SURVEILLANCE DEVICE AND METHOD FOR MAN-OVERBOARD MONITORING

Description

State of the art

[0001] A man-overboard monitoring device is proposed for a section of a ship. This section is monitored by at least one camera, the first camera being configured to provide video data. The monitoring device includes an evaluation unit designed to detect a moving object within the ship section based on the video data.

[0002] On ships, and especially on passenger ships and cruise ships, it is a well-known problem that passengers fall overboard unnoticed while underway. Such events are known as man-overboard incidents. The chances of survival in such an event decrease significantly with the time that passes before the incident is discovered.

[0003] Shipping companies have a strong interest in detecting man-overboard incidents as quickly as possible. Government regulations and / or insurance companies also exert increased pressure on shipping companies to monitor such incidents as effectively as possible.

[0004] For cargo ships, for example, the plan is to equip personnel with special wristbands or other technology that they must wear to detect man-overboard incidents. Radar-based methods are relatively expensive and require a large number of sensors.

[0005] Image- and / or video-based approaches have the problem of being prone to interference due to the multitude of possible events, objects, and a dynamic background during the journey, with many events being detected and evaluated as false man-overboard events.

[0006] Publication D1 (WO 2017 / 187407 A1) discloses a combination of camera and radar for man-overboard monitoring on ships. Changing pixels are determined by calculating differences. These changing pixels are then analyzed for movement, and accelerations are calculated. Furthermore, distances are determined using radar, and relative sizes are inferred from the pixel data. If a man-overboard is detected, an alarm is triggered with video data. If a man-overboard is not detected, no alarm is generated. Disclosure of the invention

[0007] A man-overboard monitoring device with the features of claim 1 is proposed. Furthermore, a man-overboard monitoring method with the features of claim 14 and a computer program with the features of claim 15 are provided. Preferred and / or advantageous embodiments of the invention are described in the dependent claims, the description, and the accompanying figures.

[0008] A monitoring device for man-overboard surveillance in a section of a ship is proposed. In particular, the monitoring device can be configured to perform man-overboard surveillance in multiple ship sections, especially connected or disconnected sections. Specifically, the monitoring device is designed for video and / or image-based man-overboard surveillance. The monitoring device can determine whether a person has fallen overboard or whether it is another non-human object. The ship section is, for example, the area surrounding a camera or multiple cameras. The ship section constitutes a spatial segment. Preferably, the ship section includes an area beyond the ship's railing and / or a portion of the sea.

[0009] The section of the ship is monitored by video equipment using at least one camera. In particular, the section can be monitored by multiple cameras using video and / or image technology. The cameras can have a common overlapping area; alternatively, the monitoring areas of the cameras are designed to be non-overlapping. The cameras are preferably color and / or black-and-white cameras. Alternatively, the camera can be a thermal imaging camera. The camera is designed to record an image, a video, and / or an image sequence of the section of the ship. Furthermore, the camera is designed to be able to provide the images, the video, and / or the image sequence as video data to an external system. The cameras are preferably arranged so that they have a horizontal viewing direction. Furthermore, it can be provided that the cameras specifically include a viewing angle to the horizontal of less than twenty degrees.The camera can be used to detect and / or record objects that fall overboard and / or land in the sea. The sections of the ship being monitored are specifically safety-relevant areas, such as sections accessible to people and / or passengers, with a risk of falling into the sea. A large number of cameras are installed on the ship to monitor a wide variety of sections, and these cameras are designed to transmit their video data to the monitoring system.

[0010] The monitoring device includes an evaluation unit. The evaluation unit can be a hardware module or, alternatively, a software module. For example, the evaluation unit may be a computer unit, a processor, or a microchip. The evaluation unit has an interface for receiving the video data. The interface can be physical, such as a cable interface, wireless, or virtual. Preferably, the video data from multiple cameras and / or all cameras used to monitor sections of the ship are provided to the evaluation unit. In particular, the evaluation unit can also be configured to store the video data. Furthermore, the evaluation unit may be a central evaluation unit, for example, located in a ship's control center.

[0011] The evaluation unit is designed to detect a moving object within the ship's section based on video data. For example, the evaluation unit analyzes the video data for movements and / or objects. This involves, for instance, comparing images and / or image sequences within the video data, particularly from one or more cameras. Specifically, the evaluation unit can be configured to detect downward-moving objects within the ship's section.

