WO2021120867A1

WO2021120867A1 - High toss act monitoring method and device, electronic device and storage medium

Info

Publication number: WO2021120867A1
Application number: PCT/CN2020/124096
Authority: WO
Inventors: 丁旭; 胡文泽
Original assignee: 深圳云天励飞技术股份有限公司
Priority date: 2019-12-19
Filing date: 2020-10-27
Publication date: 2021-06-24
Also published as: CN111079663B; CN111079663A

Abstract

The present invention provides a high toss act monitoring method and device, an electronic device and a storage medium, wherein the method comprises: acquiring video information of a current monitoring scene, and performing dynamic background modeling of the current monitoring scene through a normal distribution, so as to obtain a background image of the monitoring scene; determining whether a foreground image appears in the image information according to the background image; when the foreground image appears in the image information, continuing to obtain motion information of the foreground image, and calculating a motion trajectory of the foreground image according to the motion information; and determining whether the foreground image is a high toss act according to the motion trajectory of the foreground image. By modeling the background of the current monitoring scene, the background image is separated from the foreground image, and it can be determined separately whether the foreground image is a high toss act, which can determine the presence of the high toss act in real time, thereby improving the high toss act monitoring effect.

Description

High-altitude parabolic monitoring method, device, electronic equipment and storage medium

This application claims the priority of the Chinese patent application filed with the Chinese Patent Office on December 19, 2019, the application number is 201911320015.5, and the invention title is "High-altitude parabolic monitoring methods, devices, electronic equipment and storage media". The entire content of the Chinese patent application is approved. The reference is incorporated in this application.

Technical field

The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, electronic equipment and storage medium for monitoring high-altitude parabolas.

Background technique

With the development of real estate, the floors of newly-built residential quarters are getting higher and higher, and the problem of high-altitude parabolic is becoming more and more prominent. Most of the current residential communities are equipped with surveillance cameras to monitor the situation in the community. When a high-altitude parabolic event occurs, the relevant personnel can call and check according to the monitoring video collected by the high-altitude parabolic event, but the specific high-altitude parabolic situation requires Viewing manually frame by frame or through slow-motion cameras is not only a lot of work, but also easy to miss, and it is impossible to find high-altitude parabolic situations in time. Therefore, the existing high-altitude parabolic event monitoring effect is not good.

Summary of the invention

The embodiment of the present invention provides a method for monitoring high-altitude parabolic events, which can improve the monitoring effect of high-altitude parabolic events.

In the first aspect, an embodiment of the present invention provides a method for monitoring a parabola at high altitude, including:

Obtain the video information of the current monitoring scene, and perform dynamic background modeling of the current monitoring scene through normal distribution to obtain the background image of the monitoring scene, and the dynamic background modeling is to model each of the video information Frame images are all background modeling;

Judging whether a foreground image appears in the video information according to the background image;

When a foreground image appears in the video information, the motion information of the foreground image is continuously acquired, and the motion trajectory of the foreground image is calculated according to the motion information;

Based on the movement trajectory of the foreground image, it is determined whether the foreground image is a parabola at high altitude.

Optionally, the acquiring video information of the current monitoring scene and performing dynamic background modeling of the current monitoring scene through normal distribution to obtain a background image of the monitoring scene includes:

Acquiring continuous frame images in the real-time video information, wherein each pixel in the continuous frame image corresponds to K normal distributions, K is greater than 1, and the normal distribution includes a mean parameter, a variance parameter, and a weight parameter;

Match the pixel value of each pixel of the current frame image with the corresponding K normal distributions, and determine whether each pixel matches the normal distribution that meets the preset conditions;

If there are pixels whose pixel values match the M normal distributions that meet the preset conditions, then update the first parameter of the M normal distributions, and keep the parameters of the remaining KM normal distributions unchanged, where, M is greater than or equal to 1, and M is less than or equal to K;

If there is a pixel whose pixel value does not match the normal distribution that meets the preset condition, select the normal distribution with the largest mean distance from the K normal distributions corresponding to the pixel to perform weight assignment, and assign the value based on the weight Performing a second parameter update on the K normal distributions, where the mean distance is the difference between the pixel value of the pixel and the mean parameter in the normal distribution;

Based on the variance parameter and/or weight parameter of the normal distribution, select N normal distributions, and determine whether the corresponding pixel is a background pixel according to the N normal distributions, where N is greater than or equal to 1, and N Less than or equal to K;

Based on the background pixels, the frame background of the current frame image is constructed, and the frame background of the current frame image is updated to the background image of the monitoring scene.

Optionally, the determining whether a foreground image appears in the image information according to the background image includes:

Matching the pixel value of each pixel of the current frame image with the corresponding N normal distributions, and determining whether each pixel matches a normal distribution that meets a preset condition;

If there is a pixel that does not match the normal distribution that meets the preset condition, it is determined that the pixel that does not match the normal distribution that meets the preset condition is a foreground pixel;

Based on the foreground pixel points, the frame foreground of the current frame image is constructed, and the frame foreground of the frame image is updated to the foreground image of the monitoring scene.

