An intelligent detection method, device and storage medium for a motorboat

Abstract

The present application discloses a method, device and storage medium for intelligent detection of motorboats, which are used to improve the accuracy of motorboat detection. The intelligent detection method of motorboats disclosed in the present application includes: detecting the boats appearing in the monitoring video according to a set boat detection model; extracting boat features from the boats appearing in the monitoring video, where the boat features include boat position, hull color and boat size features; tracking the trajectory of the boats according to the boat features to obtain the trajectory information of the boats; obtaining boat splash information according to a set splash area detection model; and detecting whether the boats are motorboats according to the trajectory information and boat splash information of the boats. The present application also provides an intelligent detection device and storage medium for motorboats.

Description

Technical Field

[0001] The present application relates to the field of computing technology, and in particular to a method, device and storage medium for intelligent detection of motorboats. Background Art

[0002] Large motorboats (also called big boats) that have been modified by adding engines, etc., have a sailing speed of up to 60 knots (about 110 kilometers per hour), which seriously affects the normal order of water production, operations, and transportation, and threatens the life and property safety of water transportation and operators. They are water transportation tools that are expressly prohibited by relevant national departments. At present, in order to severely crack down on and curb the "big boats" in the waters, the public security, maritime police and other departments mainly adopt the smart coastal defense front-end perception system combined with 24-hour uninterrupted patrol and investigation to maintain the security environment in the waters.

[0003] However, the existing technology has a large number of false positives and missed positives in the detection of large motorboats. Summary of the invention

[0004] In response to the above technical problems, the embodiments of the present application provide a motorboat intelligent detection method, device and storage medium for intelligently detecting motorboats and improving the accuracy of motorboat detection.

[0005] In a first aspect, an embodiment of the present application provides a motorboat intelligent detection method, comprising:

[0006] Detect the ships appearing in the surveillance video according to the set ship detection model;

[0007] Extracting ship features from ships appearing in surveillance videos, where the ship features include ship location, ship color, and ship size features;

[0008] Tracking the trajectory of the vessel according to the characteristics of the vessel to obtain trajectory information of the vessel;

[0009] The ship splash information is obtained according to the set splash area detection model;

[0010] Whether the vessel is a motorboat is detected according to the track information and the water splash information of the vessel.

[0011] Preferably, the set ship detection model includes:

[0012] Collect image data of motorboats and passing ships in the monitored waters, and mark the types of ships, including motorboats and non-motorboats;

[0013] The labeled ship image data is trained based on the YOLO model to obtain the set ship detection model.

[0014] Preferably, in the present invention, vessel features are extracted from the monitored videos, where the vessel features include vessel position, hull color, and vessel size features, including:

[0015] The method for extracting the vessel size features includes:

[0016] If personnel information is detected in the monitoring area, the size information (w1, h1) of the personnel and the monitoring area information (x2, y2, w2, h2) are extracted, and the vessel size features are calculated as follows:

[0017]

[0018] where r1 is the vessel size feature, x2 is the abscissa of the upper left fixed point of the vessel box, y2 is the ordinate of the upper left fixed point of the vessel box, w2 is the width of the vessel box, h2 is the height of the vessel box, w1 is the width of the personnel box, h1 is the height of the personnel box, and c1 is the first correction constant obtained from historical data training;

[0019] If no personnel information is detected in the detection area, the monitoring area information (x2, y2, w2, h2) is extracted, and the vessel size features are calculated as follows:

[0020]

[0021] where c2 is the second correction constant obtained from historical data training, and c3 is the third correction constant obtained from historical data training.

[0022] Furthermore, the trajectory tracking of the vessel based on the vessel features to obtain the trajectory information of the vessel includes:

[0023] Performing trajectory tracking on the vessel according to the position feature, hull color, and vessel size feature of the vessel to obtain the trajectory information {(lx i , ly i )} of the vessel,

[0024] where (lx i , ly i ) is the vessel position feature, and:

[0025] (lx i , ly i ) = (x2 + 0.5w2, y2 + 0.5h2), and i represents the frame number of the video.

[0026] Preferably, the set splash area detection model includes:

[0027] Collecting the splash data when the vessel passes by and annotating the splash;

[0028] The water splash area detection model is obtained by training the labeled water splash data based on the YOLO model.

