A remote monitoring method for ship status based on data analysis

Abstract

The present invention relates to the field of data analysis, and in particular to a method for remote monitoring of ship status based on data analysis, comprising: periodically collecting each operating data of each target ship equipment, determining an abnormal data extraction method according to the associated abnormal state of the operating data; determining whether the operating data is abnormal operating data according to whether each operating data is in a corresponding preset range; performing risk factor detection and analysis on each monitoring device respectively, and determining an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device; determining a data transmission state according to data parameters and detail parameters of data to be uploaded; determining an abnormal monitoring method according to the data transmission state of data to be uploaded; and determining whether there is a communication network abnormality according to whether a reconstruction error is greater than a preset judgment threshold. The present invention improves the reliability of communication security monitoring and the real-time performance of the data transmission process during the remote monitoring of the ship status.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to a method for remotely monitoring ship status based on data analysis. Background Art

[0002] Remote monitoring of ship status ensures the safety of the navigation process by real-time collection and data analysis of various data during the ship's navigation. The current existing remote monitoring methods for ship status often only send abnormal or all operating data or video information to the user, resulting in the inability to balance data transmission efficiency and the effectiveness of abnormal operating data and abnormal monitoring videos, which in turn leads to low efficiency in users' processing of the acquired abnormal operating data and abnormal monitoring videos. Therefore, how to ensure data transmission efficiency while taking into account the effectiveness of abnormal operating data and abnormal monitoring videos is an urgent problem to be solved by technical personnel in this field.

[0003] Chinese patent application publication number CN112255946A discloses a remote monitoring system and method for ship status based on cloud services and big data, including: a ship real-time monitoring module, a ship data acquisition module, a data preprocessing module, a situation awareness module, an autonomous driving and navigation module, a main control module, a communication and navigation module, a ship-shore communication module, a remote monitoring and management module, a ship information management module, a ship shore station VDR module, a KPI auxiliary analysis module, a remote diagnosis and maintenance module, and a display module. The above scheme manages the ship's energy consumption, navigation route, equipment status, operating specifications, etc., to improve the ship's navigation efficiency and safety. However, the above scheme has the following problems: it fails to select abnormal data in a targeted manner according to the actual working scenario, and a large amount of data transmission easily increases the risk of leakage in the data transmission process, making the data transmission efficiency unable to effectively meet user needs. Summary of the invention

[0004] To this end, the present invention provides a remote monitoring method for ship status based on data analysis, so as to overcome the problem that the prior art fails to specifically select the method for extracting abnormal operation data and abnormal monitoring videos according to the actual working scenario, resulting in low effectiveness and integrity of the abnormal operation data and abnormal monitoring videos sent to users.

[0005] To achieve the above object, the present invention provides a method for remote monitoring of ship status based on data analysis, comprising:

[0006] Continuously collect the operating data of each target ship equipment;

[0007] For a single target ship equipment, if the operating data is not within the reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operating data, and the abnormal data extraction method is determined according to the associated abnormal state of the operating data. The abnormal data extraction method is to set both the operating data and its associated data as abnormal operating data or only the operating data as abnormal operating data;

[0008] Perform risk factor detection and analysis on each monitoring device respectively, and determine an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device. The image extraction strategy is to determine a video extraction method according to the risk factors corresponding to the monitoring image and the related monitoring images and the quality of the monitoring images, or to determine whether to adjust the preset video extraction time of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image;

[0009] Obtaining the data transmission volume of the data to be uploaded and the data transmission frequency of the data transmission network, and determining the data transmission status according to the data parameters and detail parameters of the data to be uploaded;

[0010] Determine an abnormality monitoring method according to the data transmission status of the data to be uploaded, and the abnormality monitoring method is to use a single feature model or a multi-feature model to monitor the communication network;

[0011] Whether there is a communication network abnormality is determined based on whether the reconstruction error is greater than a preset judgment threshold. If there is a communication network abnormality, a data communication warning is sent to the user. If there is no communication network abnormality, the data to be uploaded is sent to the user normally.

[0012] Furthermore, for an operation data of a single target ship equipment, if the operation data is not within the corresponding reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operation data, and the associated abnormal state includes:

[0013] A type of associated abnormal state in which the proportion of abnormal associated data in the operating data is greater than the preset proportion of abnormal associated data;

[0014] The second type of associated abnormal state where the proportion of abnormal associated data in the operating data is less than or equal to the preset abnormal associated data proportion;

[0015] Each operating data is individually set with a corresponding reference preset range.

[0016] Furthermore, the abnormal data extraction method is determined according to the associated abnormal state of the operation data, and the abnormal data extraction method includes:

[0017] If the operation data is in a related abnormal state, the operation data and its related data are set as abnormal operation data;

[0018] If the operation data is in the second-class associated abnormal state, only the operation data is set as abnormal operation data.

