dynamic area of ​​a visitor in the frame-based smart signage monitoring image, the Sobel operator is used to obtain the gradient magnitude and gradient direction of each pixel in that dynamic area of ​​a visitor.

[0070] Wherein, since having high gradient amplitude is not enough to determine as a tourist crowded area; such as a tourist dynamic area containing building clear right-angle edge, the gradient amplitude of its internal pixel points may also be high; but its texture is regular, single; and the texture feature of crowded tourists is that it is chaotic, that is, the gradient direction is diverse and uniformly distributed; Sobel operator is prior art, this embodiment does not make too much description here.

[0071] Pre-set a direction parameter Wherein, this embodiment takes as an example for description, and the embodiment is not specifically limited, wherein According to the specific implementation condition;

[0072] Divide the direction angle of to degrees into intervals, and count the frequency of the gradient direction of all pixel points in the arbitrary tourist dynamic area falling into the intervals;

[0073] The cumulative sum of the difference between the probability of the gradient direction of the pixel points in the arbitrary tourist dynamic area falling into the first interval and the probability of falling into other intervals is recorded as the frequency difference of the first interval; and the inverse proportional normalized value of the cumulative sum of the frequency difference of all intervals is recorded as the texture direction uniformity in the arbitrary tourist dynamic area.

[0074] The product between the mean value of the gradient amplitude of all pixel points in the arbitrary tourist dynamic area and the texture direction uniformity in the arbitrary tourist dynamic area is taken as the tourist crowded information value in the arbitrary tourist dynamic area.

[0075] The specific formula is as follows:

[0076]

[0077] In the formula, represents the tourist crowded information value of the tourist dynamic area; represents the mean value of the gradient amplitude of all pixel points in the tourist dynamic area; represents the probability of the gradient direction of the pixel points in the tourist dynamic area falling into the first interval; represents the probability of the gradient direction of the pixel points in the tourist dynamic area falling into the first interval; represents the exponential function with natural constant as base number.

[0078] Preferably, in some embodiments of the present invention, the larger the local texture feature distribution value and the larger the tourist congestion information value within the tourist dynamic area, the more likely the tourist dynamic area contains a large amount of frequently active tourist information; the specific method for obtaining the richness of tourist information in the tourist dynamic area based on the tourist congestion information value is as follows:

[0079] Preset a window parameter In this embodiment, This example is used for illustration; no specific limitations are set in this embodiment. It depends on the specific implementation situation;

[0080] For any pixel within any tourist dynamic area, construct a region of size [missing information] centered on that pixel. The neighborhood window of any pixel is denoted as the local neighborhood window of any pixel; the variance of the gray values ​​of all pixels in the local neighborhood window of any pixel is denoted as the local texture value of any pixel; the mean of the local texture values ​​of all pixels in any dynamic area of ​​a tourist is denoted as the local texture feature distribution value of any dynamic area of ​​a tourist.

[0081] The normalized value of the product between the local texture feature distribution value in any dynamic tourist area and the tourist crowding information value in any dynamic tourist area is taken as the tourist information richness in any dynamic tourist area.

[0082] The specific formula is as follows:

[0083]

[0084] In the formula, This indicates the richness of visitor information in the visitor dynamic area; This indicates the crowding information value for the dynamic area of ​​tourists; This represents the distribution values ​​of local texture features within the dynamic area of ​​the tourist area; This represents the linear normalization function.

[0085] Preferably, in some embodiments of the present invention, the higher the richness of tourist information in the tourist dynamic area, the more active behaviors of tourists in the tourist dynamic area. When there are behaviors such as tourists running in the opposite direction or multiple people running in a chaotic and disorderly manner, it is highly likely to cause potential dangers. The specific method for obtaining the degree of disorder of behavior in the tourist dynamic area based on the richness of tourist information and the movement and distribution of corner points in the tourist dynamic area is as follows:

[0086] The first The first frame of the smart signage monitoring image The first visitor dynamic area The position coordinates of the dynamic corner point and the first The first frame of the smart signage monitoring image The first visitor dynamic area The Euclidean distance between the position coordinates of the first dynamic corner point and the direction of the straight line constitute the first... Motion feature vectors of dynamic corner points;

[0087] Among them, the The first frame of the smart signage monitoring image The first visitor dynamic area The position coordinates of the _th dynamic corner point are obtained through the pyramid optical flow algorithm. The first frame of the smart signage monitoring image The first visitor dynamic area The corresponding positions of each dynamic corner point.

