A method for crowd detection

Abstract

The present invention discloses a crowd detection method based on location services. By performing dimensionality reduction decomposition processing on coordinates, crowded people can be quickly and efficiently found, effectively reducing the detection calculation amount. At the same time, this method is simple and easy to integrate and can be used as a pre-judgment means to assist other algorithms. For example, the camera to be called can be locked first through this method, and then the crowd detection method based on image / video processing can be used for further refined judgment.

Description

Technical Field

[0001] The present invention relates to the field of location-based services, and in particular to a crowd detection method based on location services. Background Art

[0002] With the rapid development of information technology, a large amount of business data is constantly migrating to the network environment. Many public security and judicial units are increasingly relying on information technology to support the operation of their businesses, and the concept of "smart justice" has been proposed. Among them, due to the particularity of the industry and the object, the risk prediction and response capabilities of abnormal behaviors are very high. By introducing advanced technologies such as high-precision location services, video processing, and AI recognition, abnormal behaviors can be discovered in time, and linked with automated supervision facilities such as video and access control, emergency response speed can be effectively improved, and passive supervision can be changed to active supervision.

[0003] Mob behavior refers to an abnormal behavior in which multiple people participate in certain activities at the same time and place. It has high potential risks and hazards, so it needs to be paid special attention. At present, the detection of mob behavior usually uses image / video processing technology, but the camera has high requirements for the environment, it is easy to have blind spots, and the recognition algorithm is greatly affected by light intensity or occlusion, and the stability is poor. If high-precision location information can be introduced for early prediction, both the detection accuracy and the algorithm calculation amount will be greatly improved.

[0004] If we start from the location information, the simplest and most direct method is to first determine the number of people in the crowd around each person based on the X and Y coordinates, and then group the people in the crowd around the candidate who exceeds the lower limit of the crowd according to the permutations and combinations, and then determine whether the other people meet the crowd distance in pairs before they are considered to meet the crowd behavior. One of the main problems of this direct method is that the amount of calculation increases sharply as the total number of people increases. Summary of the invention

[0005] The invention discloses a crowd detection method based on location service, which can quickly and efficiently find out the crowd by dimensionality reduction decomposition processing of coordinates.

[0006] An embodiment of the present invention provides a method for detecting crowd gathering, including:

[0007] Obtain the horizontal coordinates of all personnel in the area based on the positioning terminals worn by personnel;

[0008] Calculate the projection coordinates of the horizontal coordinates of all persons in the area on the first linear dimension, and screen candidate persons who are likely to gather in a crowd according to the projection coordinates of the first linear dimension;

[0009] Calculate the projection coordinates of the candidate personnel with the possibility of gathering in the second straight line dimension perpendicular to the first straight line dimension, and determine the personnel with gathering behavior according to the projection coordinates in the second straight line dimension.

[0010] Optionally, the screening of candidate personnel with the possibility of gathering according to the projection coordinates in the first straight line dimension includes: sorting all the personnel in the area according to the size of the projection coordinates in the first straight line dimension;

[0011] Traverse each sorted personnel, and count the number of people Pi falling within the range [Xi, Xi+L], where Xi is the projection coordinate of the currently traversed personnel in the first straight line dimension, and L is the set side length of the gathering range;

[0012] If Pi exceeds the threshold of the limited number of people for gathering, it is considered that the Pi people falling within the range [Xi, Xi+L] are the candidate personnel with the possibility of gathering, otherwise continue to the next personnel.

[0013] Optionally, determining the personnel with gathering behavior according to the projection coordinates in the second straight line dimension includes:

[0014] Sort the candidate personnel with the possibility of gathering according to the size of the projection coordinates in the second straight line dimension;

[0015] Traverse each sorted candidate personnel, and count the number of people Qj falling within the range [Yj, Yj+L], where Yj is the projection coordinate of the currently traversed candidate personnel in the second straight line dimension;

[0016] If Qj exceeds the threshold of the limited number of people for gathering, it is considered that the Qj candidate personnel falling within the range [Yj, Yj+L] are the personnel with gathering behavior, otherwise continue to the next candidate personnel.

[0017] Optionally, before obtaining the horizontal coordinates of all the personnel in the area, it further includes: distinguishing the personnel on different floors indoors according to the elevation coordinates of the positioning terminals worn by the personnel.

[0018] The above technical solutions of the present invention have at least the following beneficial effects:

[0019] (1) Through the dimensionality reduction decomposition processing of the position, compared with the method of directly calculating the pairwise Euclidean distances and then counting the number of people within the gathering range, the algorithm is faster and more efficient, and can effectively reduce the calculation amount;

[0020] (2) This method is simple and easy to integrate, and can be used as a pre-judgment means to assist other algorithms. For example, the camera to be called can be locked first through this method, and then the gathering detection method based on image / video processing can be used for further refined judgment. Description of the Drawings

[0021] Figure 1 This is a schematic flowchart of a crowd detection method provided in the first embodiment of the present invention. Specific embodiments

[0022] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0023] Figure 1 This is a schematic flowchart of a crowd detection method provided in the first embodiment of the present invention. As Figure 1 shown, it includes the following steps:

[0024] S101: Obtain the horizontal coordinates of all personnel in the area according to the positioning terminals worn by the personnel;

[0025] S102: Calculate the projection coordinates of the horizontal coordinates of all personnel in the area in the first linear dimension, and screen out candidate personnel with the possibility of crowd gathering according to the projection coordinates in the first linear dimension;

[0026] S103: Calculate the projection coordinates of the candidate personnel with the possibility of crowd gathering in the second linear dimension perpendicular to the first linear dimension, and determine the personnel with crowd gathering behavior according to the projection coordinates in the second linear dimension.