[0019] Further, monitoring images of each monitoring device are obtained, and risk factor detection and analysis are performed on each monitoring device respectively. In the risk factor detection and analysis of a single monitoring device, an image extraction strategy is determined according to the risk factor of each monitoring image corresponding to the monitoring device. The image extraction strategy includes:

[0020] If the risk factor is within a first preset risk factor range, determining a video extraction method according to the risk factor of the monitoring image and its related monitoring images and the quality of the monitoring images;

[0021] If the risk factor is within a second preset risk factor range, it is determined whether to adjust the preset video extraction time length of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image.

[0022] Further, if the risk factor is within the first preset risk factor range, a video extraction method is determined according to the risk factors corresponding to the monitoring image and the related monitoring image and the quality of the monitoring image, and the video extraction method includes:

[0023] If there is a related monitoring image with a risk factor greater than a first preset risk factor among the related monitoring images of the monitoring image, an abnormal monitoring image is determined according to the image quality reference value of the monitoring image and its related monitoring images, and an abnormal monitoring video with a preset video extraction time is extracted for the abnormal monitoring image;

[0024] If there is no related monitoring image with a risk factor greater than the first preset risk factor among the related monitoring images of the monitoring image, abnormal monitoring video extraction is not performed for the monitoring image;

[0025] The abnormal monitoring image is an image corresponding to the maximum value of the image quality reference value in the monitoring image and its related monitoring images, the related monitoring image is a monitoring image collected by the relevant monitoring device at the monitoring time corresponding to the corresponding dangerous image, and the relevant monitoring device is a monitoring device whose monitoring range overlaps with the monitoring device that took the corresponding dangerous image greater than the preset monitoring range overlap.

[0026] Further, if the risk factor is within a second preset risk factor range, determining whether to adjust the preset video extraction time length for the monitoring image according to the risk factor of the monitoring image of the adjacent frame of the monitoring image;

[0027] If there is an adjacent frame monitoring image whose risk factor is greater than the first preset risk factor in the monitoring image, the preset video extraction time length of the monitoring image is increased and adjusted;

[0028] The increase in the preset video extraction time is positively correlated with the risk factor of the adjacent frame monitoring image.

[0029] Further, the data transmission status is determined according to the data parameters and detail parameters of the data to be uploaded, and the data transmission status includes:

[0030] A first preset transmission state in which the data parameter is greater than a preset data parameter and the detail parameter is less than or equal to a preset detail parameter;

[0031] A second preset transmission state in which the data parameter is less than or equal to the preset data parameter or the detail parameter is greater than the preset detail parameter;

[0032] The data to be uploaded includes abnormal operation data and abnormal monitoring video;

[0033] The data parameters are determined according to the data transmission volume and the data transmission frequency of the data to be uploaded, and the detail parameters are determined according to the low-frequency components of the data to be uploaded.

[0034] Furthermore, the abnormality monitoring method is determined according to the data transmission status of the data to be uploaded, and the abnormality monitoring method includes:

[0035] The data to be uploaded is in a first preset transmission state, and a single feature model is used to monitor the communication network;

[0036] The data to be uploaded is in a second preset transmission state, and a multi-feature model is used to monitor the communication network.

[0037] Furthermore, the observation scale of the multi-feature model is set according to the proportion of the reference low-frequency component;

[0038] The proportion of one type of observation scale in the observation scale of the multi-feature model is positively correlated with the proportion of the reference low-frequency component;

[0039] The first type of observation scale is an observation scale with a time span greater than a preset time span.

[0040] Furthermore, the reconstruction error is compared with a preset judgment threshold, and whether there is a communication network abnormality is determined based on whether the reconstruction error is greater than the preset judgment threshold. If the reconstruction error is greater than the preset judgment threshold, the communication network is determined to be abnormal, and a data communication warning is sent to the user.

[0041] Compared with the prior art, the beneficial effect of the present invention lies in that the technical solution of the present invention not only periodically collects and analyzes the operating data of each target ship equipment and the monitoring images of each monitoring device, but also monitors the communication security during the data transmission process, thereby avoiding the leakage of transmission content or the loss of transmission efficiency due to attacks and intrusions during the data transmission process. The present invention improves the accuracy of remote monitoring results and the reliability of communication security monitoring.

[0042] Furthermore, the present invention determines the associated abnormal state according to the proportion of abnormal associated data of the operation data, and determines the abnormal data extraction method according to the associated abnormal state of the operation data. If the operation data is in a first type of associated abnormal state, the operation data and its associated data are both set as abnormal operation data; if the operation data is in a second type of associated abnormal state, only the operation data is set as abnormal operation data. The present invention improves the validity and integrity of the collected abnormal operation data.