[0012] The evaluation unit is configured to determine a kinematic quantity for a moving object. This kinematic quantity could be, for example, the object's velocity as a vector, as a magnitude, and / or as a direction. Alternatively, the kinematic quantity could be a path, an momentum, a starting point, or an endpoint. Preferably, several or all kinematic quantities of the object necessary to describe its motion are determined. In particular, the evaluation unit determines the kinematic quantity for a plurality and / or all frames of an image sequence and / or a video. Alternatively and / or additionally, the evaluation unit may be configured to determine the kinematic quantity at predetermined time intervals, for example, at intervals of 1 millisecond, 500 microseconds, or one second.

[0013] The invention provides that the evaluation unit is configured to determine the moving object as a man-overboard event based on its dimensions. The dimensions are, in particular, an area dimension, for example, a diameter or a surface area. Alternatively, the dimensions can be a volume dimension. The dimensions are, for example, the length, width, or diameter of the object. The dimensions are, in particular, a measure of the object's actual size. Preferably, the dimensions are the maximum longitudinal extent of an object. The object's dimensions are determined, in particular, based on the video data and the kinematic size. Preferably, the evaluation unit determines a scale, for example, a projection scale, using the kinematic size and the video data, and the object's dimensions are then determined based on this scale.For example, the object in the video data is measured and / or determined based on its scale. The object's dimensions are specifically its actual dimensions. For example, the evaluation unit is configured to analyze the video data and / or the kinematic size of the object and / or the event associated with it, and to examine it for the possibility of a man-overboard incident, with verification or rejection as a man-overboard incident based on the object's dimensions. For example, the evaluation unit is configured to assess all accelerated movements, particularly vertically directed and / or seaward accelerated movements, as potential man-overboard incidents, with exclusion and / or verification of whether it is a true man-overboard incident determined based on the dimensions.In this way, for example, it can be ruled out that a falling object, such as a cigarette, is falsely detected as a man-overboard event.

[0014] The invention is based on the premise that, while purely video-based approaches in the prior art are cost-effective to install, the image processing algorithms are not yet sensitive and robust enough to suppress false detections. Weather conditions and lighting scenarios, in particular, lead to an increased number of false alarms. Furthermore, dynamic environments with significant movement, such as waves, spray, moving objects on water or land, people moving on board, birds, and other events, complicate the application of video-based approaches. However, the present invention provides a cost-effective and flexibly deployable monitoring device that enables highly reliable and robust automatic detection of man-overboard incidents using video cameras.In particular, the monitoring device is characterized by a low probability of false alarms.

[0015] One embodiment of the invention provides that the evaluation unit is configured to determine a gravitational acceleration as a kinematic quantity. For example, the evaluation unit is configured to determine the object's accelerations in the video data, for instance, by temporal tracking and / or by evaluating the trajectory. Preferably, the gravitational acceleration and / or acceleration in the vertical direction, particularly perpendicular to the sea surface, is determined. Specifically, it can be provided that the trajectory is modeled as a projectile parabola, with the gravitational acceleration being determined from the trajectory and / or projectile parabola. For example, the motion is determined as a quadratic function and / or as a polynomial, with the gravitational acceleration being interpreted, in particular, as the coefficient of the quadratic term.The measured and / or determined acceleration due to gravity is compared, in particular, with a reference acceleration due to gravity. This comparison is performed primarily by the evaluation unit. The reference acceleration due to gravity is understood to be the physical value of the acceleration due to gravity at the respective location. For example, the value 9.81 m / s² is used as the reference acceleration. Based on the finding that the measured acceleration due to gravity in the video data corresponds to the acceleration due to gravity of 9.81 meters per second and / or the reference acceleration, the evaluation unit can, for example, determine a scale in the video data and / or images and / or, in particular, thus determine the dimensions of the object. This finding is based on the consideration that, with a measured acceleration due to gravity and a comparison with a reference acceleration due to gravity, it is possible to assign pixels to actual meters.

[0016] It is particularly preferred that the evaluation unit includes a tracking module. The tracking module can be implemented as a software module or, alternatively, as a hardware module such as a processor or a microchip. The tracking module is configured to determine a trajectory for the moving object based on the video data. For example, a sequence of video images is evaluated at fixed and / or varying time intervals, and the object's path and / or trajectory is reconstructed and / or defined. Tracking the object can be based on a graphical method, a numerical method, or, in particular, a model, such as a parabolic trajectory. The evaluation unit is further configured to determine the kinematic quantity based on the trajectory.For example, by analyzing the trajectory, the direction of the velocity, the velocity itself, and / or the acceleration due to gravity can be determined. For instance, the velocity is defined as the tangent vector to the trajectory at the given time. In particular, the trajectory can be displayed to a user and / or be viewable by the user, allowing them to visually assess and / or track the event.