[0088] For the For any two dynamic corner points in a dynamic area of ​​a tourist, the difference in the cosine similarity between the motion feature vectors of the two dynamic corner points and the first point is denoted as the clustering distance between the two dynamic corner points; based on the first... The clustering distance between all dynamic corner points in a tourist dynamic area is used to cluster all dynamic corner points using the DBSCAN clustering algorithm to obtain several activity action clusters;

[0089] In the Among all activity clusters within a visitor's dynamic area, the activity cluster with the largest number of dynamic corner points is denoted as the i-th cluster. The main activity clusters within a visitor's dynamic area; all other activity clusters besides the main activity clusters are categorized as non-main activity clusters.

[0090] Among them, when the first The more activity clusters there are in the clustering results of dynamic corner points within a tourist's dynamic area, the more it indicates that the... The more complex and chaotic the behavior of tourists in a tourist dynamic area, and the greater the distance between clusters, the greater the difference in tourist behavior, and the more attention needs to be paid to the degree of disorder in tourist behavior.

[0091] The first The first tourist dynamic area The distance between the cluster of non-primary activity actions and the cluster center is denoted as the i-th. The action differences of non-subjective activity action clusters; the first The action differences of each non-subjective activity action cluster and the first The ratio between the action differences of the main activity action clusters within a tourist dynamic area is denoted as the i-th The difference factor for each non-subjective activity cluster; the first The sum of the difference factors of each non-subjective activity cluster, the first The number of clusters of all types of activities within a tourist dynamic area, and the number of clusters .... The richness of visitor information in each visitor dynamic area, the normalized value of the product of these three factors, is used as the first... The degree of behavioral disorder in the dynamic areas for individual tourists;

[0092] The specific formula is as follows:

[0093]

[0094] In the formula, Indicates the first The degree of behavioral disorder in the dynamic areas for individual tourists; Indicates the first The richness of visitor information in each visitor dynamic area; Indicates the first The number of clusters of all types of activities and actions within a tourist's dynamic area; Indicates the first The number of clusters in each visitor dynamic area; Indicates the first The first tourist dynamic area The distance between a cluster of non-subjective activity actions and its cluster center; Indicates the first The distance between the main activity clusters and the cluster centers within a tourist dynamic area; This represents the linear normalization function.

[0095] Thus, the degree of behavioral disorder in the dynamic area of ​​tourists can be obtained through the above methods.

[0096] The light influence feature acquisition module is used to obtain the light-dark boundary feature value of the dynamic area of ​​tourists based on the brightness change of pixels on the boundary of the dynamic area of ​​tourists; to obtain the light influence recognition degree of the dynamic area of ​​tourists based on the tourist crowding information value and the light-dark boundary feature value; and to obtain the behavior recognition demand degree of the dynamic area of ​​tourists based on the degree of behavior disorder and the light influence recognition degree.

[0097] It should be noted that due to the change of light at the evening, the monitoring video picture of the scenic spot is affected by the lighting, causing the alternation of bright and dark areas; the bright places are clearly visible, while the shadow part presents different degrees of blur, making the outline of the tourists in these areas become unclear or difficult to identify; that is, due to the dynamic change of video monitoring in this light environment, the details of the overall picture are relatively blurred, especially in places where the crowd is dense or the light is insufficient, the human body outline cannot be clearly presented.