[0043] Furthermore, the present invention determines an image extraction strategy according to the risk coefficient of each monitoring image corresponding to the monitoring device. If the risk coefficient of the monitoring image is within a first preset risk coefficient range, the video extraction method is determined according to the risk coefficient of the monitoring image and its related monitoring images and the quality of the monitoring image. For monitoring images suspected of having dangerous behaviors, the risk coefficients of their related monitoring images are combined to determine whether there are dangerous behaviors within the monitoring range, so that the collected abnormal monitoring videos are more representative. The present invention improves the effectiveness and integrity of the collected abnormal monitoring videos.

[0044] Furthermore, the present invention determines whether to adjust the preset video extraction time for the monitoring image according to the danger coefficient of the adjacent frame monitoring image when the danger coefficient of the monitoring image is within the second preset danger coefficient range; when the danger coefficient of the adjacent frame monitoring image is relatively large, the preset video extraction time is increased and adjusted, so that the information contained in the acquired abnormal monitoring video is more complete. The present invention improves the validity and integrity of the collected abnormal monitoring video.

[0045] Furthermore, the present invention determines the abnormal monitoring mode according to the data transmission status of the data to be uploaded, so that the abnormal monitoring mode is more in line with the actual working scenario. The data to be uploaded is in a first preset transmission state, and a single feature model is used to monitor the communication network, thereby ensuring the efficiency of data processing during the communication monitoring process. The data to be uploaded is in a second preset transmission state, and a multi-feature model is used to monitor the communication network, thereby ensuring the accuracy of the monitoring results. The present invention improves the reliability of communication security monitoring during the remote monitoring of the ship status and the real-time performance of the data transmission process. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of a method for remote monitoring of ship status based on data analysis of the present invention;

[0047] Figure 2 This is a flow chart of the present invention for determining an abnormal data extraction method according to an associated abnormal state of operating data;

[0048] Figure 3 This is a flow chart of the present invention for determining an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device;

[0049] Figure 4 This is a flow chart of the present invention for determining an abnormality monitoring method according to the data transmission status of data to be uploaded. DETAILED DESCRIPTION

[0050] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0052] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0053] See also Figures 1 to 4 As shown, the present invention provides a method for remote monitoring of ship status based on data analysis, comprising:

[0054] Continuously collect the operating data of each target ship equipment;

[0055] For a single target ship equipment, if the operating data is not within the reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operating data, and the abnormal data extraction method is determined according to the associated abnormal state of the operating data. The abnormal data extraction method is to set both the operating data and its associated data as abnormal operating data or only the operating data as abnormal operating data;

[0056] Perform risk factor detection and analysis on each monitoring device respectively, and determine an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device. The image extraction strategy is to determine a video extraction method according to the risk factors corresponding to the monitoring image and the related monitoring images and the quality of the monitoring images, or to determine whether to adjust the preset video extraction time of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image;

[0057] Obtaining the data transmission volume of the data to be uploaded and the data transmission frequency of the data transmission network, and determining the data transmission status according to the data parameters and detail parameters of the data to be uploaded;

[0058] Determine an abnormality monitoring method according to the data transmission status of the data to be uploaded, and the abnormality monitoring method is to use a single feature model or a multi-feature model to monitor the communication network;

[0059] Whether there is a communication network abnormality is determined based on whether the reconstruction error is greater than a preset judgment threshold. If there is a communication network abnormality, a data communication warning is sent to the user. If there is no communication network abnormality, the data to be uploaded is sent to the user normally.

[0060] Among them, the target ship is a ship that needs to be remotely monitored, and the target ship equipment is the equipment on the target ship that needs to be remotely monitored. The target ship equipment includes but is not limited to the main engine, auxiliary engine, boiler, emergency generator, emergency air compressor and various sensors of the target ship. The present invention does not specifically limit the model and installation location of the monitoring device. The user sets it according to actual needs. How to select a suitable model of the monitoring device and how to set the installation location of the monitoring device are contents that are easy to understand for technicians in this field and will not be elaborated here.

[0061] Specifically, for an operation data of a single target ship equipment, if the operation data is not within the corresponding reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operation data, and the associated abnormal state includes:

[0062] A type of associated abnormal state in which the proportion of abnormal associated data in the operating data is greater than the preset proportion of abnormal associated data;

[0063] The second type of associated abnormal state where the proportion of abnormal associated data in the operating data is less than or equal to the preset abnormal associated data proportion;

[0064] Each operating data is individually set with a corresponding reference preset range.

