Surveillance system

Abstract

To provide a monitoring system which can be operated without stopping an object detection function even when work in a monitoring area is scheduled.SOLUTION: The monitoring system of the present example includes a 3D sensor 20 for detecting an object present in a monitoring area and a server 40 for detecting the object in the monitoring area based on point cloud data acquired by the 3D sensor 20, and a specific person (for example, a worker) whose action in the monitoring area is permitted is requested to wear a predetermined reflector, and the server 40 determines whether or not the object in the monitoring area is the specific person based on a reflection intensity of a point cloud included in the point cloud data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a surveillance system that uses a 3D sensor to detect objects within a surveillance area. [Background technology]

[0002] Conventionally, systems that monitor a monitoring area using video captured by cameras have been in practical use. In such monitoring systems, sensors such as fence sensors and infrared sensors, as well as image processing detection technology, are used as items for object detection functions to detect objects that have entered the monitoring area. In recent years, 3D sensors have also been used in conjunction with these systems to confirm the distance to an object. 3D sensors use lasers or ultrasound to obtain three-dimensional information about a subject, and are characterized by relatively high measurement accuracy. Using 3D sensors, it is possible to obtain point cloud data that includes distance information for each of multiple points on the subject.

[0003] Prior art in the technical field related to the present invention includes the following: For example, Patent Document 1 discloses an invention in which, in an obstacle detection system including a plurality of radar devices installed at predetermined intervals along a travel route and each scanning a predetermined section of the travel route, and a determination device that determines the presence or absence of an obstacle on the travel route based on scan data from the radar devices, the determination device determines the presence or absence of an obstacle based on changes in the scan data obtained at predetermined scan intervals. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-65721 Summary of the Invention [Problem to be solved by the invention]

[0005] The object detection function in conventional surveillance systems is only used to detect people or vehicles that enter the surveillance area, and the identity of the object is generally determined by a security officer who separately checks the video from the visible light camera. However, if there is a possibility that a worker may enter the surveillance area for scheduled work, it is necessary to take measures such as temporarily suspending the object detection function.

[0006] Without such measures, even workers who are authorized to be in the monitored area will be detected, and the monitors will have to check the footage every time, which will lead to inefficient operations. Also, if the task of monitors checking the footage every time a worker is detected becomes a regular occurrence, it is possible that an intruder may be mistaken for a worker when in fact there is an intruder.

[0007] While it is possible to disable the object detection function while scheduled work is being carried out, this poses a security problem as it would not be possible to detect an intruder if one were to occur at that time. Therefore, it is desirable to operate the system so that the object detection function is not disabled even when scheduled work is being carried out, but so that workers are not detected as intruders.

[0008] The present invention has been made in consideration of the above-mentioned conventional circumstances, and aims to provide a surveillance system that can be operated without stopping the object detection function even when work is scheduled to be carried out within the surveillance area. [Means for solving the problem]

[0009] A surveillance system according to one embodiment of the present invention comprises a 3D sensor for detecting objects present in a surveillance area, and a server for detecting objects within the surveillance area based on point cloud data acquired by the 3D sensor, wherein specific persons permitted to move about within the surveillance area are required to wear a specified reflector, and the server determines whether an object within the surveillance area is the specific person based on the reflection intensity of the point cloud contained in the point cloud data.

[0010] Here, in the above-mentioned monitoring system, the server may determine whether the object is the specified person based on the shape, size, or position of the point cloud portion in the point cloud data where a reflection intensity equal to or greater than a reference value is obtained.

[0011] In addition, in the above surveillance system, when an object is detected within the surveillance area, the server may be configured to issue a first alarm if the object is the specified person, and to issue a second alarm if the object is not the specified person.