[0098] Preferably, in some embodiments of the present application, when the tourist dynamic area is located in the light and dark alternating area, the brightness value of the pixel point on the boundary of the tourist dynamic area will present the characteristics of sharp and high-frequency fluctuation, which greatly interferes with the stability of the tourist contour recognition; according to the brightness change of the pixel point on the boundary of the tourist dynamic area, the specific method for obtaining the light-dark junction characteristic value of the tourist dynamic area is:

[0099] the first frame of intelligent signboard monitoring image is converted into YUV color space to obtain the brightness value of each pixel point; on the boundary of the first tourist dynamic area, starting from the pixel point with the maximum brightness value, clockwise rotation for one week is performed to traverse to obtain all the pixel points on the boundary of the first tourist dynamic area, and recorded as the edge brightness sequence of the first tourist dynamic area;

[0100] the brightness values of all the pixel points in the edge brightness sequence of the first tourist dynamic area are curve fitted by using the least square method to obtain the edge brightness fitting curve of the first tourist dynamic area;

[0101] three characteristic parameters , and are preset, wherein , the present embodiment is described taking , and as examples, and the present embodiment is not specifically limited, wherein , and are determined according to the specific implementation;

[0102] for any one data point on the edge brightness fitting curve of the first tourist dynamic area, if the slope of the arbitrary one data point is negative, the brightness characteristic value of the pixel point corresponding to the arbitrary one data point is recorded as ; if the slope of the arbitrary one data point is 0, the brightness characteristic value of the pixel point corresponding to the arbitrary one data point is recorded as If the slope of the arbitrary data point is positive, the luminance feature value of the pixel corresponding to the arbitrary data point is recorded as The sequence of the luminance feature values of all the pixels in the edge luminance sequence of the first tourist dynamic area is taken as the edge luminance sequence of the first tourist dynamic area.

[0103] The sequence segment composed of the luminance feature values of and in order is recorded as a luminance abrupt change feature segment.The absolute value of the difference between the luminance value corresponding to the first luminance feature value and the luminance value corresponding to the second luminance feature value in the first luminance abrupt change feature segment in the edge luminance feature sequence of the first tourist dynamic area is recorded as the light-dark difference degree of the first luminance abrupt change feature segment.The cumulative sum of the light-dark difference degrees of all the luminance abrupt change feature segments in the edge luminance feature sequence of the first tourist dynamic area is recorded as the boundary light-dark difference value of the first tourist dynamic area.

[0104] The ratio between the number of all the luminance feature values in all the luminance abrupt change feature segments in the edge luminance feature sequence of the first tourist dynamic area and the total number of all the pixels in the edge luminance feature sequence of the first tourist dynamic area is recorded as the light-dark alternation feature value of the first tourist dynamic area.

[0105] The normalized value of the product between the light-dark alternation feature value of the first tourist dynamic area and the boundary light-dark difference value is taken as the light-dark junction feature value of the first tourist dynamic area.

[0106] The specific formula is as follows:

[0107]

[0108] In the formula, L represents the light-dark junction feature value of the first tourist dynamic area.N represents the number of all the luminance feature values in all the luminance abrupt change feature segments in the edge luminance feature sequence of the first tourist dynamic area.M represents the total number of all the pixels in the edge luminance feature sequence of the first tourist dynamic area. The number of all brightness-changing feature segments in the edge brightness feature sequence of a tourist dynamic area; Indicates the first The edge brightness feature sequence of the dynamic tourist area is the first The brightness value corresponding to the first brightness feature value within a brightness variation feature segment; Indicates the first The edge brightness feature sequence of the dynamic tourist area is the first The brightness value corresponding to the second brightness feature value within a brightness drastic change feature segment; This represents the linear normalization function.

[0109] Among them, if the first When the edge brightness feature sequence of a tourist dynamic area contains a large number of brightness drastic change feature segments, it indicates that the brightness value of the edge pixel on its boundary drops sharply from a high value to a low value in a short distance, and then rises again, forming an unstable sawtooth fluctuation; then the probability that its location is at the junction of light and dark is greater; the least squares method is an existing technology, and will not be described in detail here.