[0029] Specifically, obtaining the ship water splash information according to the set water splash area detection model includes:

[0030] For any frame i of the picture, use the water splash area detection model to detect the set of water splash areas in the picture {(qx j , qy j , qw j , qh j )};

[0031] Calculate the matching score gq of each water splash area j ;

[0032] Select the water splash area with the maximum matching score gq j as the ship water splash information of the ship in the i-th frame of the picture (wx i , wy i , ww i , wh i );

[0033] Among them,

[0034] gq j = g1 j (c4g2 j + g3 j )

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041] Among them, ts1 is the set first judgment threshold, ts2 is the set second judgment threshold, ts3 is the set third judgment threshold, ts4 is the set fourth judgment threshold, ts5 is the set fifth judgment threshold, ts6 is the set sixth judgment threshold, and c4 is the fourth correction constant obtained by training historical data;

[0042] qx j is the abscissa of the upper left corner fixed point of the water splash area, qy j is the ordinate of the upper left corner fixed point of the water splash area, qw jIt is the width of the water splash area, qh j It is the height of the water splash area, and j is the number of the water splash area;

[0043] wx i It is the abscissa of the upper left fixed point of the water splash area, wy i It is the ordinate of the upper left fixed point of the water splash area, ww i It is the width of the water splash area, wh i It is the height of the water splash area, and i is the frame number of the monitoring screen.

[0044] Preferably, the detection of whether the ship is a motorboat according to the trajectory information and ship water splash information of the ship includes:

[0045] According to the trajectory information of the ship {(lx i , ly i )} and the ship water splash information {(wx i , wy i , ww i , wh i ), the number of trajectory points N is counted, and the ship speed score g is calculated v ;

[0046] When the ship speed score g v is greater than the set seventh judgment threshold ts7, it is judged that the detected ship is a motorboat;

[0047] Among them,

[0048]

[0049]

[0050]

[0051] c5 is the fifth correction constant obtained by training historical data, and c6 is the sixth correction constant obtained by training historical data.

[0052] Furthermore, after it is judged that the detected ship is a motorboat, it further includes:

[0053] The early warning information is sent to the management department in real time through a preset communication device.

[0054] In a second aspect, the embodiment of the present application further provides a motorboat intelligent detection device, including:

[0055] A ship detection module configured to detect ships appearing in a monitoring video according to a set ship detection model;

[0056] A vessel feature extraction module, configured to extract vessel features of vessels appearing in a surveillance video, where the vessel features include vessel position, hull color, and vessel size features;

[0057] A vessel trajectory tracking module, configured to track the trajectory of the vessel according to the vessel features to obtain the trajectory information of the vessel;

[0058] A vessel splash detection module, configured to obtain vessel splash information according to a set splash area detection model;

[0059] A judgment module, configured to detect whether the vessel is a speedboat according to the trajectory information and the vessel splash information of the vessel.

[0060] In a third aspect, an embodiment of the present application further provides a speedboat intelligent detection device, including: a memory, a processor, and a user interface;

[0061] The memory is used to store a computer program;

[0062] The user interface is used to interact with the user;

[0063] The processor is used to read the computer program in the memory. When the processor executes the computer program, the speedboat intelligent detection method provided by the present invention is implemented.

[0064] In a fourth aspect, an embodiment of the present application further provides a processor-readable storage medium. The processor-readable storage medium stores a computer program. When the processor executes the computer program, the speedboat intelligent detection method provided by the present invention is implemented.

[0065] Using the method of the present invention, first, speedboat picture data and picture data of passing vessels in a monitored water area are collected, the vessel types are labeled, and a set vessel detection model is obtained by training the labeled vessel picture data based on the YOLO model; splash data when the vessel passes by is collected, and the splashes are labeled, and the splash area detection model is obtained by training the labeled splash data based on the YOLO model; then, the vessels appearing in the surveillance video are detected according to the vessel detection model, and the vessel splash information is obtained according to the splash area detection model; finally, it is detected whether the vessel is a speedboat according to the trajectory information and the vessel splash information of the vessel. Using the method of the present invention, intelligent detection of speedboats can be achieved, the detection accuracy can be improved, and false alarms and missed detections of speedboat detection can be reduced. Description of the Drawings