[0017] Optionally, the evaluation unit may include a selection module. This selection module is designed to determine the kinematic size of objects with parabolic trajectories. Specifically, the selection module is designed to determine the trajectories and / or kinematic sizes of objects with accelerated motion in the vertical direction. Furthermore, the selection module is designed to exclude objects with non-parabolic trajectories and / or objects without accelerated motion in the vertical direction from the determination of their kinematic size. For example, such objects and / or events with unaccelerated vertical motion and / or without parabolic trajectories are excluded as irrelevant events, and / or their data is not monitored further. This allows for data reduction for the actual evaluation and / or for determining the extent.This design is based on the idea that by filtering beforehand into objects and / or events that exhibit accelerated motion and those that do not, objects that do not fall freely, such as birds, spray and / or waves, can be excluded.

[0018] One embodiment of the invention provides that the evaluation unit includes a segmentation module. The segmentation module is specifically designed as a software module. The segmentation module is configured to segment a moving object in the video data, individual frames, and / or the video image sequence as an object against a background. In particular, it is possible for the segmentation module to be configured to segment the object against a dynamic background. The segmentation module recognizes, for example, object segments that form a coherent relationship and / or belong together. For example, a local significant difference and / or a deviation from the image background is determined. In particular, additional information can be used for segmentation, wherein the additional information presupposes, for example, a continuous and / or steady path and / or trajectory.This design is based on the consideration that a moving object forms an extended object and has an object relationship that can be defined by segmentation.

[0019] It is particularly preferred that the tracking module is configured to determine the trajectory based on the segmented object and, in particular, on the center of gravity of the segmented object. Furthermore, the evaluation unit may be configured to determine the kinematic value for the center of gravity of the segmented object. This configuration is based on the consideration that freely falling objects often have a rotational degree of freedom, meaning that a point on the object's surface changes over time, whereas the center of gravity follows the intended parabola and / or trajectory.

[0020] It is particularly preferred that the evaluation device is designed to determine a resolution in the video data images based on the kinematic quantity. Specifically, resolution is understood as the mapping of pixels to meters and / or, more generally, pixels to units of length. For example, an acceleration value and / or acceleration in the form of pixels per square centimeter is defined as the kinematic quantity. Since a reference acceleration due to gravity, for example, 9.81 meters per second squared, is known, a resolution can be determined by dividing the determined acceleration value by the reference acceleration due to gravity. By evaluating the object size in the video data and / or individual frames, for example, as pixel size, the actual extent and / or a real size of the falling object can thus be determined.Thus, it is possible to determine the size of the falling and / or moving object without explicit calibration of the camera and / or video data, and thus, for example, to distinguish a falling person from a falling cigarette.

[0021] It is particularly preferred that the monitoring device includes a first camera for providing the video data. The first camera is, in particular, a color camera, a black-and-white camera, or a grayscale camera. The first camera is configured to monitor the section of the ship using video technology and to provide the video data. The monitoring of the ship section by the first video camera takes place in the visible wavelength range. In particular, the monitoring device can also include a plurality of first cameras for providing a plurality of video data. It is particularly recommended that the evaluation unit is a central evaluation unit, for example, located in a monitoring room, with the first cameras being connected to the evaluation unit via data transmission.

[0022] It is particularly preferred that the monitoring device includes a second camera for providing additional video data. In particular, the second camera is configured to monitor the same section of the ship as the first camera. Thus, the first and second cameras preferably image a common section of the ship. The second camera is, in particular, an infrared camera or a thermal imaging camera. Specifically, it is provided that the first and second cameras image the common section of the ship in different wavelength ranges. By combining a first and a second camera with different wavelength ranges, the usable wavelength range is increased, and the detection reliability can be significantly improved.For example, a thermal imaging camera can offer significant advantages over a standard camera at night, in fog, and / or in bad weather, whereas a thermal imaging camera is inferior to a standard camera on hot summer days. In particular, the use of both the first and second cameras allows the ship's crew to better verify an alarm and / or an incident. Furthermore, it is preferable for the first and second cameras to form a single camera module.