[0065] Among them, the present invention applies a cyclic monitoring cycle, and the duration of the monitoring cycle can be determined by the user. The higher the user's requirements for ship safety, the shorter the duration of the monitoring cycle. A monitoring cycle duration is provided, and the monitoring cycle is 5 minutes. At the end of each monitoring cycle, a determination is made as to whether the operating data is within the corresponding reference preset range.

[0066] The reference preset range is determined according to each operating data of the data reference record, the data reference record is a historical record that meets the user's requirements for ship safety, the average value of the minimum values ​​of each operating data in the data reference record is recorded as the minimum value of the reference preset range of the corresponding operating data, and the average value of the maximum values ​​of each operating data in the data reference record is recorded as the maximum value of the reference preset range of the corresponding operating data;

[0067] For any operation data, the abnormal correlation data ratio = (the number of abnormal correlation data of the operation data / the number of correlation data of the operation data) × 100%. The abnormal correlation data is the correlation data that is not within the corresponding reference preset range. The value of the preset abnormal correlation data ratio can be set by the user according to actual needs and historical records. The higher the user's requirements for ship safety, the smaller the value of the preset abnormal correlation data ratio. A value of the preset abnormal correlation data ratio is provided, and the value of the preset abnormal correlation data ratio is 50%;

[0068] For any operating data, the associated data is operating data that has an associated relationship with the operating data. The associated relationship includes, but is not limited to, two operating data are used to characterize the same performance of the same equipment, two operating data belong to the same equipment and one operating data directly or indirectly affects the other operating data, and two operating data belong to different equipment and one operating data directly or indirectly affects the other operating data. Several groups of operating data with an associated relationship are provided herein, for example, the real-time speed of the main engine and the real-time power output of the main engine, the ship speed and the ship fuel consumption rate, and the cabin temperature and the hull structure stress.

[0069] Specifically, the abnormal data extraction method is determined according to the associated abnormal state of the operation data, and the abnormal data extraction method includes:

[0070] If the operation data is in a related abnormal state, the operation data and its related data are set as abnormal operation data;

[0071] If the operation data is in the second-class associated abnormal state, only the operation data is set as abnormal operation data.

[0072] Specifically, the monitoring images of each monitoring device are obtained, and the risk factor detection and analysis are performed on each monitoring device respectively. In the risk factor detection and analysis of a single monitoring device, an image extraction strategy is determined according to the risk factor of each monitoring image corresponding to the monitoring device. The image extraction strategy includes:

[0073] If the risk factor is within a first preset risk factor range, determining a video extraction method according to the risk factor of the monitoring image and its related monitoring images and the quality of the monitoring images;

[0074] If the risk factor is within a second preset risk factor range, it is determined whether to adjust the preset video extraction time length of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image.

[0075] Among them, according to the above-mentioned monitoring cycle, the monitoring image of each monitoring device is subjected to risk factor detection and analysis, and the monitoring image is obtained at the same interval in each monitoring cycle, and the interval time is the preset extraction time. The preset extraction time can be set by the user according to actual needs and historical records. The higher the user's requirements for ship safety, the shorter the preset extraction time. A method for determining the value of the preset extraction time is provided, and the average value of the time intervals for each acquisition of the monitoring image of the monitoring device in the historical records that meet the user's requirements for ship safety is recorded as the preset extraction time. A value for the preset extraction time is provided, and the value of the preset extraction time is 10s;

[0076] The risk factor detection analysis is to detect the risk factor of each monitoring image collected by each monitoring device. The risk factor detection process includes extracting behavior features for each monitoring image, comparing the extracted behavior features with each dangerous behavior feature in a preset dangerous behavior feature set, and calculating the dangerous behavior feature similarity. If the dangerous behavior feature similarity is greater than the preset feature similarity, the risk factor = (dangerous behavior feature similarity - preset feature similarity) / preset feature similarity. If the dangerous behavior feature similarity is less than the preset feature similarity, the risk factor of the monitoring image is recorded as 0. The user can set the value of the preset feature similarity according to actual needs and historical records. The higher the user's requirements for ship safety, the smaller the value of the preset feature similarity. A method for setting the value of the preset feature similarity is provided, and the minimum value of the dangerous behavior feature similarity of each abnormal detection image in the historical record that meets the user's requirements for ship safety is recorded as the preset feature similarity.

[0077] The process of extracting behavioral features and determining the similarity of dangerous behavioral features for each monitoring image includes:

[0078] Preprocess each monitoring image through contrast enhancement, noise removal, and grayscale conversion;

[0079] Perform edge detection and feature point detection on the preprocessed image, and perform multi-feature fusion on features of different types (such as shape, texture, and motion) to obtain behavioral features;

[0080] The extracted features are converted into numerical vectors and normalized to have a length of 1;

[0081] The cosine similarity is obtained by calculating the dot product of two normalized vectors and then dividing by the product of the moduli of the two vectors;

[0082] The value range of the cosine similarity is [-1, 1]. It can be understood that a similarity close to 1 indicates that the directions of the two vectors are very similar, that is, the images are similar, and a similarity close to -1 indicates that the directions of the two vectors are opposite, that is, the images are not similar.