Optionally, the judging whether the foreground image is a parabola at high altitude based on the movement trajectory of the foreground image includes:

Judging whether the motion trajectory of the foreground image conforms to a preset parabolic trajectory;

If the movement trajectory of the foreground image conforms to the preset parabolic trajectory, it is determined that the foreground image is a parabola at high altitude;

If the motion trajectory of the foreground image does not conform to the preset parabolic trajectory, it is determined that the foreground image is not a high-altitude parabola.

Constructing multiple horizontal detection lines in the background image;

Judging whether the number of intersections between the movement trajectory of the foreground image and the horizontal detection line is greater than a preset threshold for the number of intersections;

If the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is greater than a preset threshold of the number of intersections, determining that the foreground image is a parabola at high altitude;

If the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is less than a preset number of intersection thresholds, it is determined that the foreground image is not a parabola at high altitude.

Optionally, the method further includes:

If the foreground image is a high-altitude parabola, a high-altitude parabola prompt alarm is issued to the current monitoring scene and/or the management department.

Optionally, if the foreground image is a high-altitude parabola, sending a high-altitude parabola to the current monitoring scene and/or management department includes:

Extracting the foreground image;

Performing feature recognition on the foreground image to identify the category to which the foreground image belongs;

Match the corresponding high-altitude parabola level according to the category of the foreground image;

Based on the high-altitude parabolic level, a corresponding-level high-altitude parabolic alert is issued to the current monitoring scene and/or the management department.

In the second aspect, an embodiment of the present invention provides a high-altitude parabolic monitoring device, including:

The first acquisition module is used to acquire video information of the current monitoring scene, and perform dynamic background modeling of the current monitoring scene through normal distribution to obtain a background image of the monitoring scene, and the dynamic background modeling is a pair Background modeling is performed on each frame of image in the video information;

The first judgment module is configured to judge whether a foreground image appears in the video information according to the background image;

The second acquisition module is configured to continuously acquire the motion information of the foreground image when a foreground image appears in the video information, and calculate the motion trajectory of the foreground image according to the motion information;

The second judgment module is used for judging whether the foreground image is a parabola at high altitude based on the movement trajectory of the foreground image.

In a third aspect, an embodiment of the present invention provides an electronic device including: a memory, a processor, and a computer program stored on the memory and capable of running on the processor, and when the processor executes the computer program The steps in the high-altitude parabola monitoring method provided by the embodiment of the present invention are realized.

In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored on the computer-readable storage medium, and when the computer program is executed by a processor, the method for monitoring a parabola at high altitude provided by the embodiment of the invention is implemented Steps in.

In the embodiment of the present invention, real-time video information of the current monitoring scene is continuously acquired, and dynamic background modeling of the current monitoring scene is performed according to the real-time video information to obtain a background image of the monitoring scene; according to the background image , Determine whether a foreground image appears in the image information; when a foreground image appears in the image information, continue to acquire the motion information of the foreground image, and calculate the motion trajectory of the foreground image according to the motion information; based on The motion track of the foreground image determines whether the foreground image is a parabola at high altitude. By modeling the background of the current monitoring scene, the background image is separated from the foreground image, and whether the foreground image is a high-altitude parabola is judged separately without manual judgment. Since the background image is obtained by dynamic modeling, it can be judged in real time whether there is a high-altitude parabola. Circumstances, thereby improving the monitoring effect of high-altitude parabolic.

Description of the drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

FIG. 1 is a flowchart of a method for monitoring a parabola at high altitude according to an embodiment of the present invention;

2 is a flowchart of a dynamic background modeling method provided by an embodiment of the present invention;

Figure 3 is a schematic structural diagram of a high-altitude parabolic monitoring device provided by an embodiment of the present invention;

4 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

5 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

6 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

Figure 7 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

8 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

9 is a schematic structural diagram of another high-altitude parabolic monitoring device provided by an embodiment of the present invention;

FIG. 10 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.

Detailed ways

Please refer to FIG. 1. FIG. 1 is a flowchart of a method for monitoring a parabola at high altitude according to an embodiment of the present invention. As shown in FIG. 1, it includes the following steps:

101. Obtain video information of the current monitoring scene, and perform dynamic background modeling on the current monitoring scene through normal distribution to obtain a background image of the monitoring scene.

Among them, the foregoing current monitoring scene may be a residential building, a commercial building, or an office building that is being monitored by the camera. The above-mentioned camera monitoring range can be all floors of the building or floors above a certain number of floors, such as floors above the 4th floor. When installing the camera, it can be determined as needed, and the shooting angle of the camera can be adjusted so that the camera can Monitor the floors in the corresponding range.

The above-mentioned video information can be understood as a continuous image sequence captured by a camera. The above-mentioned video information may be target video information captured by the camera in real time, or target video information captured by the camera periodically, or target video information uploaded after the user retrieves the video information captured by the camera.