[0023] One embodiment of the invention provides that the first camera, the second camera, and / or the camera module includes and / or comprises the evaluation unit. This allows for the determination of the extension and / or the determination of the kinematic quantity to take place in the first camera, the second camera, and / or the camera module. Preferably, the evaluation in the first camera and the second camera is performed independently of each other. For example, the video data from the first camera and the video data from the second camera are evaluated independently of each other, and the kinematic quantities and / or the extension of the object are determined independently. Furthermore, it can be provided that the first camera and / or the second camera, specifically the camera module, is configured to output the video data and / or the kinematic quantity and / or the extension of the object as data.This design is based on the consideration that in confined spaces on the ship it makes sense to integrate the computing power into the cameras themselves, so that space for central, extensive computing units can be reduced.

[0024] Optionally, the evaluation unit is designed to issue an alarm upon detection of a man-overboard event. For example, the monitoring device is designed to issue an alarm if a kinematic quantity exceeds or falls below a certain value, such as acceleration due to gravity and / or velocity. Alternatively and / or additionally, the alarm may be issued if the size corresponds to a minimum size and it can be assumed that the object is a person. In particular, the first camera, the second camera, and / or the third camera are designed to output an alarm or a signal to issue an alarm in addition to the video data.

[0025] It is particularly desirable that the evaluation unit, the first camera, the second camera, and / or the camera module is configured to output the corresponding video data when an alarm is triggered. For example, this makes it possible to provide a user with relevant video data directly when an alarm is triggered and / or a man-overboard incident occurs, enabling them to quickly and efficiently decide whether further action is necessary.

[0026] It is particularly preferred that the evaluation unit includes a model module. This model module contains additional information, especially regarding the section of the ship and / or its surroundings. For example, the model module includes information on the location of particularly dangerous areas and / or areas where a person might fall overboard. The evaluation unit is designed to determine the kinematic magnitude and / or the detection of the man-overboard event by incorporating this additional information. For example, the evaluation unit can exclude areas of the ship section where man-overboard events cannot occur, for instance, because there is no access to the sea and / or sufficient safety measures are in place to prevent a person from falling overboard there.

[0027] Another aspect of the invention is a method for determining a man-overboard incident. For this purpose, a section of a ship is monitored using one or more cameras. The video monitoring provides video data, individual images, and / or image sequences. By evaluating the video data, a kinematic quantity, for example, the acceleration due to gravity in the vertical direction, is determined. Based on the determined kinematic quantity, the size of the object is determined, for example, by comparing a reference acceleration due to gravity with a specific acceleration, whereby, in particular, if a minimum size is exceeded, a man-overboard incident is inferred.

[0028] A further aspect of the invention is a computer program, in particular a computer program stored on a data carrier. The computer program is configured to perform all steps of the method when the computer program is executed on a computer, the monitoring device, and / or the evaluation unit.

[0029] Another object of the invention is a ship, in particular a passenger ship, with a ship section and the described monitoring device for man-overboard monitoring in the ship section.

[0030] Further advantages, effects, and features are shown in the attached figures and their descriptions. These show: Figure 1 schematically a monitoring device; Figure 2 Example camera distribution on a ship; Figure 3 exemplary trajectory; Figure 4 Image with additional information.

[0031] Figure 1Figure 1 shows a monitoring device 1. The monitoring device 1 is used to monitor a ship 12 and / or a section of the ship 5. The monitoring device 1 can detect when a person 6 goes overboard. In particular, the monitoring device 1 serves to locate and / or monitor man-overboard incidents. The monitoring device 1 can distinguish a person going overboard from other objects going overboard, such as cigarettes or trash, especially by comparing sizes and / or assessing the size of the object going overboard. Furthermore, the monitoring device 1 can differentiate between objects going overboard, especially falling objects, and objects that move in other ways or upwards, such as a bird.

[0032] The monitoring device 1 comprises two cameras, namely a first camera 2 and a second camera 3. The first camera 2 and the second camera 3 are arranged in a camera module 4. The camera module 4 is positioned on the ship such that a section of the ship 5 can be monitored video-technically using this module and, in particular, also using cameras 2 and 3. The first camera 2 is designed as a color video camera to capture images in the visible wavelength range. The second camera 3 is designed as a thermal imaging camera and monitors the ship section 5 in the IR or NIR range. The first camera 2 and the second camera 3 are designed to provide video data. The video data comprises the video-technical monitoring of the ship section 5.