[0083] How to extract behavioral features and calculate feature similarity for monitoring images is easy to understand for those skilled in the art and will not be elaborated here;

[0084] The values ​​within the first preset risk factor range are all greater than the first preset risk factor and are all less than or equal to the second preset risk factor, and the values ​​within the second preset risk factor range are all greater than the second preset risk factor. The user can set the values ​​of the first preset risk factor and the second preset risk factor according to actual needs and historical records. The higher the user's requirements for ship safety, the smaller the value of the first preset risk factor and the smaller the value of the second preset risk factor. A method for determining the value of the first preset risk factor is provided, and the minimum value of the risk factor of each abnormal monitoring image in the historical record that meets the user's requirements for ship safety is recorded as the first preset risk factor. A method for determining the value of the second preset risk factor is provided, and the average value of the risk factor of each abnormal monitoring image in the historical record that meets the user's requirements for ship safety is recorded as the second preset risk factor. Monitoring images with a risk factor less than or equal to the first preset risk factor are recorded as safe monitoring images, and the safe monitoring images are stored in a storage device in the target ship. Monitoring images with corresponding content are sent according to user instructions under safety warning conditions. The safety warning condition is that the user needs a more complete monitoring video to verify dangerous behavior in the target ship.

[0085] The user can set the value of the preset video extraction time according to actual needs and historical records. The higher the user's requirement for video integrity, the larger the value of the preset video extraction time. A method for setting the value of the preset video extraction time is provided, and the average value of the video duration of each abnormal monitoring video in the historical records that meet the user's requirement for video integrity is set as the preset video extraction time. A value of the preset video extraction time is provided, and the value of the preset video extraction time is 30s.

[0086] Specifically, if the risk factor is within the first preset risk factor range, a video extraction method is determined according to the risk factors corresponding to the monitoring image and the related monitoring image and the quality of the monitoring image. The video extraction method includes:

[0087] If there is a related monitoring image with a risk factor greater than a first preset risk factor among the related monitoring images of the monitoring image, an abnormal monitoring image is determined according to the image quality reference value of the monitoring image and its related monitoring images, and an abnormal monitoring video with a preset video extraction time is extracted for the abnormal monitoring image;

[0088] If there is no related monitoring image with a risk factor greater than the first preset risk factor among the related monitoring images of the monitoring image, abnormal monitoring video extraction is not performed for the monitoring image;

[0089] The abnormal monitoring image is an image corresponding to the maximum value of the image quality reference value in the monitoring image and its related monitoring images, the related monitoring image is a monitoring image collected by the relevant monitoring device at the monitoring time corresponding to the corresponding dangerous image, and the relevant monitoring device is a monitoring device whose monitoring range overlaps with the monitoring device that took the corresponding dangerous image greater than the preset monitoring range overlap.

[0090] The image quality reference value is determined according to the resolution reference value and the signal-to-noise ratio reference value of each monitoring image. , b is the resolution reference value of each monitoring image, θ1 is the influence coefficient corresponding to the resolution reference value, z is the signal-to-noise ratio reference value of each monitoring image, θ2 is the influence coefficient corresponding to the signal-to-noise ratio reference value, the resolution reference value is the number of pixels of each monitoring image, the signal-to-noise ratio reference value is the ratio of the signal intensity of each monitoring image to the background noise intensity, the user can set the influence coefficient corresponding to the resolution reference value and the signal-to-noise ratio reference value according to actual needs, the higher the user's requirements for image clarity, the greater the value of the influence coefficient corresponding to the resolution reference value, the higher the user's requirements for the existence of noise in the image, the greater the value of the influence coefficient corresponding to the signal-to-noise ratio reference value, a resolution reference value and a value of the influence coefficient corresponding to the signal-to-noise ratio reference value are provided, the value of the influence coefficient corresponding to the resolution reference value is 0.5, and the value of the influence coefficient corresponding to the signal-to-noise ratio reference value is 0.5;

[0091] When extracting an abnormal monitoring video of a preset video extraction time from an abnormal monitoring image, the acquisition time corresponding to the abnormal monitoring image is taken as the midpoint of the time axis of the abnormal monitoring video, the extraction start time of the abnormal monitoring video = the acquisition time corresponding to the abnormal monitoring image - (preset video extraction time / 2), the extraction interpretation time of the abnormal monitoring video = the acquisition time corresponding to the abnormal monitoring image + (preset video extraction time / 2);