The aforementioned dynamic background modeling refers to the establishment of different background images based on different current frame images, that is, each frame image corresponds to a background image. The above-mentioned background image is embodied in the continuous image sequence as: in the continuous image sequence, the pixel value of the pixel as the background image does not change or the pixel value changes within a certain range. The above-mentioned dynamic background modeling relies on the association of pixels in different frames of images in a continuous image sequence. It can be understood that the change in the pixel value of a pixel in the continuous image sequence as a background pixel obeys a normal distribution. During the change of the pixel value of the background pixel, the pixel value of the background pixel is distributed in a range, and the range is determined by the mean value of the pixel value of the background pixel. It can be considered that the pixel value of the background pixel changes distribution On both sides of the mean of this change.

Specifically, please refer to Figure 2. Figure 2 is a flowchart of a dynamic background modeling method provided by an embodiment of the present invention. As shown in Figure 2, the above dynamic background modeling method includes the following steps:

201. Acquire continuous frame images in video information.

Among them, the above-mentioned continuous frame images refer to continuous images in a time series.

202. Construct K normal distributions corresponding to each pixel in the continuous frame image.

Where K is greater than 1, the normal distribution includes a mean parameter, a variance parameter, and a weight parameter.

In this step, the K normal distributions corresponding to each pixel of the first frame of image can be initialized to make the K normal distributions, wherein the K normal distributions can be performed by the following formula expression:

Wherein, the above P(x _j ) represents the normal distribution model of the j-th pixel, and the normal distribution model includes K normal distributions of the j-th pixel, and x _j,t represents the j-th pixel. The pixel value of the pixel, the above

Represents the weight parameter of the i-th normal distribution of the j-th pixel in the t-th frame image, the above

Represents the mean parameter of the i-th normal distribution of the j-th pixel in the t-th frame image, the above

Represents the variance parameter of the i-th normal distribution of the j-th pixel in the t-th frame image, the above η is the density function of the normal distribution, the above σ is the standard deviation, and

get.

In the process of initializing the K normal distributions corresponding to each pixel of the first frame of image, one normal distribution in each pixel of the first frame of image can be initialized, and the above-mentioned initialization can be The mean parameter in a normal distribution is assigned the pixel value of the corresponding pixel, and the weight parameter is assigned a value of 1, and the variance is 0 at this time. The mean parameter and weight parameter of the other normal distributions except the normal distribution are assigned as 0. For example, a pixel has 5 normal distributions, that is, K=5. Among these 5 normal distributions, select the mean parameter and weight parameter of one normal distribution for assignment, and the mean parameters of the remaining 4 normal distributions And the weight parameter is assigned the value 0. Since each pixel in the first frame of image is not dependent on the previous sequence, it is necessary to initialize the normal distribution of each pixel in the first frame of image.

Of course, in a possible embodiment, random initialization can be used to randomly assign all the normal distributions of each pixel in the first frame of image. It should be noted that in the random assignment process, all normal distributions The sum of the assigned weight parameters of the distribution needs to be equal to 1.

203. Match the pixel value of each pixel of the current frame of image with the corresponding K normal distributions.

Wherein, the above-mentioned current frame image is not the first frame image.

Take a pixel j in the current frame image as an example. Assuming that the current frame image is the t-th frame image, it can be understood that in the previous first frame image to the t-1th frame image, each pixel point The mean and variance of the corresponding pixel values are all known. For example, as of the t-1 frame image, in the K normal distributions of the pixel point j, the mean parameter is from the first frame image to the t-1 frame image The sum of all the pixel values of the middle pixel point j, and then dividing by the data of the frame image, that is, dividing by t-1, the mean value parameter is

The variance parameter of pixel j is the pixel value corresponding to pixel j in the t-1 frame image minus the mean parameter

Then square to get the variance parameter as

Thus, the K normal distributions of the pixel point j in the t-1 frame image can be obtained:

In the current t-th frame image, if the pixel point j is a background pixel point, the pixel value x _{j of} the pixel point j satisfies one or more of the above k normal distributions. This is because in the monitoring scene, the pixel value corresponding to the background pixel is usually unchanged or changes little. That is to say, the pixel value distribution corresponding to the background pixel can be predicted within a certain range of pixel value. Yes, because the pixel value corresponding to the background pixel is sampled for a long time, the amount of pixel value data corresponding to the background pixel is large enough, so that the pixel value corresponding to the background pixel obeys the normal distribution, that is, the data is concentrated in the mean parameter In the vicinity of, a random variable that follows a normal distribution has a high probability of being a value near the mean parameter, and a small probability of being a value far from the mean parameter. For example, consider the pixel value x _{j, t of pixel j in the t-} th frame of image as a random variable. If pixel j is a background pixel, then x _{j, t} is

Near value. Therefore, we can pass x _{j, t} and

To match the K normal distributions corresponding to pixel j.

204. Determine whether each pixel matches a normal distribution that meets a preset condition.

If there are M pixels with a pixel value matching the normal distribution that meet the preset condition, then go to step 205; if there are pixels with a pixel value that does not match the normal distribution that meets the preset condition, then go to step 206 .