[0033] Ship section 5 is, for example, an area beyond the railing, particularly facing the sea and / or the exterior of the ship. Preferably, the monitoring device 4 is arranged such that the first camera 2 and the second camera 3 have a horizontal viewing direction. A person 6, as an object that goes overboard and / or could go overboard, can be detected by means of video surveillance through the camera module 4. A person 6 who goes overboard describes a parabolic path 7 as a trajectory. The parabolic path 7 corresponds to a projectile motion, exhibiting a constant horizontal velocity VX and a vertical velocity VY, where the vertical velocity VY represents an accelerated motion.Furthermore, person 6 has an extent 8, where the extent 8 can be, for example, the height and / or size of person 6 or a diameter; for example, a rectangle is circumscribed around person 6, where the extent 8 would then be the diagonal of the rectangle.

[0034] The falling person 6 in ship section 5 is recorded by the first camera 2 and the second camera 3. This video surveillance is provided as video data to an evaluation unit 9. The video data is provided to an interface 10, which is included in the evaluation unit 9.

[0035] The evaluation unit 9 is designed to detect the moving object 6 in ship section 5 based on the video data. Specifically, the movement of object 6 is tracked, for example, in successive frames within the video data. The determination of the moving object and / or the trajectory is performed independently for the video data from the second camera 3 and the first camera 2. Based on the tracked object 6, a kinematic quantity of the object's movement, in this case, the falling person 6, is determined. This kinematic quantity is the acceleration due to gravity and / or the velocity profile in the vertical direction. Furthermore, the kinematic quantity can include a velocity or another physical quantity. Based on this kinematic quantity, in this case, the measured acceleration due to gravity, the video data is evaluated to determine the size and / or extent 8 of the falling object.For example, the measured acceleration is compared with a reference acceleration due to gravity, such as the acceleration due to Earth's gravity. Once the value of the reference acceleration due to gravity is known, for example, 9.81 m / s², and the number of pixels per second squared in the image representing the acceleration due to gravity is known, a scale can be established, in this case, the mapping of pixels to meters. By measuring the dimensions of the object and / or person 6 in the image, based on pixels, the dimensions 8 in the real world can be determined. By comparing these dimensions 8 with a minimum dimension, for example, a minimum size for a human, it can be determined whether a potential man-overboard event is present and, if so, an alarm can be triggered. If the measured and / or determined dimensions 8 are below the minimum dimension, the event is classified as not critical and dismissed.In the event of a man-overboard incident or suspected man-overboard incident, an alarm 11, for example an audible alarm, can be triggered. Preferably, the corresponding video footage from the relevant location and / or camera 2, 3 is provided simultaneously with the alarm 11.

[0036] Figure 2Figure 1 shows a ship 12. The ship 12 is shown in a top view. The ship 12 is, in particular, a passenger ship, preferably a cruise ship. The exterior areas 13 of the ship are monitored by camera modules 4. The camera modules 4 are designed, in particular, as previously described. The camera modules 4 are arranged on the exterior of the ship 12 such that they have a horizontal viewing direction and each monitors a section of the ship using video technology. Each camera module 4 monitors one section 5 of the ship. At least as many camera modules 4 are arranged as there are monitoring areas 5. By mounting the camera modules 4 in this way, the entire exterior of the ship is monitored, so that man-overboard incidents can be detected.The camera modules 4 can be connected to a central evaluation unit, which is located, for example, in a monitoring room where all alarms from the camera modules 4 converge and the video material can be viewed if necessary.

[0037] Figure 3Image 14 shows the video data, recorded with one of cameras 2 or 3. Image 14 depicts ship section 5, which is divided into an exterior area 15 and ship section 16. Ship section 16 shows parts of the ship, in this case, an outer wall. The image also shows the path 7 taken by a falling person. The falling person jumps over the deck and initially follows a rising parabolic trajectory 17a, which transitions into a descending parabolic section 17c after the apex 17b. The path 7 has been tracked or modeled as a trajectory by the evaluation unit 9. From the trajectory, here the parabolic shape of path 7, the acceleration due to gravity A can be determined. This acceleration is then compared to a reference acceleration due to gravity g, allowing the size of the falling object to be inferred from image 14, thus determining the object's extent.

[0038] Figure 4Image 14, taken by one of cameras 2 or 3, also shows the ship section 5 divided into an exterior area 15 and a ship area 16. Additional information can be stored and / or entered using a model module, which is part of the monitoring device 1. In this case, a model of ship section 5 is stored, with ship area 16 excluded from analysis. This allows for a robust, data-efficient, and streamlined evaluation of the video data. A further exclusion criterion for a man-overboard event is the distance of the apex of the fall trajectory from the ship's edge. This exclusion can be achieved by marking an exclusion zone.Preferably, the monitoring device 1 is configured such that a user can mark ship area 16 in image 14 via a human-machine interface when configuring the monitoring device 1, with the model module being configured to store this as additional information. Furthermore, the monitoring device 1 is configured such that a user can mark image regions of particularly dangerous areas and / or areas from which a person can go overboard via the human-machine interface when configuring the monitoring device 1. Alternatively or additionally, the monitoring device 1 is configured such that a user can mark image regions of areas where man-overboard events cannot occur when configuring the monitoring device 1 via the human-machine interface.