[0110] Preferably, in some embodiments of the present invention, when the dynamic area of ​​tourists becomes crowded, the alternating mutual occlusion and gaps between people will also drastically increase the high-frequency oscillation of the edge brightness of the dynamic area of ​​tourists; therefore, if the dynamic area of ​​tourists spans both light and dark areas, and the greater the tourist crowding information value, the more obvious the bimodal distribution of edge pixel brightness will be, and the greater its light influence recognition will be; the specific method for obtaining the light influence recognition within the dynamic area of ​​tourists based on the tourist crowding information value and the light-dark boundary feature value is as follows:

[0111] Through the first The brightness values ​​of all pixels in the edge brightness sequence of the dynamic tourist area constitute the first... The edge brightness histogram of the dynamic tourist area; the first... The root mean square error of the edge brightness histogram of each dynamic tourist area is denoted as the i-th Edge distribution characteristic values ​​of a dynamic tourist area;

[0112] This will be recorded as the number The edge distribution characteristic value of the dynamic tourist area, the first Visitor congestion information values ​​for each visitor dynamic area, and the first visitor congestion information value for the second visitor dynamic area. The characteristic values ​​of the light and dark boundary of each dynamic tourist area, and the normalized value of the product of these three values ​​are used as the first... Light within a visitor's dynamic area affects visibility;

[0113] The specific formula is as follows:

[0114]

[0115] wherein, represents the light influence recognition degree in the first tourist dynamic area; represents the edge distribution feature value of the first tourist dynamic area; represents the tourist congestion information value of the first tourist dynamic area; represents the light-dark boundary feature value of the first tourist dynamic area; represents the light influence recognition degree in the first tourist dynamic area; represents the edge distribution feature value of the first tourist dynamic area; represents the tourist congestion information value of the first tourist dynamic area; represents the light-dark boundary feature value of the first tourist dynamic area; represents a linear normalization function.

[0116] It should be noted that, in the evening scenic environment, the greater the light influence recognition degree in the tourist dynamic area and the greater the behavior confusion degree, the more complex the behavior actions in the tourist dynamic area, and the greater the security risks, the higher the behavior recognition requirement degree of the tourist dynamic area.

[0117] Preferably, in some embodiments of the present application, according to the behavior confusion degree and the light influence recognition degree, the specific method for obtaining the behavior recognition requirement degree of the tourist dynamic area is as follows:

[0118] the light influence recognition degree in the first tourist dynamic area is multiplied by the mean value of the behavior confusion degree in the first tourist dynamic area, and the product is normalized to obtain the behavior recognition requirement degree of the first tourist dynamic area; the light influence recognition degree in the first tourist dynamic area is multiplied by the mean value of the behavior confusion degree in the first tourist dynamic area, and the product is normalized to obtain the behavior recognition requirement degree of the first tourist dynamic area;

[0119] The specific formula is as follows:

[0120]

[0121] wherein, represents the behavior recognition requirement degree of the first tourist dynamic area; represents the light influence recognition degree in the first tourist dynamic area; represents the behavior confusion degree in the first tourist dynamic area; represents the light influence recognition degree in the first tourist dynamic area; represents the behavior confusion degree in the first tourist dynamic area; represents the mean value of the brightness values of all pixel points in the first tourist dynamic area. Similarly, the behavior recognition requirement degree of each tourist dynamic area in each frame of intelligent signboard monitoring image is obtained.

[0122] Similarly, the behavior recognition requirement degree of each tourist dynamic area in each frame of intelligent signboard monitoring image is obtained.

[0123] ​​​​Until now, the behavior recognition demand degree of each tourist dynamic area in each frame of intelligent signboard monitoring image is obtained by the above method.

[0124] The multifunctional intelligent module is used for enhancing each frame of intelligent signboard monitoring image based on the behavior recognition demand degree, and the enhanced image of the intelligent signboard monitoring; and the tourists are guided by the enhanced image of the intelligent signboard monitoring.