The above preset conditions can be x _{j, t} and

The difference value satisfies the preset difference value threshold. The preset difference value threshold value may be determined according to the standard deviation in the normal distribution of _{x j, t-1, and the standard deviation is determined by}

get. Specifically, it can be judged that x _{j, t} and

Whether the difference of is less than the coefficient multiple of the standard deviation, such as judging x _{j, t} and

Whether the difference of is less than 1.5 times, 2.5 times and other standard deviations.

If x _{j, t} and

Whether the difference of is less than the coefficient multiple of the standard deviation, it indicates that the pixel j in the t-th frame obeys the normal distribution, that is, matches the normal distribution that satisfies the preset condition. The traversal judges whether to obey the K normal distributions of the pixel, so as to judge the number of the pixel j in the t-th frame that obeys the K normal distributions. Traverse each pixel in the t-th frame to determine the normal distribution of the matching of each pixel.

205. Update the first parameter of the M normal distributions, and keep the parameters of the remaining K-M normal distributions unchanged.

Among them, M is greater than or equal to 1, and M is less than or equal to K.

In this step, the M normal distributions that meet the preset conditions are updated. The above-mentioned first parameter update refers to the update of the mean parameter and the variance parameter in the normal distribution, for example,

Update to the new mean

will

Update to the new mean

Then the current normal distribution of pixel j in the t-th frame can be obtained. For a pixel, only M normal distributions meeting the preset conditions are updated, and the parameters of the remaining KM normal distributions are kept unchanged.

206. Select the normal distribution with the largest mean distance from the K normal distributions corresponding to the pixels to perform weight assignment, and perform the second parameter update on the K normal distributions based on the weight assignment.

The aforementioned mean distance is the difference between the pixel value of the pixel and the mean parameter in the normal distribution.

In this step, when a pixel does not match any one of the corresponding K normal distributions, x _{j, t} and

The second parameter update is performed on the normal distribution with the largest difference, and the remaining K-1 normal distributions remain unchanged.

The above-mentioned second parameter update refers to the update of the weight parameter in the normal distribution, for example,

Update to the new mean

Specifically, after the update, it is determined again whether the pixel matches the new normal distribution. The weight parameter in the normal distribution can be updated by the following formula:

ω _i,t =(1-a)·ω _i,t-1 +a·M _i,t

Among them, the above a is the learning rate of the algorithm, and the above Mi _,t is the updated matching result. If the pixel can match the new normal distribution after the weight is updated _{, the} value of Mi,t is 1, If the pixel still cannot match the new normal distribution after the weight is updated _{, the} value of Mi,t is 0.

Since the background pixel is subject to a normal distribution, if the above-mentioned pixel can match the new normal distribution, it means that the pixel is a background point, if it cannot match the new normal distribution, it means This pixel is the former scenic spot. Specifically, according to the above weight parameter update formula, it can be known that if the pixel can match the new normal distribution, the weight parameter in the final normal distribution is increased. If the pixel cannot match the new normal distribution Normal distribution, the weight parameter in the final normal distribution is reduced.

207. Select N normal distributions based on the variance parameter and/or weight parameter of the normal distribution, and determine whether the corresponding pixel is a background pixel according to the N normal distributions.

Among them, N is that the ratio of the weight parameter to the variance parameter in K normal distributions is the largest than N normal distributions, N is greater than or equal to 1, and N is less than or equal to K.

The aforementioned variance parameter characterizes the degree of dispersion of the data distribution. The greater the variance, the greater the degree of dispersion, the smaller the variance, and the smaller the degree of dispersion. The smaller the degree of dispersion, it means that the data is concentrated in a small area, and the characteristics are more obvious. Therefore, a background pixel can select N normal distributions with the smallest variance parameter among K normal distributions as the best description of the background.

The above weight parameters represent the degree of data support of each normal distribution. When the background continues to remain unchanged, the distribution data corresponding to the background pixels in the background will continue to accumulate, and the supported normal distribution weight ratio will be higher. The higher the probability of falling into the normal distribution. Therefore, a background pixel can select the N normal distributions with the largest weight parameters among the K normal distributions as the best description of the background. It should be noted that in the K normal distributions corresponding to a pixel, the sum of the K weight parameters is 1.

As an embodiment of the present invention, it can also be selected based on the ratio of the weight parameter to the variance parameter. A background pixel can select the N normal distributions with the largest ratio of the weight parameter to the variance parameter among K broken distributions as the background. The best description.

After determining the N normal distributions corresponding to each pixel, each pixel in the current t-th frame image is matched with the corresponding N normal distributions again. When at least one normal distribution is matched, the corresponding Pixels are background pixels, and go to step 208. If it fails to match any normal distribution, it means that the pixel is a foreground pixel, and then go to step 209.

208. Based on the background pixels, construct the frame background of the current frame image, and update the frame background of the current frame image to the background image of the monitoring scene.

When it is determined that the pixel is the background pixel of the current frame image, the background pixel of the current frame image can be masked to distinguish it from the foreground part, and the frame background corresponding to the current frame image is obtained, and the frame background is updated to the video Corresponding frame images in the information, so as to obtain the background image of each frame of the monitoring scene.