[0066] To more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0067] Figure 1 Schematic flowchart of the intelligent detection method for motorboats provided by the embodiments of the present application;

[0068] Figure 2 Schematic flowchart of the method for extracting the size characteristics of a ship provided by the embodiments of the present application;

[0069] Figure 3 Schematic diagram of the intelligent detection device for motorboats provided by the embodiments of the present application;

[0070] Figure 4 Schematic diagram of the structure of another intelligent detection device for motorboats provided by the embodiments of the present application. Detailed implementation manners

[0071] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0072] The following explains some terms appearing in the text:

[0073] 1. In the embodiments of the present invention, the term "and / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0074] 2. In the embodiments of the present application, the term "plurality" refers to two or more, and other quantifiers are similar.

[0075] 3. YOLO model: Abbreviation for You Only Look Once, which is an object detection model.

[0076] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0077] It should be noted that the display order of the embodiments of the present application only represents the sequence of the embodiments, and does not represent the superiority or inferiority of the technical solutions provided by the embodiments.

[0078] Embodiment 1

[0079] See Figure 1 , a schematic diagram of an intelligent detection method for a motorboat provided by an embodiment of the present application, as Figure 1 shown, this method includes steps S101 to S103:

[0080] S101. Detect the boats appearing in the surveillance video according to the set boat detection model;

[0081] As a preferred example, in this step, the set boat detection model is determined in the following manner:

[0082] Collect motorboat picture data and picture data of passing boats in the monitored water area, and label the boat types, where the boat types include motorboats and non-motorboats;

[0083] Train the labeled boat picture data based on the YOLO model to obtain the set boat detection model.

[0084] It should be noted that in the process of determining the boat detection model in this embodiment, after collecting motorboat picture data and picture data of passing boats in the monitored water area and labeling the boat types, it can be trained based on the YOLO model, or it can be trained based on other models, such as R-CNN, VovNet, DenseNet, which is not limited in the present invention.

[0085] S102. Extract boat features from the boats appearing in the surveillance video, where the boat features include boat position, hull color, and boat size features;

[0086] In this step, as Figure 2 shown, the method for extracting the boat size features includes:

[0087] S201. Determine whether personnel information is detected in the monitoring area. If so, execute S202, that is, extract the boat size features by Method 1, otherwise execute S203, that is, extract the boat size features by Method 2. Specifically, Method 1 is:

[0088] Extract the size information (w1, h1) of the person and the monitoring area information (x2, y2, w2, h2), and calculate the ship size feature in the following way:

[0089]

[0090] Among them, r1 is the ship size feature, x2 is the abscissa of the upper left corner of the ship box, y2 is the ordinate of the upper left corner of the ship box, w2 is the width of the ship box, h2 is the height of the ship box, w1 is the width of the person box, h1 is the height of the person box, and c1 is the first correction constant obtained from historical data training;

[0091] The second method is:

[0092] Extract the monitoring area information (x2, y2, w2, h2), and calculate the ship size feature in the following way:

[0093]

[0094] Among them, c2 is the second correction constant obtained from historical data training, and c3 is the third correction constant obtained from historical data training.

[0095] That is to say, in this step, in the monitored water area, use the ship detection model to detect the ships appearing in the monitoring video, and according to the detected ship types, extract the ship images of the ships. Perform feature extraction on the ships appearing in the monitoring video, extract the ship position, hull color and ship size feature, and the detection method of the ship size feature is as follows:

[0096] Detect the pedestrian information in the ship area (x2, y2, w2, h2), where x2 is the abscissa of the upper left corner of the ship box, y2 is the ordinate of the upper left corner of the ship box, w2 is the width of the ship box, and h2 is the height of the ship box;

[0097] When person information is detected in the area, extract the size information (w1, h1) of the person box, where w1 is the width of the person box and h1 is the height of the person box, and calculate the ship size feature:

[0098]

[0099] Among them, c1 is the first correction constant obtained from historical data training;

[0100] When no person information is detected in the area, calculate the ship size feature:

[0101]