[0125] Preferably, in some embodiments of the present application, the specific method for enhancing each frame of intelligent signboard monitoring image based on the behavior recognition demand degree is that:

[0126] It should be noted that the adaptive Laplacian sharpening algorithm is used to enhance each frame of intelligent signboard monitoring image to remove the edge contour blur of the tourists; the adaptive Laplacian sharpening algorithm gives each pixel point a sharpening intensity, the value range of the sharpening intensity is between 0 and 1, and the value of the sharpening intensity determines the mixing degree of the original image and the sharpened image; when the sharpening intensity is 0, there is no sharpening effect, and the output image is the same as the original image; when the sharpening intensity is 1, the output image is completely determined by the sharpened image, and the sharpening effect is the strongest; other values between 0 and 1 represent different degrees of sharpening effect.

[0127] Taking any one frame of intelligent signboard monitoring image as an example, the sharpening intensity of each pixel point not in the tourist dynamic area is 0; the behavior recognition demand degree of the tourist dynamic area is used as the sharpening intensity of each pixel point in the tourist dynamic area; and then the sharpening intensity of each pixel point in the intelligent signboard monitoring image is obtained; the Laplacian operator of eight neighborhoods is used to obtain the Laplacian operator value of each pixel point in the intelligent signboard monitoring image; the gray value of all pixel points in the intelligent signboard monitoring image after adaptive Laplacian sharpening is used to form a sharpening enhanced image, which is used as the enhanced image of the intelligent signboard monitoring.

[0128] Preferably, in some embodiments of the present application, the specific method for guiding the tourists by the enhanced image of the intelligent signboard monitoring is that:

[0129] According to the sharpening enhanced image of the intelligent signboard monitoring image, the number level (high, medium and low) of the tourists is recognized in real time by combining the deep learning model EITNet.

[0130] If the number of tourists is high according to the real-time monitoring video of the smart signboard, the smart signboard can easily cause congestion and pushing behavior in the tourist dynamic area, which can cause accidents such as stampede. Therefore, the smart signboard adjusts the light brightness to the highest level, and the voice broadcast content is: Dear tourists, the number of tourists in the current area is large, please keep order and do not crowd; for your safety, please keep a proper distance to avoid accidents; and the signboard guides tourists to move to the crowd dispersal area.

[0131] If the number of tourists is medium according to the real-time monitoring video of the smart signboard, the smart signboard adjusts the light brightness to the medium level, and the voice broadcast content is: There is a gathering of people in the scenic area, please follow the order, keep a proper distance, and ensure safe passage; if there is an emergency, please ask for help from the staff immediately; and the signboard guides tourists to move to the crowd dispersal area.

[0132] If the number of tourists is low according to the real-time monitoring video of the smart signboard, the smart signboard adjusts the light brightness to the low level to avoid wasting energy when there are few tourists, and to maintain a comfortable visual experience. The voice broadcast content is: Dear tourists, there are few tourists in the current area, please enjoy the scenic area with peace of mind, and please follow the regulations of the scenic area to maintain good order, and please consult the staff at any time if you need help; and the signboard guides tourists to move to a more lively or popular area to balance the flow of people in the scenic area, avoid some areas being too empty due to few tourists, and improve the overall experience of tourists and the operation efficiency of the scenic area.

[0133] Among them, the adaptive Laplacian sharpening algorithm, the deep learning model EITNet and the eight-neighbor Laplacian operator are prior art, and this embodiment will not be described in detail.

[0134] At this point, the embodiment is completed; please refer to Figure 2 which shows a feature relationship flowchart of a smart signboard multifunctional object system with video monitoring.