In the embodiment of the present invention, the background pixel of the background image is judged by normal distribution, and the past data distribution of a pixel can be used to predict whether the pixel is a background pixel, and the accuracy of dynamic background modeling can be improved.

102. Determine whether a foreground image appears in the video information according to the background image.

In this step, the pixel value of each pixel of the current frame image is matched with the corresponding N normal distributions, and it is judged whether each pixel matches the normal distribution that meets the preset conditions; if there is a pixel matching If the normal distribution does not meet the preset condition, it means that the pixel does not obey the normal distribution of the background pixel, and then it is determined that the pixel that does not match the normal distribution that meets the preset condition is the foreground pixel. Through the foreground pixels, it can be judged whether there is a foreground image in the video information. Specifically, it can be judged whether there are foreground pixels in the frame image of the foreground image, thereby judging whether there is a foreground frame image in the video information, and then judging whether the video information is in the foreground image. The foreground image appears.

209. Based on the foreground pixels, construct the frame foreground of the current frame image, and update the frame foreground of the frame image to the foreground image of the monitoring scene.

In this step, the pixel value of each pixel of the current frame image is matched with the corresponding N normal distributions, and it is judged whether each pixel matches the normal distribution that meets the preset conditions; if there is a pixel matching If the normal distribution does not meet the preset conditions, it is judged that the pixels that do not match the normal distribution that meets the preset conditions are foreground pixels; based on the foreground pixels, the frame foreground of the current frame image is constructed, and the frame image The frame foreground is updated to the foreground image of the monitoring scene. In the same way, when the pixel is determined to be the foreground pixel, the foreground pixel of the current frame image can be masked to distinguish it from the background part to obtain the frame foreground corresponding to the current frame image, and update the frame foreground to the video information In the corresponding frame image, the foreground image of each frame of the monitoring scene is thus obtained.

103. When a foreground image appears in the video information, the motion information of the foreground image is continuously acquired, and the motion trajectory of the foreground image is calculated according to the motion information.

In this step, when a foreground image appears in the video information, the foreground image can be continuously tracked by the tracking algorithm to obtain the displacement data of the foreground image in the background image in the sequence corresponding to the continuous frame image, and calculate according to the displacement data The trajectory of the foreground image. Wherein, the aforementioned displacement data refers to the displacement data of the pixel point coordinates of the foreground image in the frame image.

104. Based on the movement trajectory of the foreground image, determine whether the foreground image is a parabola at high altitude.

In this step, the movement trajectory of the foreground image can be compared with a preset parabolic trajectory, and if the movement trajectory of the foreground image conforms to the preset parabolic trajectory, it can be determined that the foreground image is a parabola at high altitude. If the motion trajectory of the foreground image does not conform to the preset parabolic trajectory, it is determined that the foreground image is not a parabola at high altitude. The aforementioned preset motion trajectory may be a downward straight line or a parabolic trajectory.

Optionally, the motion trajectory can be preset based on the position of the camera. For example, when the camera is shooting the building to be monitored, the preset motion trajectory can be a downward linear trajectory, a parabolic trajectory toward the lower left or right Three types of motion trajectories such as the parabolic trajectory below. When the camera is shooting on the side of the building to be monitored, the aforementioned preset motion trajectory can be a downward linear trajectory, or a characteristic line trajectory that is biased to one side, etc. In this case, only two types can be preset Movement trajectory. It should be noted that the above-mentioned building to be monitored may also be referred to as a monitoring scene.

Optionally, in this step, it is also possible to determine whether the foreground image is a high-altitude parabola by constructing a horizontal detection line in the background image. Specifically, multiple horizontal detection lines are constructed in the background image; it is determined whether the number of intersections between the movement trajectory of the foreground image and the horizontal detection line is greater than the preset number of intersection thresholds; if the number of intersections between the movement trajectory of the foreground image and the horizontal detection line is greater than The preset number of intersection thresholds determines that the foreground image is a parabola at high altitude; if the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is less than the threshold number of intersections, it is determined that the foreground image is not a parabola at high altitude. Furthermore, the above-mentioned horizontal detection line can be constructed according to the floor. For example, each floor can construct a horizontal detection line, and the horizontal detection line can be constructed according to the upper window edge or the lower window edge outside each floor. In this case, the height of the high-altitude parabola can be determined according to the number of intersections between the movement trajectory of the foreground image and the horizontal detection line, that is, which floor is performing the high-altitude parabolic behavior. In addition, according to the number of intersections of the horizontal detection line, different high-altitude parabolic levels can be set.

Optionally, after step 104, when it is determined that the foreground image is a high-altitude parabola, a warning of high-altitude parabola can be automatically issued to the current monitoring scene and/or the management department.

The above-mentioned current monitoring scene refers to the scene where the corresponding camera is deployed, such as unit C in building B in residential area A. When the camera of unit C of building B in residential area A detects high-altitude parabolic objects, a hazard warning will be issued at unit C of building B in residential area A to alert people in the vicinity of unit C in building B in residential area A.