[0102] Among them, c2 is the second correction constant obtained by training historical data, and c3 is the third correction constant obtained by training historical data;

[0103] S103. Perform trajectory tracking on the vessel according to the vessel characteristics to obtain the trajectory information of the vessel;

[0104] In this step, the performing trajectory tracking on the vessel according to the vessel characteristics to obtain the trajectory information of the vessel includes:

[0105] Perform trajectory tracking on the vessel according to the position characteristics, hull color, and vessel size characteristics of the vessel to obtain the trajectory information of the vessel {(lx i , ly i )},

[0106] wherein, (lx i , ly i ) are the position characteristics of the vessel, and:

[0107] (lx i , ly i ) = (x2 + 0.5w2, y2 + 0.5h2), and i represents the frame number of the video.

[0108] It should be noted that the trajectory information of the vessel {(lx i , ly i )} is a set of the position characteristics (lx i , ly i ) of the vessel, including all the position characteristics of the collected vessel. For example, if 1000 frames of images are collected, that is, the frame number of the video is 1000, then for each frame of image, the position characteristics (lx i , ly i ) of the vessel in each frame of image are calculated according to the vessel area (x2, y2, w2, h2) of each frame of image. After all 1000 frames of images are calculated, the set composed of 1000 vessel position characteristics is the trajectory information of the vessel.

[0109] S104. Obtain the vessel splash information according to the set splash area detection model;

[0110] In this step, collect the splash data when the motorboat passes by and label the splashes;

[0111] Train the labeled splash data based on the YOLO model to obtain the splash area detection model.

[0112] It should be noted that in the process of determining the water splash area detection model in this embodiment, after collecting the water splash data when the motorboat passes by and annotating the water splashes, it can be trained based on the YOLO model or other models, such as R-CNN, VovNet, DenseNet, which is not limited in the present invention.

[0113] In this step, the obtaining of the ship water splash information according to the set water splash area detection model includes:

[0114] For any frame i, use the water splash area detection model to detect the set of water splash areas in the frame {(qx j , qy j , qw j , qh j )};

[0115] Calculate the matching score gq of each water splash area j ;

[0116] Select the water splash area with the maximum matching score gq j as the ship water splash information (wx i , wy i , ww i , wh i ) of the ship in the i-th frame;

[0117] Among them,

[0118] gq j = g1 j (c4g2 j + g3 j )

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125] Among them, ts1 is the set first judgment threshold, ts2 is the set second judgment threshold, ts3 is the set third judgment threshold, ts4 is the set fourth judgment threshold, ts5 is the set fifth judgment threshold, ts6 is the set sixth judgment threshold, and c4 is the fourth correction constant obtained by training historical data;

[0126] qx j is the abscissa of the upper-left fixed point of the water splash area, qy j is the ordinate of the upper-left fixed point of the water splash area, qw j is the width of the water splash area, qh j is the height of the water splash area, and j is the number of the water splash area;

[0127] wx i is the abscissa of the upper-left fixed point of the water splash area, wy i is the ordinate of the upper-left fixed point of the water splash area, ww i is the width of the water splash area, wh i is the height of the water splash area, and i is the frame number of the monitoring screen.

[0128] That is to say, in this step, the trajectory of the ship is tracked according to the position characteristics, hull color and ship size characteristics of the ship, and the trajectory information of the ship {(lx i , ly i )} is obtained. The water splash data passed by the ship is collected, the water splashes are labeled, and the labeled water splash data is trained based on the YOLO model to obtain a water splash area detection model.

[0129] For any frame i of the screen, use the water splash area detection model to detect the set of water splash areas in the screen {(qx j , qy j , qw j , qh j )}, where qx j is the abscissa of the upper-left fixed point of the water splash area, qy j is the ordinate of the upper-left fixed point of the water splash area, qw j is the width of the water splash area, qh j is the height of the water splash area, and j is the number of the water splash area.

[0130] Calculate the matching score of each water splash area:

[0131] gq j = g1 j (c4g2 j + g3 j )

[0132] Where:

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139] ts1 is the set first judgment threshold, ts2 is the set first judgment threshold, ts3 is the set third judgment threshold, ts4 is the set fourth judgment threshold, ts5 is the set fifth judgment threshold, ts6 is the set sixth judgment threshold, and c4 is the fourth correction constant obtained by training historical data.