[0027] Optionally, the screening of candidate personnel with the possibility of crowd gathering according to the projection coordinates in the first linear dimension includes: sorting all personnel in the area according to the size of the projection coordinates in the first linear dimension;

[0028] Traverse each sorted personnel, and count the number of people Pi falling within the range [Xi, Xi + L], where Xi is the projection coordinate of the current traversed personnel in the first linear dimension, and L is the set side length of the crowd gathering range;

[0029] If Pi exceeds the crowd gathering limit number threshold, it is considered that the Pi personnel falling within the range [Xi, Xi + L] are the candidate personnel with the possibility of crowd gathering, otherwise continue with the next personnel.

[0030] Optionally, determining the personnel with crowd gathering behavior according to the projection coordinates in the second linear dimension includes:

[0031] Sorting the candidate personnel with the possibility of crowd gathering according to the size of the projection coordinates in the second linear dimension;

[0032] Traverse each sorted candidate personnel, and count the number of people Qj falling within the range [Yj, Yj + L], where Yj is the projection coordinate of the current traversed candidate personnel in the second linear dimension;

[0033] If Qj exceeds the threshold of the crowd gathering limit number, it is considered that the Qj candidate persons falling within the range of [Yj, Yj+L] are the persons with crowd gathering behavior, otherwise, continue to the next candidate person.

[0034] Optionally, before obtaining the horizontal coordinates of all persons in the area, it further includes: distinguishing the persons on different floors indoors according to the elevation coordinates of the positioning terminals worn by the persons.

[0035] In this embodiment, the positioning terminal is a terminal device with one or more positioning modules such as GNSS, UWB, WiFi, and Bluetooth, and can be a general device such as a smart phone, a tablet computer, or a personal digital assistant, or can also be a tag, a wrist watch, or other wearable dedicated devices. It should be noted that the specific type of the positioning terminal is not limited in the embodiments of the present invention. In this embodiment, if the horizontal coordinates adopt a non-rectangular coordinate system, such as the longitude and latitude of a geodetic coordinate system, it needs to be converted to a rectangular coordinate system first. The first straight line dimension and the second straight line dimension preferably adopt two number axes of the rectangular coordinate system to simplify the process of calculating the projection coordinates.

[0036] In this embodiment, it is assumed that the total number of all persons in the area is N. If the direct method is used to calculate, first judge the number P of people in the crowd gathering range around each person according to the X and Y coordinates, and the time complexity is 2N×(N-1); then, group the number of people around the candidate persons exceeding the lower limit K of the crowd gathering number by C P (K-1) Grouping, and sequentially judge whether the remaining (K-1) people in each group satisfy the crowd gathering range in pairs before it is considered that the crowd gathering behavior is satisfied. If the method of this embodiment is used, sort all the persons in the area according to the size of the projection coordinates of the first straight line dimension, and an average of N×log(N) comparison judgments are required; if in the worst case all persons are candidate persons with the possibility of crowd gathering, then sort the candidate persons with the possibility of crowd gathering according to the size of the projection coordinates of the second straight line dimension, and an average of N×log(N) more comparison judgments are required, with a total of 2N×log(N) comparison judgments. When the value of N is large, the method of this embodiment has obvious advantages in terms of calculation amount.

[0037] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A crowd detection method, comprising: Obtaining the horizontal coordinates of all personnel in the area according to the positioning terminals worn by the personnel; Calculating the projected coordinates of the horizontal coordinates of all personnel in the area in the first linear dimension, and screening candidate personnel with the possibility of crowding according to the projected coordinates in the first linear dimension, including: Sorting all personnel in the area according to the magnitude of the projected coordinates in the first linear dimension; traversing each sorted personnel, and counting the number of people Pi falling within the range [Xi, Xi+L], where Xi is the projected coordinate of the current traversed personnel in the first linear dimension, and L is the set side length of the crowding range; if Pi exceeds the crowding limit number threshold, it is considered that the Pi personnel falling within the range [Xi, Xi+L] are the candidate personnel with the possibility of crowding, otherwise continue with the next personnel; Calculating the projected coordinates of the candidate personnel with the possibility of crowding in the second linear dimension perpendicular to the first linear dimension, and determining the personnel with crowding behavior according to the projected coordinates in the second linear dimension, including: Sorting the candidate personnel with the possibility of crowding according to the magnitude of the projected coordinates in the second linear dimension; traversing each sorted candidate personnel, and counting the number of people Qj falling within the range [Yj, Yj+L], where Yj is the projected coordinate of the current traversed candidate personnel in the second linear dimension; if Qj exceeds the crowding limit number threshold, it is considered that the Qj candidate personnel falling within the range [Yj, Yj+L] are the personnel with crowding behavior, otherwise continue with the next candidate personnel.

Images