[0012] In the above surveillance system, the server may be configured to issue an alarm when an object is detected in the surveillance area only if the object is not the specific person. [Effects of the Invention]

[0013] According to the present invention, it is possible to provide a monitoring system that can be operated without stopping the object detection function even when work is scheduled to be performed within the monitoring area. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a diagram showing a schematic configuration of a monitoring system according to an embodiment of the present invention; [Figure 2] FIG. 2 is a diagram illustrating an example of the configuration of a server in the monitoring system of FIG. [Figure 3] FIG. 2 is a diagram illustrating an example of a processing flow of a server in the monitoring system of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0015] An embodiment of the present invention will be described with reference to the drawings. Here, a configuration will be described in which an object detection function using a 3D sensor is added to a monitoring system that performs 24-hour monitoring using a general visible light camera. An example of such a monitoring system is a track monitoring system that detects objects such as intruders and flying objects in restricted areas such as railroad tracks to ensure the safety of train operations. For example, the track monitoring system monitors restricted areas using a 3D sensor, and when an object such as an intruder or flying object is detected, the system photographs the object using a visible light camera and displays the image on a monitoring terminal.

[0016] Fig. 1 shows a schematic configuration of a monitoring system according to an embodiment of the present invention. The monitoring system in Fig. 1 includes a visible light camera 10, a 3D sensor 20, a network switch 30, a server 40, and a monitoring terminal 50.

[0017] The server 40 and the monitoring terminal 50 are realized by a computer equipped with hardware resources such as a processor and memory, and are configured to implement functions and processes related to the present invention by reading a predetermined program from the memory and executing it with the processor. The server 40 and the monitoring terminal 50 may be realized by a single computer, or may be realized by multiple computers operating in cooperation with each other.

[0018] One or more visible light cameras 10 are installed so as to capture images of a monitored area, including restricted areas. The camera images captured by the visible light cameras 10 are transmitted to a server 40 and a monitoring terminal 50 via a network switch 30. Users such as monitors can check the status of the monitored area by viewing the camera images displayed on the monitoring terminal 50.

[0019] One or more 3D sensors 20 are installed near the visible camera 10 so as to detect the status of objects within the monitoring area. The 3D sensors 20 are used to detect foreign objects such as intruders and flying objects. In this example, a LiDAR (Light Detection and Ranging) sensor is used as the 3D sensor 20, but this is just one example, and various types of 3D sensors can be used. The 3D sensor 20 generates sensor information by irradiating the monitoring area with laser light and acquiring the return value (e.g., reflection intensity). The sensor information is generated for each frame in which the monitoring area is scanned once. Examples of the sensor information include point cloud data representing three-dimensional coordinate information and brightness information of points where reflection intensity equal to or greater than a predetermined value is obtained. This sensor information is transmitted to the server 40 and the monitoring terminal 50 via the network switch 30.

[0020] The server 40 determines the status of objects in the monitoring area based on the sensor information received from the 3D sensor 20. The server 40 can determine the presence or absence of objects in the monitoring area and acquire characteristic information about the objects (such as the object's position, distance, and shape) by, for example, using a background subtraction method that calculates the difference between the current sensor information received from the 3D sensor 20 and sensor information (background information) previously obtained when there is no object to be detected (such as an intruder or flying object). Note that the background information data that can be acquired from the 3D sensor 20 is based on the use of a sensor that can consistently acquire stable data for each frame.

[0021] Furthermore, when it is determined that an object is present, the server 40 controls the visible camera 10 to capture an image of the object based on the object's characteristic information, and displays the camera image on the monitoring terminal 50 for a user, such as a monitor, to confirm. At this time, the sensor information obtained by the 3D sensor 20 may also be displayed on the monitoring terminal 50. Note that it is difficult to determine the state of the object in detail by simply visualizing and displaying sensor information such as point cloud data. Therefore, in order to be able to determine the state of the object in more detail, for example, a detailed data determination process may be performed on the application side of the monitoring terminal 50, and the results may also be displayed.

[0022] The above is an explanation of object detection, which is the basic operation of the surveillance system in this example. In a surveillance system that must constantly detect intruders and other intruders 24 hours a day, it is desirable to operate the object detection function in the same way as normal, even when scheduled work is being carried out in the surveillance area, in order to minimize the time surveillance is stopped. However, since it is necessary to prevent workers from being detected as intruders, this invention utilizes point cloud data that can be acquired by a 3D sensor to solve this problem.