[0140] Select the matching score gq of the water splash area j The largest water splash area is used as the water splash area of the suspected smuggler ship in the i-th frame (wx i , wy i , ww i , wh i ), where wx i is the abscissa of the fixed point at the upper left corner of the water splash area, wy i is the ordinate of the fixed point at the upper left corner of the water splash area, ww i is the width of the water splash area, and wh i is the height of the water splash area.

[0141] S105. Detect whether the ship is a motorboat according to the trajectory information and ship water splash information of the ship.

[0142] In this step, the detection of whether the ship is a motorboat according to the trajectory information and ship water splash information of the ship includes:

[0143] According to the trajectory information of the ship {(lx i , ly i )} and the ship water splash information {(wx i , wy i , ww i , wh i ), count the number of trajectory points N, and calculate the ship speed score g v ;

[0144] When the ship speed score g v is greater than the set seventh judgment threshold ts7, it is determined that the detected ship is a motorboat;

[0145] Among them,

[0146] g v = g s + g w

[0147]

[0148]

[0149]

[0150] c5 is the fifth correction constant obtained by training with historical data, and c6 is the sixth correction constant obtained by training with historical data.

[0151] As a preferred example, after determining that the detected vessel is a motorboat, it further includes:

[0152] Real-time sending of early warning information to the management department through a preset communication device.

[0153] Using the method of the present invention, first collect motorboat picture data and picture data of passing vessels in the monitored water area, label the vessel types, and train the labeled vessel picture data based on the YOLO model to obtain a set vessel detection model; collect the water splash data when the vessels pass by, and label the water splashes, and train the labeled water splash data based on the YOLO model to obtain the water splash area detection model; then, detect the vessels appearing in the monitoring video according to the vessel detection model, and obtain the vessel water splash information according to the water splash area detection model; finally, detect whether the vessel is a motorboat according to the trajectory information and the vessel water splash information of the vessel. Using the method of the present invention, intelligent detection of motorboats can be achieved, the detection accuracy can be improved, and false alarms and missed alarms in the detection of motorboats can be reduced.

[0154] Embodiment 2

[0155] Based on the same inventive concept, the embodiment of the present invention further provides a motorboat intelligent detection device, as Figure 3 shown, the device includes:

[0156] A vessel detection module 301, configured to detect vessels appearing in the monitoring video according to a set vessel detection model;

[0157] A vessel feature extraction module 302, configured to extract vessel features from vessels appearing in the monitoring video, where the vessel features include vessel position, hull color, and vessel size features;

[0158] A vessel trajectory tracking module 303, configured to track the trajectory of the vessel according to the vessel features to obtain the trajectory information of the vessel;

[0159] A vessel water splash detection module 304, configured to obtain vessel water splash information according to a set water splash area detection model;

[0160] A judgment module 305, configured to detect whether the vessel is a motorboat according to the trajectory information and the vessel water splash information of the vessel.

[0161] As a preferred example, the set vessel detection model is determined in the following manner:

[0162] Collect motorboat picture data and picture data of passing vessels in the monitored water area, and label the vessel types, where the vessel types include motorboats and non-motorboats;

[0163] Train the labeled vessel picture data based on the YOLO model to obtain the set vessel detection model.

[0164] As a preferred example, the set water splash area detection model is determined in the following manner:

[0165] Collect water splash data when vessels pass by, and label the water splashes;

[0166] Train the labeled water splash data based on the YOLO model to obtain the water splash area detection model.

[0167] As a preferred example, the vessel feature extraction module 302 is further configured to extract the vessel size feature according to the following manner:

[0168] If personnel information is detected in the monitoring area, extract the size information (w1, h1) of the personnel and the monitoring area information (x2, y2, w2, h2), and calculate the vessel size feature in the following manner:

[0169]

[0170] where r1 is the vessel size feature, x2 is the abscissa of the upper left vertex of the vessel box, y2 is the ordinate of the upper left vertex of the vessel box, w2 is the width of the vessel box, h2 is the height of the vessel box, w1 is the width of the personnel box, h1 is the height of the personnel box, and c1 is the first correction constant obtained from historical data training;

[0171] If no personnel information is detected in the detection area, extract the monitoring area information (x2, y2, w2, h2), and calculate the vessel size feature in the following manner:

[0172]

[0173] where c2 is the second correction constant obtained from historical data training, and c3 is the third correction constant obtained from historical data training.