The above-mentioned management department may be the property management department or the city management department or other organizations with management authority, and other organizations with management authority, such as the owners' committee, the garden club, etc. In a possible implementation manner, the reminder alarm sent to the management department also includes video information of the current monitoring scene. The video information includes continuous frame images of high-altitude parabolic events. The alarm can be notified through various contact methods. The information is sent to the contact terminal of the management department or related personnel, such as mail, mobile APP or WeChat official account push, etc.

Optionally, when the foreground image is detected, the foreground image can also be extracted and feature recognition is performed on the foreground image to identify the category to which the foreground image belongs; according to the category of the foreground image, the corresponding high-altitude parabolic level is matched; Based on the matched high-altitude parabolic level, a corresponding level of high-altitude parabolic alert is issued to the current monitoring scene and/or the management department. For example, when it is recognized that the foreground image is a paper sheet or plastic bag, it can be judged that the high-altitude parabola level of the foreground image is low; when the foreground image is recognized as a flower pot or a mobile phone, it can be judged that the high-altitude parabola level of the foreground image is high . The above-mentioned high-altitude parabolic level can be positively correlated with the hazard level. Different high-altitude parabolic reminders can be set for different high-altitude parabolic levels.

In the embodiment of the present invention, real-time video information of the current monitoring scene is continuously acquired, and dynamic background modeling of the current monitoring scene is performed according to the real-time video information to obtain the background image of the monitoring scene; according to the background Image, judging whether a foreground image appears in the image information; when a foreground image appears in the image information, continuously acquiring the motion information of the foreground image, and calculating the motion trajectory of the foreground image according to the motion information; Based on the movement trajectory of the foreground image, it is determined whether the foreground image is a parabola at high altitude. By modeling the background of the current monitoring scene, the background image is separated from the foreground image, and whether the foreground image is a high-altitude parabola is judged separately without manual judgment. Since the background image is obtained by dynamic modeling, it can be judged in real time whether there is a high-altitude parabola. Circumstances, thereby improving the monitoring effect of high-altitude parabolic.

It should be noted that the high-altitude parabolic monitoring method provided by the embodiment of the present invention can be applied to mobile terminals, monitors, computers, servers and other equipment that need to monitor high-altitude parabolic behavior.

Please refer to FIG. 3, which is a schematic structural diagram of a high-altitude parabolic monitoring device provided by an embodiment of the present invention. As shown in FIG. 3, the device includes:

The first acquisition module 301 is configured to acquire video information of the current monitoring scene, and perform dynamic background modeling of the current monitoring scene through a normal distribution to obtain a background image of the monitoring scene, and the dynamic background modeling is Perform background modeling on each frame of image in the video information;

The first determining module 302 is configured to determine whether a foreground image appears in the image information according to the background image;

The second acquisition module 303 is configured to continuously acquire the motion information of the foreground image when a foreground image appears in the video information, and calculate the motion trajectory of the foreground image according to the motion information;

The second judgment module 304 is configured to judge whether the foreground image is a parabola at high altitude based on the movement trajectory of the foreground image.

Optionally, the above-mentioned first acquisition module 301 and the second acquisition module 302 may be the same acquisition module, and the above-mentioned first acquisition module 301 and the second acquisition module 302 may also be integrated in the same acquisition module.

Optionally, as shown in FIG. 4, the first obtaining module 301 includes:

The acquiring unit 3011 is configured to acquire continuous frame images in the video information, wherein each pixel in the continuous frame image corresponds to K normal distributions, K is greater than 1, and the normal distribution includes a mean parameter and a variance Parameters and weight parameters;

The first determining unit 3012 is configured to match the pixel value of each pixel of the current frame image with the corresponding K normal distributions, and determine whether each pixel matches a normal distribution that meets a preset condition;

The first update unit 3013 is configured to update the first parameter of the M normal distributions if there are pixel values matching the M normal distributions that meet the preset condition, and keep the remaining KM normal distributions The parameters of the distribution remain unchanged, where M is greater than or equal to 1, and M is less than or equal to K;

The second update unit 3014 is configured to select the normal distribution with the largest mean distance from the K normal distributions corresponding to the pixel points if there are pixels whose pixel values do not match the normal distribution that meets the preset conditions. Weight assignment, performing a second parameter update on the K normal distributions based on the weight assignment, and the mean distance is the difference between the pixel value of the pixel and the mean parameter in the normal distribution;

The second determining unit 3015 is configured to select N normal distributions based on the variance parameter and/or weight parameter of the normal distribution, and determine whether the corresponding pixel is a background pixel according to the N normal distributions, wherein , N is greater than or equal to 1, and N is less than or equal to K;

The first construction unit 3016 is configured to construct the frame background of the current frame image based on the background pixels, and update the frame background of the current frame image to the background image of the monitoring scene.

Optionally, as shown in FIG. 5, the first judgment module 302 includes:

The third determining unit 3021 is configured to match the pixel value of each pixel of the current frame image with the corresponding N normal distributions, and determine whether each pixel matches a normal distribution that meets a preset condition;

The fourth determining unit 3022 is configured to determine that if there is a pixel that does not match the normal distribution that meets the preset condition, determine that the pixel that does not match the normal distribution that meets the preset condition is a foreground pixel;

The second construction unit 3023 is configured to construct the frame foreground of the current frame image based on the foreground pixels, and update the frame foreground of the frame image to the foreground image of the monitoring scene.