[0023] 3D sensors can not only accurately obtain distance information to specific points on an object's surface (the point where the laser is aimed), but can also simultaneously obtain reflection intensity information. In other words, the point cloud data acquired by a 3D sensor contains distance information and reflection intensity information measured at multiple points on the object's surface. Using this reflection intensity information makes it possible to recognize the presence of objects that exhibit strong reflection intensity even when they should not exist. Furthermore, since determining the presence of a single point of data alone is likely to result in a false alarm, analyzing the reflection intensity and shape of multiple points (i.e., a collection of points) can more reliably identify an object. For example, since workers always wear reflectors (such as fluorescent reflective strips or helmets with reflective stickers), the presence or absence of these can be used to determine whether or not an object is a worker. Furthermore, requiring workers to wear reflectors of a specific shape can more reliably determine whether or not an object is a worker. Furthermore, it is possible to determine whether or not an object is a worker not only from multiple point data but also from multiple frames, thereby further increasing the accuracy of the determination.

[0024] 2 shows an example of the configuration of processing units related to the above-described operations among the processing units included in server 40. Server 40 shown in FIG. 2 includes a sensor information acquisition unit 41, an object detection unit 42, a reflector reference storage unit 43, a specified person determination unit 44, and an alarm output unit 45.

[0025] The sensor information acquisition unit 41 acquires sensor information from the 3D sensor 20 via the network switch 30. In this example, point cloud data including distance information and reflection intensity information for each of a plurality of points (point cloud) within the monitoring area is acquired from the 3D sensor 20 as sensor information.

[0026] The object detection unit 42 detects objects within the monitoring area based on the point cloud data acquired from the 3D sensor 20. In this example, object detection is performed using the background subtraction method described above, but object detection may also be performed using other methods.

[0027] The reflector reference storage unit 43 pre-stores reflector reference information regarding a predetermined reflector that should be worn by a specific person (e.g., a worker) who is permitted to act within the monitoring area. The reflector reference information includes reference information such as the shape of the reflector, the size of the reflector, and the position of the reflector (relative position to the person). Examples of reflectors include fluorescent reflective strips and helmets with reflective stickers that provide strong reflection intensity as described above, but other forms are also acceptable. However, it is desirable that the reflector used has a form that is easily detected by the 3D sensor 20 regardless of the wearer's movement or posture.

[0028] The specified person determination unit 44 determines whether an object detected by the object detection unit 42 is a specified person based on the point cloud data acquired from the 3D sensor 20. In this example, first, a point cloud portion in the point cloud data that exhibits a reflection intensity equal to or greater than a reference value is identified. The reference value is a reflection intensity that cannot be obtained from a typical object. Next, the specified point cloud portion is compared with information on a reflector standard to determine whether the point cloud portion is a reflector. That is, if the shape, size, position, etc. of the specified point cloud portion meet the reflector standard, the point cloud portion is deemed to be a reflector. If not, the point cloud portion is deemed not to be a reflector. Here, a point cloud portion may be deemed to be a reflector if only one of the elements of the shape, size, position, etc. of the point cloud portion meets the reflector standard, or if a predetermined number or more elements meet the reflector standard. If the point cloud portion is a reflector, the object detected by the object detection unit 42 is determined to be a specified person. If not, the object is determined to be a foreign object such as an intruder or a flying object.

[0029] Note that the above-described method for determining a specific person is an example, and the present invention is not limited to this. For example, an AI model that has previously learned point cloud data related to a person wearing a reflector may be prepared, and the AI ​​model may be used to analyze the shape, size, position, etc. of the point cloud portion to determine whether an object detected by the object detection unit 42 is a specific person.

[0030] The alarm output unit 45 outputs alarm information according to the processing results of the object detection unit 42 and the specified person determination unit 44. In this example, when an object is detected by the object detection unit 42, if the specified person determination unit 44 determines that the object is a specified person, a minor alarm (first alarm) is issued, and if not, a major alarm (second alarm) is issued. The minor alarm is an alarm that notifies that a specified person (e.g., a worker) whose behavior is permitted within the monitored area has been detected. The major alarm is an alarm that notifies that a foreign object such as an intruder or a flying object has been detected.

[0031] The alarm information output from the alarm output unit 45 includes information such as an instruction to display a text message or an image message, an instruction to output a voice message, an instruction to turn on a warning lamp, etc. Such alarm information is transmitted to an output device (monitoring terminal 50 in this example) via the network switch 30, and the output device performs output processing according to the alarm information.