[0174] As a preferred example, the vessel trajectory tracking module 303 is further configured to obtain the trajectory information of the vessel in the following manner:

[0175] Perform trajectory tracking on the vessel according to the position feature, hull color, and vessel size feature of the vessel to obtain the trajectory information of the vessel {(lxi ,ly i ),

[0176] where (lx i ,ly i ) is the ship position feature, and:

[0177] (lx i ,ly i ) = (x2 + 0.5w2, y2 + 0.5h2), and i represents the frame number of the video.

[0178] As a preferred example, the ship splash detection module 304 is further configured to obtain ship splash information according to a set splash area detection model:

[0179] For any frame i, use the splash area detection model to detect the set of splash areas in the frame {(qx j ,qy j ,qw j ,qh j )};

[0180] Calculate the matching score gq of each splash area j ;

[0181] Select the splash area with the highest matching score gq j as the ship splash information (wx i ,wy i ,ww i ,wh i ) of the ship in the i-th frame;

[0182] where,

[0183] gq j = g1 j (c4g2 j + g3 j )

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190] Among them, ts1 is the set first judgment threshold, ts2 is the set second judgment threshold, ts3 is the set third judgment threshold, ts4 is the set fourth judgment threshold, ts5 is the set fifth judgment threshold, ts6 is the set sixth judgment threshold, and c4 is the fourth correction constant obtained by training historical data;

[0191] qx j is the abscissa of the upper left fixed point of the water splash area, qy j is the ordinate of the upper left fixed point of the water splash area, qw j is the width of the water splash area, qh j is the height of the water splash area, and j is the number of the water splash area;

[0192] wx i is the abscissa of the upper left fixed point of the water splash area, wy i is the ordinate of the upper left fixed point of the water splash area, ww i is the width of the water splash area, wh i is the height of the water splash area, and i is the frame number of the monitoring screen.

[0193] As a preferred example, the judgment module 305 is further configured to detect whether the ship is a motorboat according to the trajectory information of the ship and the ship water splash information:

[0194] According to the trajectory information of the ship {(lx i ,ly i )} and the ship water splash information {(wx i ,wy i ,ww i ,wh i )}, count the number of trajectory points N, and calculate the ship speed score g v ;

[0195] When the ship speed score g v is greater than the set seventh judgment threshold ts7, it is judged that the detected ship is a motorboat; among them,

[0196] g v = g s + g w

[0197]

[0198]

[0199]

[0200] c5 is the fifth correction constant obtained by training historical data, and c6 is the sixth correction constant obtained by training historical data.

[0201] As a preferred example, the determination module 305 is further configured to, after determining that the detected vessel is a motorboat, transmit a warning message to the management department in real time through a preset communication device.

[0202] It should be noted that the vessel detection module 301 provided in this embodiment can implement all the functions included in step S101 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be elaborated here;

[0203] It should be noted that the vessel feature extraction module 302 provided in this embodiment can implement all the functions included in step S102 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be elaborated here;

[0204] It should be noted that the vessel trajectory tracking module 303 provided in this embodiment can implement all the functions included in step S103 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be elaborated here;

[0205] It should be noted that the vessel water splash detection module 304 provided in this embodiment can implement all the functions included in step S104 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be elaborated here;

[0206] It should be noted that the determination module 305 provided in this embodiment can implement all the functions included in step S105 in Embodiment 1, solve the same technical problems, and achieve the same technical effects, which will not be elaborated here;

[0207] It should be noted that the device provided in Embodiment 2 and the method provided in Embodiment 1 belong to the same inventive concept, solve the same technical problems, and achieve the same technical effects. The device provided in Embodiment 2 can implement all the methods of Embodiment 1, and the same parts will not be elaborated here.