Claims

1. A monitoring device (1) for monitoring a man overboard situation in a ship section (5), wherein the ship section (5) is monitored by video means of at least one camera (2), the camera being configured to provide the monitoring in the form of video data, with an analysis device (9), the analysis device (9) including an interface (10) for transferring the video data, wherein the analysis device (9) is configured to detect a moving object in the ship section (5) based on the video data and determine the kinematic quantity of the moving object, wherein the analysis device (9) is configured to determine a scaling for determining an extent (8) of the moving object based on the video data and the kinematic variable and evaluate the moving object as a man-overboard event based on its extent (8), characterized in that the analysis device (9) comprises a model module, the model module including additional information about the ship section (5) as a model of the ship section (5), the ship section (5) being divided into two parts, in an outer area (15) and a ship area (16), wherein the analysis device (9) is configured to determine the kinematic quantity taking into account the additional information (15), the ship area (16) being excluded from the evaluation.

2. The monitoring device (1) according to claim 1, characterized in that the analysis device (9) is configured to determine a measured free-fall acceleration as a kinematic variable and, in order to determine the extent (8), to compare the measured free-fall acceleration (A) with a reference free-fall acceleration.

3. The monitoring device (1) according to any of the previous claims, characterized in that the analysis device (9) comprises a tracking module, wherein the tracking module is configured to determine a trajectory for the moving object based on the video data, wherein the kinematic variable can be determined from the trajectory.

4. The monitoring device (1) according to claim 3, characterized in that the analysis device (9) comprises a selection module, the selection module being configured to determine the kinematic quantity of objects with parabolic trajectories and / or to exclude objects with non-parabolic trajectories from the determination of the kinematic quantity.

5. The monitoring device (1) according to any of the previous claims, characterized in that the analysis device (9) has a segmentation module, the segmentation module being configured to segment a moving object in the video data and / or individual frames of the video data as an object against a background.

6. The monitoring device (1) according to claim 5, characterized in that the tracking module is configured to determine the trajectory based on the segmented object, in particular the center of gravity of the segmented object.

7. Monitoring device (1) according to one of the preceding claims, characterized in that the analysis device (9) is configured to determine a resolution in images (14) of the video data based on the kinematic quantity.

8. The monitoring device (1) according to any of the previous claims, characterized by a first camera (2) for providing the video data.

9. The monitoring device (1) according to claim 8, characterized by a second camera (3) for providing further video data, wherein the first camera (2) and the second camera (3) monitor a common region, the analysis device (9) being configured to analyze the video data of the first camera (2) and the second camera (3) separately and / or independently.

10. The monitoring device (1) according to claim 8 or 9, characterized in that the analysis device (9) is comprised of the first and / or second camera (2, 3).

11. The monitoring device (1) according to any of the preceding claims, characterized in that the monitoring device (1) is configured to issue an alarm when a man-overboard event is detected.

12. The monitoring device (1) according to claim 11, characterized in that the analysis device (9) is configured to output video data corresponding to an alarm.

13. The monitoring device (1) according to any of the preceding claims, characterized in that the additional information to the ship section (5) can be provided.

14. A method for determining a man-overboard event, in particular using the monitoring device (1) according to any of the preceding claims, wherein a ship section (5) is monitored by video means of a camera (2) and video data is provided, wherein based on the video data, a kinematic quantity of a moving object is determined, wherein a scaling and / or an extent of the object is determined based on the video data and the kinematic quantity, wherein the moving object is evaluated as the man-overboard event on the basis of its extent (8), characterized in that a model module comprises additional information to the ship section (5) as a model of the ship section (5), wherein the ship section (5) is divided into two parts, into an outer area (15) and a ship area (16), wherein the kinematic quantity is determined taking into account the additional information (15), with the ship area (16) being excluded from an evaluation.

15. A computer program, in particular a computer program on a data carrier, wherein the computer program is configured to perform all steps of the method according to claim 14 when the computer program is executed on a computer, the monitoring device (1) and / or the analysis device (9).