[0092] For any two monitoring devices, the corresponding monitoring range overlap = the area of ​​the overlapping monitoring areas of the two monitoring devices / the area of ​​the monitoring area of ​​the monitoring device corresponding to the dangerous image. The monitoring area is the effective area that the monitoring device can capture, and the effective area is the horizontal area corresponding to the space that the monitoring device can clearly capture. It can be understood that whether the monitoring device can clearly capture the space is determined by the accuracy of its own equipment. The user can determine the effective area of ​​the monitoring device based on his or her own professional experience. The overlapping monitoring area is the overlapping part of the monitoring areas of the two monitoring devices. The user can set the value of the preset monitoring range overlap according to actual needs and historical records. The higher the user's requirements for the validity of the acquired abnormal monitoring video, the greater the value of the preset monitoring range overlap. A method for determining the value of the preset monitoring range overlap is provided. The average value of the monitoring range overlap of each relevant monitoring device in the historical records that meet the user's requirements for the validity of the acquired abnormal monitoring video is recorded as the preset monitoring range overlap. A value of the preset monitoring range overlap is provided, and the value of the preset monitoring range overlap is 0.2.

[0093] Specifically, if the risk factor is within the second preset risk factor range, determining whether to adjust the preset video extraction time length for the monitoring image according to the risk factor of the monitoring image of the adjacent frame of the monitoring image;

[0094] If there is an adjacent frame monitoring image whose risk factor is greater than the first preset risk factor in the monitoring image, the preset video extraction time length of the monitoring image is increased and adjusted;

[0095] The increase in the preset video extraction time is positively correlated with the risk factor of the adjacent frame monitoring image.

[0096] Wherein, monitoring images of each monitoring device are acquired periodically, and the time interval between the acquisition moments corresponding to any monitoring image and its adjacent frame monitoring images is a preset image monitoring period.

[0097] Specifically, the data transmission status is determined according to the data parameters and detail parameters of the data to be uploaded. The data transmission status includes:

[0098] A first preset transmission state in which the data parameter is greater than a preset data parameter and the detail parameter is less than or equal to a preset detail parameter;

[0099] A second preset transmission state in which the data parameter is less than or equal to the preset data parameter or the detail parameter is greater than the preset detail parameter;

[0100] The data to be uploaded includes abnormal operation data and abnormal monitoring video;

[0101] The data parameters are determined according to the data transmission volume and the data transmission frequency of the data to be uploaded, and the detail parameters are determined according to the low-frequency components of the data to be uploaded.

[0102] Among them, the data parameters , m is the data transmission volume of the data to be uploaded, f is the data transmission frequency of the data transmission network, and the detailed parameters , d is the proportion of low-frequency components of the data to be uploaded, d0 is the proportion of reference low-frequency components, the values ​​of the preset data parameters and the preset detail parameters can be set by the user according to actual needs and historical records, the higher the user's real-time requirements for the data transmission process, the smaller the value of the preset data parameter, the higher the user's reliability requirements for communication security monitoring, the smaller the value of the preset detail parameter, and a method for taking the value of the preset data parameter is provided, the historical records of using a multi-feature model to monitor the communication network are recorded as data reference records, and the maximum value of the data parameters of each time the data to be uploaded in the data reference record that meets the user's real-time requirements for the data transmission process is recorded as a preset data reference value, and a method for taking the value of the preset detail reference value is provided, the historical records of using a single-feature model to monitor the communication network are recorded as detail reference records, and the maximum value of the detail parameters of each time the data to be uploaded in the detail reference record that meets the user's reliability requirements for communication security monitoring is recorded as a preset detail reference value.

[0103] Specifically, the abnormality monitoring method is determined according to the data transmission status of the data to be uploaded, and the abnormality monitoring method includes:

[0104] The data to be uploaded is in a first preset transmission state, and a single feature model is used to monitor the communication network;

[0105] The data to be uploaded is in a second preset transmission state, and a multi-feature model is used to monitor the communication network.

[0106] Among them, the difference between the process of constructing a single-feature model and a multi-feature model in the present invention is that the multi-feature model selects multiple observation scales and observation steps corresponding to each observation scale according to actual conditions during the model construction process, and extracts traffic features for the data flow based on the selected multiple observation scales and observation steps corresponding to each observation scale, so as to obtain traffic features at multiple scales, and performs feature fusion on the traffic features obtained at each scale, and constructs a multi-feature model based on the traffic features after feature fusion. The single-feature model uses a single observation scale and observation step to collect traffic features for the data during the model construction process, and constructs a single-feature model based on the traffic features of the collected single scale. How to construct a model based on the traffic features of the acquired single scale or multiple scales is easy for technicians in this field to understand and will not be elaborated here.