Optionally, as shown in FIG. 6, the second judgment module 304 includes:

The fifth judging unit 3041 is used to judge whether the motion trajectory of the foreground image conforms to a preset parabolic trajectory;

The sixth determining unit 3042 is configured to determine that the foreground image is a parabola at high altitude if the motion trajectory of the foreground image conforms to the preset parabolic trajectory;

The seventh determining unit 3043 is configured to determine that the foreground image is not a high-altitude parabola if the motion trajectory of the foreground image does not conform to the preset parabolic trajectory.

Optionally, as shown in FIG. 7, the second judgment module 304 includes:

The constructing unit 3044 is configured to construct multiple horizontal detection lines in the background image;

An eighth judging unit 3045, configured to judge whether the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is greater than a preset threshold for the number of intersections;

A ninth determining unit 3046, configured to determine that the foreground image is a parabola at high altitude if the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is greater than a preset number of intersection thresholds;

The tenth determining unit 3047 is configured to determine that the foreground image is not a parabola at high altitude if the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is less than a preset number of intersection thresholds.

Optionally, as shown in FIG. 8, the device further includes:

The prompting module 305 is configured to, if the foreground image is a high-altitude parabola, send a high-altitude parabola to the current monitoring scene and/or the management department.

Optionally, as shown in FIG. 9, the prompt module 305 includes:

The extraction unit 3051 is used to extract the foreground image;

The recognition unit 3052 is configured to perform feature recognition on the foreground image to recognize the category to which the foreground image belongs;

The matching unit 3053 is configured to match the corresponding high-altitude parabolic level according to the category of the foreground image;

The prompting unit 3054 is configured to issue a corresponding level of high-altitude parabolic prompt alarm to the current monitoring scene and/or management department based on the high-altitude parabolic level.

It should be noted that the high-altitude parabolic monitoring device provided in the embodiment of the present invention can be applied to mobile terminals, monitors, computers, servers and other equipment that need to monitor high-altitude parabolic behavior.

The high-altitude parabolic monitoring device provided by the embodiment of the present invention can realize the various processes realized by the high-altitude parabolic monitoring method in the foregoing method embodiment, and can achieve the same beneficial effects. To avoid repetition, I won’t repeat them here.

Referring to FIG. 10, FIG. 10 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. As shown in FIG. 10, it includes: a memory 1002, a processor 1001, and a memory 1002 that is stored on the memory 1002 and can be stored in the processor. A computer program running on 1001, where:

The processor 1001 is configured to call a computer program stored in the memory 1002, and execute the following steps:

Optionally, the acquiring video information of the current monitoring scene performed by the processor 1001 and performing dynamic background modeling of the current monitoring scene through normal distribution to obtain a background image of the monitoring scene includes:

Acquiring continuous frame images in the video information, wherein each pixel in the continuous frame image corresponds to K normal distributions, K is greater than 1, and the normal distribution includes a mean parameter, a variance parameter, and a weight parameter;

Matching the pixel value of each pixel of the current frame image with the corresponding K normal distributions, and determining whether each pixel matches a normal distribution that meets a preset condition;

Optionally, the determining whether a foreground image appears in the video information according to the background image performed by the processor 1001 includes:

Optionally, the determination by the processor 1001 to determine whether the foreground image is a high-altitude parabola based on the movement trajectory of the foreground image includes:

If the motion trajectory of the foreground image conforms to the preset parabolic trajectory, determining that the foreground image is a high-altitude parabola;

Constructing multiple horizontal detection lines in the background image;

Optionally, the processor 1001 further executes including:

Optionally, if the foreground image is a high-altitude parabola executed by the processor 1001, issuing a high-altitude parabola to the current monitoring scene and/or management department includes:

Extracting the foreground image;

It should be noted that the electronic equipment provided by the embodiments of the present invention can be applied to mobile terminals, monitors, computers, servers and other equipment that need to monitor parabolic behavior at high altitude.

The electronic device provided by the embodiment of the present invention can implement each process implemented by the method for monitoring parabola at high altitude in the foregoing method embodiment, and can achieve the same beneficial effects. To avoid repetition, details are not repeated here.

The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, each process of the method for monitoring a parabolic at high altitude provided by the embodiment of the present invention is implemented, and To achieve the same technical effect, in order to avoid repetition, I will not repeat them here.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through a computer program. The program can be stored in a computer readable storage medium, and the program can be stored in a computer readable storage medium. During execution, it may include the procedures of the above-mentioned method embodiments. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM for short), etc.