[0208] Embodiment 3

[0209] Based on the same inventive concept, an embodiment of the present invention further provides a motorboat intelligent detection device, as Figure 4 shown, the device includes:

[0210] including a memory 402, a processor 401, and a user interface 403;

[0211] The memory 402 is used to store computer programs;

[0212] The user interface 403 is used to interact with the user;

[0213] The processor 401 is used to read the computer program in the memory 402. When the processor 401 executes the computer program, it realizes:

[0214] Detect the ships appearing in the surveillance video according to the set ship detection model;

[0215] Extract the ship features of the ships appearing in the surveillance video, where the ship features include ship position, hull color, and ship size features;

[0216] Track the trajectory of the ship according to the ship features to obtain the trajectory information of the ship;

[0217] Obtain the ship splash information according to the set splash area detection model;

[0218] Detect whether the ship is a motorboat according to the trajectory information and ship splash information of the ship.

[0219] Among them, in Figure 4 In, the bus architecture may include any number of interconnected buses and bridges, specifically, various circuits of one or more processors represented by the processor 401 and the memory represented by the memory 402 are linked together. The bus architecture can also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art, so they will not be further described herein. The bus interface provides an interface. The processor 401 is responsible for managing the bus architecture and general processing, and the memory 402 can store the data used by the processor 501 when executing operations.

[0220] The processor 401 can be a CPU, ASIC, FPGA or CPLD, and the processor 401 can also adopt a multi-core architecture.

[0221] When the processor 401 executes the computer program stored in the memory 402, it implements any of the motorboat intelligent detection methods in Embodiment 1.

[0222] It should be noted that the device provided in Embodiment 3 and the method provided in Embodiment 1 belong to the same inventive concept, solve the same technical problems, and achieve the same technical effects. The device provided in Embodiment 3 can implement all the methods in Embodiment 1, and the same parts will not be described in detail.

[0223] This application also proposes a processor-readable storage medium. Among them, the processor-readable storage medium stores a computer program, and when the processor executes the computer program, it implements any of the motorboat intelligent detection methods in Embodiment 1.

[0224] It should be noted that the division of units in the embodiments of the present application is illustrative, merely a logical function division, and there may be other division methods in actual implementation. In addition, in each embodiment of the present application, the functional units may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit. The above integrated units may be implemented in the form of hardware or in the form of software functional units.

[0225] Those skilled in the art should understand that the embodiments of the present application may be provided as a method, a system, or a computer program product. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0226] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0227] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0228] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent detection method for a motorboat, characterized in that, Including: Detecting the ships appearing in the surveillance video according to the set ship detection model; Extracting the ship features of the ships appearing in the surveillance video, where the ship features include ship position, hull color, and ship size features; Tracking the trajectory of the ship according to the ship features to obtain the trajectory information of the ship; Obtaining the ship splash information according to the set splash area detection model; Detecting whether the ship is a motorboat according to the trajectory information and ship splash information of the ship; The set ship detection model includes: Collecting motorboat picture data and picture data of passing ships in the monitored waters, and labeling the ship types, where the ship types include motorboats and non-motorboats; Training the labeled ship picture data based on the YOLO model to obtain the set ship detection model; The tracking the trajectory of the ship according to the ship features to obtain the trajectory information of the ship includes: Track the vessel according to the position characteristics, hull color, and vessel size characteristics of the vessel to obtain the trajectory information {(lx i , ly i )} of the vessel. where (lx i , ly i ) is the vessel position feature, and: (lx i ,ly i ) = (x2 + 0.5w2, y2 + 0.5h2), where i represents the frame number of the video; The obtaining the ship splash information according to the set splash area detection model includes: For any frame i, use the water splash area detection model to detect the set of water splash areas in the frame {(qx j , qy j , qw j , qh j )}; Calculate the matching score gq for each splash region j ; Select the matching score gq of the water splash area j The largest water splash area serves as the ship's water splash information (wx i , wy i , ww i , wh i ) for the ship in the i-th frame Wherein, gq j = g1 j (c4g2 j + g3 j ) ts1 is the set first judgment threshold, ts2 is the set second judgment threshold, ts3 is the set third judgment threshold, ts4 is the set fourth judgment threshold, ts5 is the set fifth judgment threshold, ts6 is the set sixth judgment threshold, and c4 is the fourth correction constant obtained by training historical data; qx j is the abscissa of the upper left corner fixed point of the water splash area, qy j is the ordinate of the upper left corner fixed point of the water splash area, qw j is the width of the water splash area, qh j is the height of the water splash area, and j is the number of the water splash area; wx i is the abscissa of the upper left corner fixed point of the water splash area, wy i is the ordinate of the upper left corner fixed point of the water splash area, ww i is the width of the water splash area, wh i is the height of the water splash area, and i is the frame number of the monitoring screen; The detecting whether the ship is a motorboat according to the trajectory information and ship splash information of the ship includes: Based on the trajectory information of the vessel {(lx i , ly i )} and the water splash information of the vessel {(wx i , wy i , ww i , wh i ), count the number of trajectory points N, and calculate the vessel speed score g v ; When the vessel speed score g v is greater than the set seventh judgment threshold ts7, it is determined that the detected vessel is a motorboat; Wherein, g v = g s + g w c5 is the fifth correction constant obtained by training historical data, and c6 is the sixth correction constant obtained by training historical data.