[0135] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A smart signboard multifunctional object system with video monitoring, characterized in that, The system comprises the following modules: An image acquisition module is configured to acquire a plurality of frames of intelligent signboard monitoring images; A tourist behavior feature acquisition module is configured to acquire a tourist dynamic area in the intelligent signboard monitoring images according to a change in a corner point in adjacent frames of the intelligent signboard monitoring images; acquire a tourist congestion information value of the tourist dynamic area according to a gradient performance of a pixel point in the tourist dynamic area; acquire a tourist information richness in the tourist dynamic area according to the tourist congestion information value; and acquire a behavior confusion degree in the tourist dynamic area according to the tourist information richness and a motion and distribution of the corner point in the tourist dynamic area; A light influence feature acquisition module is configured to acquire a light and dark boundary feature value of the tourist dynamic area according to a luminance change of a pixel point on a boundary of the tourist dynamic area; acquire a light influence recognition degree in the tourist dynamic area according to the tourist congestion information value and the light and dark boundary feature value; and acquire a behavior recognition demand degree of the tourist dynamic area according to the behavior confusion degree and the light influence recognition degree; A multi-functional intelligent internet module is configured to enhance each frame of the intelligent signboard monitoring images based on the behavior recognition demand degree to obtain an enhanced image of the intelligent signboard monitoring; and guide tourists through the enhanced image of the intelligent signboard monitoring. The specific method for obtaining the light-dark boundary feature value of the dynamic area of ​​tourists includes: for the first... The first frame of the smart signage monitoring image The first visitor dynamic area, obtain the first The brightness feature value of each pixel in the edge brightness feature sequence of a dynamic tourist area; three preset feature parameters. , and The The brightness feature value is and The sequentially arranged segments are denoted as the brightness drastic change characteristic segments; the first segment is... The edge brightness feature sequence of the dynamic tourist area is the first The absolute value of the difference between the brightness value corresponding to the first brightness feature value and the brightness value corresponding to the second brightness feature value within a brightness variation feature segment is denoted as the first brightness feature value. The difference in brightness between segments with significant brightness variations; the first The sum of the brightness differences of all brightness variation segments in the edge brightness feature sequence of a dynamic tourist area is denoted as the i-th. The difference in brightness at the boundary of each tourist dynamic area; the first... The number of all brightness feature values ​​within all brightness change feature segments in the edge brightness feature sequence of the dynamic tourist area is related to the number of brightness feature values ​​in the first... The ratio between the total number of pixels in the edge brightness feature sequence of each dynamic tourist area is denoted as the ratio of the total number of pixels in the sequence. The characteristic value of alternating light and dark in the dynamic area of ​​the tourist area; the first The normalized value of the product between the alternating light and dark characteristic values ​​of each dynamic tourist area and the light and dark difference value at the boundary is used as the first... The characteristic value of the light-dark boundary of a tourist dynamic area; The method for obtaining the light impact recognition degree within the dynamic area of ​​tourists based on tourist crowding information value and light-dark boundary feature value includes: through the first... The brightness values ​​of all pixels in the edge brightness sequence of the dynamic tourist area constitute the first... The edge brightness histogram of the dynamic tourist area; the first... The root mean square error of the edge brightness histogram of each dynamic tourist area is denoted as the i-th The edge distribution characteristic value of the dynamic area of ​​the tourist; the first The edge distribution characteristic value of the dynamic tourist area, the first Visitor congestion information values ​​for each visitor dynamic area, and the first visitor congestion information value for the second visitor dynamic area. The characteristic values ​​of the light and dark boundary of each dynamic tourist area, and the normalized value of the product of these three values ​​are used as the first... Light within a visitor's dynamic area affects visibility; The method for obtaining the behavior recognition demand degree of the tourist dynamic area according to the behavior confusion degree and the light influence recognition degree comprises the following specific steps: The mean value of the light influence recognition degree in the first tourist dynamic area and the behavior confusion degree in the second tourist dynamic area is denoted as a first mean value; The normalized value of the product of the mean value of the brightness value of all pixel points in the first tourist dynamic area and the first mean value is taken as the behavior recognition demand degree of the first tourist dynamic area. ​​ The method for enhancing each frame of the intelligent signboard monitoring images based on the behavior recognition demand degree to obtain the enhanced image of the intelligent signboard monitoring comprises the following steps: taking a sharpening intensity of each pixel point not in the tourist dynamic area as 0; taking the behavior recognition demand degree of the tourist dynamic area as a sharpening intensity of each pixel point in the tourist dynamic area; using a Laplacian operator of eight neighborhoods to obtain a Laplacian operator value of each pixel point in the intelligent signboard monitoring image; and performing adaptive Laplacian sharpening processing on each frame of the intelligent signboard monitoring images to obtain the enhanced image of the intelligent signboard monitoring.