Claims

A high-altitude parabolic monitoring method is characterized in that it comprises the following steps:

Obtain the video information of the current monitoring scene, and perform dynamic background modeling of the current monitoring scene through normal distribution to obtain the background image of the monitoring scene, and the dynamic background modeling is to model each of the video information Frame images are all background modeling;

Judging whether a foreground image appears in the video information according to the background image;

When a foreground image appears in the video information, the motion information of the foreground image is continuously acquired, and the motion trajectory of the foreground image is calculated according to the motion information;

Based on the movement trajectory of the foreground image, it is determined whether the foreground image is a parabola at high altitude.
The method according to claim 1, wherein said acquiring video information of the current monitoring scene, and performing dynamic background modeling of the current monitoring scene through a normal distribution to obtain a background image of the monitoring scene, include:

Acquiring continuous frame images in the video information, wherein each pixel in the continuous frame image corresponds to K normal distributions, K is greater than 1, and the normal distribution includes a mean parameter, a variance parameter, and a weight parameter;

Match the pixel value of each pixel of the current frame image with the corresponding K normal distributions, and determine whether each pixel matches the normal distribution that meets the preset conditions;

If there are pixels whose pixel values match the M normal distributions that meet the preset conditions, then update the first parameter of the M normal distributions, and keep the parameters of the remaining KM normal distributions unchanged, where, M is greater than or equal to 1, and M is less than or equal to K;

If there is a pixel whose pixel value does not match the normal distribution that meets the preset condition, select the normal distribution with the largest mean distance from the K normal distributions corresponding to the pixel to perform weight assignment, and assign the value based on the weight Performing a second parameter update on the K normal distributions, where the mean distance is the difference between the pixel value of the pixel and the mean parameter in the normal distribution;

Based on the variance parameter and/or weight parameter of the normal distribution, select N normal distributions, and determine whether the corresponding pixel is a background pixel according to the N normal distributions, where N is greater than or equal to 1, and N Less than or equal to K;

Based on the background pixels, the frame background of the current frame image is constructed, and the frame background of the current frame image is updated to the background image of the monitoring scene.
The method according to claim 2, wherein the determining whether a foreground image appears in the video information according to the background image comprises:

Matching the pixel value of each pixel of the current frame image with the corresponding N normal distributions, and determining whether each pixel matches a normal distribution that meets a preset condition;

If there is a pixel that does not match the normal distribution that meets the preset condition, it is determined that the pixel that does not match the normal distribution that meets the preset condition is a foreground pixel;

Based on the foreground pixel points, the frame foreground of the current frame image is constructed, and the frame foreground of the frame image is updated to the foreground image of the monitoring scene.
The method according to claim 1, wherein the judging whether the foreground image is a parabola at high altitude based on the movement trajectory of the foreground image comprises:

Judging whether the motion trajectory of the foreground image conforms to a preset parabolic trajectory;

If the motion trajectory of the foreground image conforms to the preset parabolic trajectory, determining that the foreground image is a high-altitude parabola;

If the motion trajectory of the foreground image does not conform to the preset parabolic trajectory, it is determined that the foreground image is not a high-altitude parabola.
The method according to claim 1, wherein the judging whether the foreground image is a parabola at high altitude based on the movement trajectory of the foreground image comprises:

Constructing multiple horizontal detection lines in the background image;

Judging whether the number of intersections between the movement trajectory of the foreground image and the horizontal detection line is greater than a preset threshold for the number of intersections;

If the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is greater than a preset threshold of the number of intersections, determining that the foreground image is a parabola at high altitude;

If the number of intersections between the motion trajectory of the foreground image and the horizontal detection line is less than a preset number of intersection thresholds, it is determined that the foreground image is not a parabola at high altitude.
The method of claim 1, wherein the method further comprises:

If the foreground image is a high-altitude parabola, a high-altitude parabola prompt alarm is issued to the current monitoring scene and/or the management department.
7. The method according to claim 6, wherein if the foreground image is a parabola at high altitude, sending a high-altitude parabolic alert to the current monitoring scene and/or management department comprises:

Extracting the foreground image;

Performing feature recognition on the foreground image to identify the category to which the foreground image belongs;

Match the corresponding high-altitude parabola level according to the category of the foreground image;

Based on the high-altitude parabolic level, a corresponding level of high-altitude parabolic alert is issued to the current monitoring scene and/or the management department.
A high-altitude parabolic monitoring device, characterized in that the device comprises:

The first acquisition module is used to acquire video information of the current monitoring scene, and perform dynamic background modeling of the current monitoring scene through normal distribution to obtain a background image of the monitoring scene, and the dynamic background modeling is a pair Background modeling is performed on each frame of image in the video information;

The first judgment module is configured to judge whether a foreground image appears in the image information according to the background image;

The second acquisition module is configured to continuously acquire the motion information of the foreground image when a foreground image appears in the video information, and calculate the motion trajectory of the foreground image according to the motion information;

The second judgment module is used for judging whether the foreground image is a parabola at high altitude based on the movement track of the foreground image.
An electronic device, characterized by comprising: a memory, a processor, and a computer program stored on the memory and capable of running on the processor. The processor executes the computer program as claimed in claim 1. To the step in the method for monitoring a parabola at high altitude as described in any one of 7.
A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the high-altitude parabola according to any one of claims 1 to 7 is realized. The steps in the monitoring method.