[0032] 3 shows an example of a processing flow of the server 40. In the server 40, first, the sensor information acquisition unit 41 acquires point cloud data of the current frame from the 3D sensor 20 (step S11). Next, the object detection unit 42 detects an object in the monitoring area based on the point cloud data acquired from the 3D sensor 20 (step S12). If no object is detected, the process returns to step S11 and acquires point cloud data of the next frame.

[0033] If an object is detected in step S12, the specified person determination unit 44 determines whether the object detected by the object detection unit 42 is a specified person based on the reflection intensity of the points included in the point cloud data (step S13). If it is determined that a specified person has been detected, a minor alarm is output from the alarm output unit 45 (step S14), and if not, that is, if it is determined that a foreign object such as an intruder or a flying object has been detected, a major alarm is output from the alarm output unit 45 (step S15).

[0034] As described above, the monitoring system of this example includes a 3D sensor 20 for detecting objects present in a monitoring area, and a server 40 for detecting objects within the monitoring area based on point cloud data acquired by the 3D sensor 20. Specific individuals (e.g., workers) permitted to be active within the monitoring area are required to wear specific reflectors, and the server 40 determines whether an object within the monitoring area is a specific individual based on the reflection intensity of the points contained in the point cloud data. This configuration makes it possible to automatically determine whether an object detected within the monitoring area is a specific individual, eliminating the need to stop the object detection function even when work is scheduled to be performed within the monitoring area, leading to safer system operation.

[0035] Furthermore, when an object is detected within the monitored area, the server 40 is configured to issue a minor alarm (first alarm) if the object is a specific person, and to issue a major alarm (second alarm) if the object is not a specific person. This allows the user (monitoring staff) to quickly determine whether a specific person or a foreign object such as an intruder or flying object has been detected. At this time, the user can also check whether a specific person has actually been detected by viewing the camera image as needed.

[0036] In the above explanation, when an object is detected in the monitored area, a minor or major alarm is issued depending on whether the object is a specific person (first operation). However, it is also possible to use an operation (second operation) in which an alarm is not issued if the object is a specific person, and an alarm is issued only in other cases. The first and second operations may also be combined. For example, the second operation is used during periods when work is scheduled to be performed in the monitored area (periods when workers frequently enter the area), and the first operation is used during other periods. This makes it possible to reduce the frequency of alarms, which is expected to reduce the burden on users.

[0037] Although the embodiments of the present invention have been described above, these embodiments are merely illustrative and do not limit the technical scope of the present invention. The present invention can take on various other embodiments, and various modifications such as omissions and substitutions can be made without departing from the spirit of the present invention. These embodiments and modifications thereof are included in the scope and spirit of the invention described in this specification, etc., and are included in the invention described in the claims and their equivalents.

[0038] Furthermore, the present invention can be provided not only as devices such as those described above or as systems composed of these devices, but also as methods executed by these devices, programs for realizing the functions of these devices using a processor, and storage media for storing such programs in a computer-readable manner. [Industrial Applicability]

[0039] The present invention can be used in a surveillance system that uses a 3D sensor to detect objects within a surveillance area. [Explanation of symbols]

[0040] 10: Visible camera, 20: 3D sensor, 30: Network switch, 40: Server, 41: Sensor information acquisition unit, 42: Object detection unit, 43: Reflector reference storage unit, 44: Specific person determination unit, 45: Alarm output unit, 50: Monitoring terminal

Claims

1. a 3D sensor for detecting objects present in the monitoring area; a server that detects objects within the monitoring area based on the point cloud data acquired by the 3D sensor; A specific person who is permitted to move within the surveillance area is required to wear a predetermined reflector, The monitoring system is characterized in that the server determines whether an object in the monitoring area is the specified person based on the reflection intensity of the point cloud included in the point cloud data.

2. 2. The monitoring system according to claim 1, A surveillance system characterized in that the server determines whether the object is the specified person based on the shape, size, or position of the point cloud portion in the point cloud data where a reflection intensity greater than a reference value is obtained.

3. 2. The monitoring system according to claim 1, A surveillance system characterized in that, when an object is detected within the surveillance area, the server issues a first alarm if the object is the specified person, and issues a second alarm if the object is not the specified person.

4. 2. The monitoring system according to claim 1, The monitoring system is characterized in that the server issues an alarm when an object is detected in the monitoring area only if the object is not the specified person.

Images