2. The intelligent signboard multifunctional object system with video monitoring according to claim 1, characterized in that, The method for acquiring the tourist dynamic area in the intelligent signboard monitoring images according to a change in a corner point in adjacent frames of the intelligent signboard monitoring images comprises the following steps: The FAST corner detection algorithm is used to obtain the first All the corners in the monitoring image of the intelligent signboard are acquired. For the The first frame of the smart signage monitoring image The corner point is tracked using the pyramid optical flow algorithm. The corner point at the th The location coordinates in the monitoring image of the smart signage; the first The corner point at the th The location coordinates in the smart signage monitoring image and the first The corner point at the th The normalized value of the Euclidean distance between the location coordinates in the frame of the smart signage monitoring image is denoted as the th frame. Motion factors of each corner point; pre-set a motion threshold , if the motion factor of the first corner point is greater than or equal to the motion threshold , the first corner point is recorded as a dynamic corner point; In the first In the frame of the intelligent signboard monitoring image, all dynamic corner points are clustered by using the DBSCAN clustering algorithm based on the position coordinates of the dynamic corner points, a plurality of clustering clusters are obtained, and the convex hull region formed by each clustering cluster is recorded as a tourist dynamic region. 3.The smart signboard multifunctional object system with video monitoring according to claim 1, wherein, The method for acquiring the tourist congestion information value of the tourist dynamic area according to a gradient performance of a pixel point in the tourist dynamic area comprises the following steps: For the first The frame intelligent signboard monitors any one tourist dynamic area in the image, and uses a Sobel operator to obtain the gradient amplitude and gradient direction of each pixel point in the any one tourist dynamic area. A direction parameter is preset , the direction angle of the image is divided into 8 intervals, and the frequency of the gradient direction of all pixel points in the dynamic region of the arbitrary tourist falling into the 8 intervals is counted ​​​​ The accumulated sum of the difference between the probability of the gradient direction of the pixel point in the arbitrary one of the tourist dynamic areas falling into the first interval and the probability of falling into other intervals is denoted as the frequency difference of the first interval. The inverse proportional normalized value of the accumulated sum of the frequency differences of all intervals is denoted as the texture direction uniformity in the arbitrary one of the tourist dynamic areas. Taking a product of a mean value of a gradient amplitude of all pixel points in the arbitrary one tourist dynamic area and a texture direction uniformity in the arbitrary one tourist dynamic area as the tourist congestion information value in the arbitrary one tourist dynamic area.

4. The multi-functional object system with intelligent signboard and video monitoring according to claim 3, characterized in that, The method for acquiring the tourist information richness in the tourist dynamic area according to the tourist congestion information value comprises the following steps: Pre-set a window parameter For any one pixel point in the arbitrary one tourist dynamic area, a neighborhood window with a size of is constructed with the arbitrary one pixel point as the center, and is recorded as the local neighborhood window of the arbitrary one pixel point; the variance of the gray values of all pixel points in the local neighborhood window of the arbitrary one pixel point is recorded as the local texture value of the arbitrary one pixel point; the mean value of the local texture values of all pixel points in the arbitrary one tourist dynamic area is recorded as the local texture feature distribution value in the arbitrary one tourist dynamic area. Taking a normalized value of a product of a local texture feature distribution value in the arbitrary one tourist dynamic area and the tourist congestion information value in the arbitrary one tourist dynamic area as the tourist information richness in the arbitrary one tourist dynamic area.