[0107] When using a multi-feature model to monitor a communication network, it is necessary to divide the currently transmitted traffic data sequence according to the observation scales determined during the construction of the multi-feature model and the observation step corresponding to each observation scale, obtain multiple sub-traffic data sequences, and reconstruct each sub-traffic data sequence. The reconstruction process uses wavelet transform technology to extract traffic features from the reconstructed sequence at each observation scale using the constructed multi-feature model. The reconstruction error is determined based on the error value between the original traffic data sequence at each observation scale and the reconstructed sequence at the corresponding observation scale and the weight coefficient corresponding to each observation scale. , is the absolute value of the difference between the original traffic data sequence and the reconstructed sequence at the u-th observation scale, is the weight coefficient corresponding to the u-th observation scale. Users can set the weight coefficient corresponding to each observation scale according to actual needs. The single feature model is used to monitor the communication network. The reconstruction error is the absolute value of the difference between the original traffic data sequence and the reconstructed sequence.

[0108] Specifically, the observation scale of the multi-feature model is set according to the proportion of the reference low-frequency component;

[0109] The proportion of one type of observation scale in the observation scale of the multi-feature model is positively correlated with the proportion of the reference low-frequency component;

[0110] The first type of observation scale is an observation scale with a time span greater than a preset time span.

[0111] Among them, the reference low-frequency component ratio is the average value of the low-frequency component ratio of each transmission data in the reference record, the reference record is a flow monitoring record that meets the user's reliability requirements for communication security monitoring, the low-frequency component ratio of each transmission data = the energy value of the low-frequency component of each transmission data / the total energy value of the corresponding transmission data, and the low-frequency component is extracted according to the wavelet transform technology, which is easy for technicians in this field to understand and will not be repeated here;

[0112] The observation scale of the multi-feature model includes a first-class observation scale and a second-class observation scale, the second-class observation scale is an observation scale whose time span is less than or equal to a preset time span, and the proportion of the first-class observation scale = the number of the first-class observation scale / the total number of observation scales;

[0113] The user can set the value of the preset time span according to actual needs and historical records. The higher the user's requirements for the reliability of communication security monitoring, the smaller the value of the preset time span. A method for setting the value of the preset time span is provided, and the average value of the time spans of each observation scale in the historical records that meet the user's requirements for the reliability of communication security monitoring is recorded as the preset time span.

[0114] Specifically, the reconstruction error is compared with a preset judgment threshold, and whether there is a communication network abnormality is determined based on whether the reconstruction error is greater than the preset judgment threshold. If the reconstruction error is greater than the preset judgment threshold, the communication network is determined to be abnormal and a data communication warning is sent to the user.

[0115] Among them, the user can set the value of the preset judgment threshold according to actual needs and historical records. The higher the user's requirement for the reliability of communication security monitoring, the smaller the value of the preset judgment threshold. A method for setting the value of the preset judgment threshold is provided, and the average value of each reconstruction error in the historical record that meets the user's requirement for the reliability of communication security monitoring is recorded as the preset judgment threshold.

[0116] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0117] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for remote monitoring of ship status based on data analysis, characterized in that: include: Continuously collect the operating data of each target ship equipment; For a single target ship equipment, if the operating data is not within the reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operating data, and the abnormal data extraction method is determined according to the associated abnormal state of the operating data. The abnormal data extraction method is to set both the operating data and its associated data as abnormal operating data or only the operating data as abnormal operating data; Perform risk factor detection and analysis on each monitoring device respectively, and determine an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device. The image extraction strategy is to determine a video extraction method according to the risk factors corresponding to the monitoring image and the related monitoring images and the quality of the monitoring images, or to determine whether to adjust the preset video extraction time of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image; Obtaining the data transmission volume of the data to be uploaded and the data transmission frequency of the data transmission network, and determining the data transmission status according to the data parameters and detail parameters of the data to be uploaded; Determine an abnormality monitoring method according to the data transmission status of the data to be uploaded, and the abnormality monitoring method is to use a single feature model or a multi-feature model to monitor the communication network; Determine whether there is a communication network anomaly based on whether the reconstruction error is greater than a preset judgment threshold. If there is a communication network anomaly, send a data communication warning to the user. If there is no communication network anomaly, send the data to be uploaded to the user normally. Data parameters , m is the data transmission volume of the data to be uploaded, f is the data transmission frequency of the data transmission network, and the detailed parameters , d is the proportion of low-frequency components of the data to be uploaded, and d0 is the proportion of reference low-frequency components.

2. The method for remote monitoring of ship status based on data analysis according to claim 1, characterized in that: For an operation data of a single target ship equipment, if the operation data is not within the corresponding reference preset range, the associated abnormal state is determined according to the abnormal associated data ratio of the operation data, and the associated abnormal state includes: A type of associated abnormal state in which the proportion of abnormal associated data in the operating data is greater than the preset proportion of abnormal associated data; The second type of associated abnormal state where the proportion of abnormal associated data in the operating data is less than or equal to the preset abnormal associated data proportion; Each operating data is individually set with a corresponding reference preset range.