5. The multi-functional object system with intelligent signboard and video monitoring according to claim 2, characterized in that, The method for acquiring the behavior confusion degree in the tourist dynamic area according to the tourist information richness and a motion and distribution of the corner point in the tourist dynamic area comprises the following steps: The first Frame intelligent signboard monitors the dynamic corner in the first Tourist dynamic area, obtains a plurality of activity motion clustering clusters in the first Tourist dynamic area; the plurality of activity motion clustering clusters include a subject activity motion clustering cluster and a plurality of non-subject activity motion clustering clusters. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The distance between the first non-agent activity motion clustering cluster and the clustering center in the first tourist dynamic area is recorded as the motion difference of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster. The ratio between the motion difference of the first non-agent activity motion clustering cluster and the motion difference of the first non-agent activity motion clustering cluster in the first tourist dynamic area is recorded as the difference factor of the first non-agent activity motion clustering cluster.

6. The intelligent signboard multifunctional object system with video monitoring according to claim 5, characterized in that, The first Frame intelligent signboard monitors the dynamic corner in the first Tourist dynamic area, and clusters the dynamic corner to obtain a first Tourist dynamic area activity motion clustering cluster, including the specific method: The first The first frame of the smart signage monitoring image The first visitor dynamic area The position coordinates of the dynamic corner point and the first The first frame of the smart signage monitoring image The first visitor dynamic area The Euclidean distance between the position coordinates of the first dynamic corner point and the direction of the straight line constitute the first... Motion feature vectors of dynamic corner points; The inverse proportional value of the cosine similarity between the motion feature vectors of the dynamic corner points is denoted as a clustering distance; based on the clustering distances between all the dynamic corner points in the dynamic region of the first tourist, all the dynamic corner points are clustered by using a DBSCAN clustering algorithm to obtain a plurality of activity motion clustering clusters. The inverse proportional value of the cosine similarity between the motion feature vectors of the dynamic corner points is denoted as a clustering distance; based on the clustering distances between all the dynamic corner points in the dynamic region of the first tourist, all the dynamic corner points are clustered by using a DBSCAN clustering algorithm to obtain a plurality of activity motion clustering clusters. In all the activity motion clustering clusters in the first dynamic area of the tourists, the activity motion clustering cluster with the largest number of dynamic corner points is recorded as the main activity motion clustering cluster in the first dynamic area of the tourists; all the other activity motion clustering clusters except the main activity motion clustering cluster are recorded as non-main activity motion clustering clusters. 7.The smart signboard multifunctional object system with video monitoring according to claim 1, wherein, The acquisition of the edge brightness feature sequence of the first The brightness feature value of each pixel point in the edge brightness feature sequence of the first tourist dynamic area includes specific methods. The first frame intelligent signboard monitoring image is converted into a YUV color space, and a brightness value of each pixel point is obtained; on the boundary of the first tourist dynamic area, a clockwise rotation is performed for one round from a pixel point with the maximum brightness value to obtain all pixel points on the boundary of the first tourist dynamic area, and the all pixel points are recorded as an edge brightness sequence of the first tourist dynamic area; Using the least squares method to apply the first Curve fitting was performed on the brightness values ​​of all pixels in the edge brightness sequence of the dynamic tourist area to obtain the first... Edge brightness fitting curve of a dynamic tourist area; For the For any data point on the edge brightness fitting curve of a dynamic tourist area, if the slope of that data point is negative, then the brightness feature value of the corresponding pixel is recorded as... If the slope of any data point is 0, then the brightness feature value of the pixel corresponding to that data point is denoted as... If the slope of any data point is positive, then the brightness feature value of the pixel corresponding to that data point is denoted as... ; will the first The sequence consisting of the brightness feature values ​​of all pixels in the edge brightness sequence of the dynamic tourist area is used as the first... The edge brightness sequence of a visitor's dynamic area.

Images