3. The method for remote monitoring of ship status based on data analysis according to claim 2 is characterized in that: Determine the abnormal data extraction method based on the associated abnormal state of the operating data. The abnormal data extraction method includes: If the operation data is in a related abnormal state, the operation data and its related data are set as abnormal operation data; If the operation data is in the second-class associated abnormal state, only the operation data is set as abnormal operation data.

4. The method for remote monitoring of ship status based on data analysis according to claim 1, characterized in that: Acquire monitoring images of each monitoring device, perform risk factor detection and analysis on each monitoring device respectively, and in the risk factor detection and analysis on a single monitoring device, determine an image extraction strategy according to the risk factor of each monitoring image corresponding to the monitoring device, the image extraction strategy includes: If the risk factor is within a first preset risk factor range, determining a video extraction method according to the risk factor of the monitoring image and its related monitoring images and the quality of the monitoring images; If the risk factor is within a second preset risk factor range, it is determined whether to adjust the preset video extraction time length of the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image.

5. The method for remote monitoring of ship status based on data analysis according to claim 4 is characterized in that: If the risk factor is within the first preset risk factor range, a video extraction method is determined according to the risk factors corresponding to the monitoring image and the related monitoring image and the quality of the monitoring image. The video extraction method includes: If there is a related monitoring image with a risk factor greater than a first preset risk factor among the related monitoring images of the monitoring image, an abnormal monitoring image is determined according to the image quality reference value of the monitoring image and its related monitoring images, and an abnormal monitoring video with a preset video extraction time is extracted for the abnormal monitoring image; If there is no related monitoring image with a risk factor greater than the first preset risk factor among the related monitoring images of the monitoring image, abnormal monitoring video extraction is not performed for the monitoring image; The abnormal monitoring image is an image corresponding to the maximum value of the image quality reference value in the monitoring image and its related monitoring images, the related monitoring image is a monitoring image collected by the relevant monitoring device at the monitoring time corresponding to the corresponding dangerous image, and the relevant monitoring device is a monitoring device whose monitoring range overlaps with the monitoring device that took the corresponding dangerous image greater than the preset monitoring range overlap.

6. The method for remote monitoring of ship status based on data analysis according to claim 5 is characterized in that: If the risk factor is within a second preset risk factor range, determining whether to adjust the preset video extraction time length for the monitoring image according to the risk factors of the monitoring images of adjacent frames of the monitoring image; If there is an adjacent frame monitoring image whose risk factor is greater than the first preset risk factor in the monitoring image, the preset video extraction time length of the monitoring image is increased and adjusted; The increase in the preset video extraction time is positively correlated with the risk factor of the adjacent frame monitoring image.

7. The method for remote monitoring of ship status based on data analysis according to claim 6, characterized in that: The data transmission status is determined according to the data parameters and detail parameters of the data to be uploaded. The data transmission status includes: A first preset transmission state in which the data parameter is greater than a preset data parameter and the detail parameter is less than or equal to a preset detail parameter; A second preset transmission state in which the data parameter is less than or equal to the preset data parameter or the detail parameter is greater than the preset detail parameter; The data to be uploaded includes abnormal operation data and abnormal monitoring video; The data parameters are determined according to the data transmission volume and the data transmission frequency of the data to be uploaded, and the detail parameters are determined according to the low-frequency components of the data to be uploaded.

8. The method for remote monitoring of ship status based on data analysis according to claim 7, characterized in that: The abnormality monitoring method is determined according to the data transmission status of the data to be uploaded, and the abnormality monitoring method includes: The data to be uploaded is in a first preset transmission state, and a single feature model is used to monitor the communication network; The data to be uploaded is in a second preset transmission state, and a multi-feature model is used to monitor the communication network.

9. The method for remote monitoring of ship status based on data analysis according to claim 8, characterized in that: The observation scale of the multi-feature model is set according to the proportion of the reference low-frequency component; The proportion of one type of observation scale in the observation scale of the multi-feature model is positively correlated with the proportion of the reference low-frequency component; The first type of observation scale is an observation scale with a time span greater than a preset time span.

10. The method for remote monitoring of ship status based on data analysis according to claim 9, characterized in that: The reconstruction error is compared with the preset judgment threshold, and whether there is a communication network abnormality is determined based on whether the reconstruction error is greater than the preset judgment threshold. If the reconstruction error is greater than the preset judgment threshold, the communication network is determined to be abnormal and a data communication warning is sent to the user.

Images