2. The method according to claim 1, wherein Extracting the ship features of the ships appearing in the surveillance video, where the ship features include ship position, hull color, and ship size features includes: The ship size feature extraction method includes: If personnel information is detected in the monitoring area, extracting the size information (w1, h1) of the personnel and the monitoring area information (x2, y2, w2, h2), and calculating the ship size features in the following manner: Wherein, r1 is the ship size feature, x2 is the abscissa of the upper left fixed point of the ship box, y2 is the ordinate of the upper left fixed point of the ship box, w2 is the width of the ship box, h2 is the height of the ship box, w1 is the width of the personnel box, h1 is the height of the personnel box, and c1 is the first correction constant obtained by training historical data; If no personnel information is detected in the detection area, extracting the monitoring area information (x2, y2, w2, h2), and calculating the ship size features in the following manner: Where c2 is the second correction constant obtained by training historical data, and c3 is the third correction constant obtained by training historical data.

3. The method according to claim 1, wherein The set splash area detection model includes: Collecting the splash data when the ship passes by, and labeling the splashes; Training the labeled splash data based on the YOLO model to obtain the splash area detection model.

4. The method according to claim 1, characterized in that After determining that the detected ship is a motorboat, it further includes: Sending a warning message to the management department in real time through a preset communication device.

5. An intelligent detection device for a motorboat, characterized in that, Including: A ship detection module configured to detect the ships appearing in the surveillance video according to the set ship detection model; A vessel feature extraction module, configured to extract vessel features of vessels appearing in a surveillance video, where the vessel features include vessel position, hull color, and vessel size features; A vessel trajectory tracking module, configured to track the trajectory of the vessel based on the vessel features to obtain the trajectory information of the vessel; A vessel splash detection module, configured to obtain vessel splash information according to a set splash area detection model; A judgment module, configured to detect whether the vessel is a motorboat according to the trajectory information and the vessel splash information of the vessel; The set vessel detection model includes: Collecting motorboat picture data and picture data of passing vessels in a monitored water area, and labeling the vessel types, where the vessel types include motorboats and non-motorboats; Training the labeled vessel picture data based on the YOLO model to obtain a set vessel detection model; The tracking the trajectory of the vessel based on the vessel features to obtain the trajectory information of the vessel includes: Track the vessel based on the position characteristics, hull color, and vessel size characteristics of the vessel to obtain the trajectory information {(lx i , ly i )} of the vessel. Among them, (lx i , ly i ) is the vessel position feature, and: (lx i ,ly i ) = (x2 + 0.5w2, y2 + 0.5h2), where i represents the frame number of the video; The obtaining the vessel splash information according to the set splash area detection model includes: For any frame i, use the water splash area detection model to detect the set of water splash areas in the frame {(qx j , qy j , qw j , qh j )}; Calculate the matching score gq for each splash region j ; Select the matching score gq of the water splash area j The largest water splash area serves as the ship's water splash information (wx i , wy i , ww i , wh i ) for the ship in the i-th frame; Wherein, gq j = g1 j (c4g2 j + g3 j ) where ts1 is a set first judgment threshold, ts2 is a set second judgment threshold, ts3 is a set third judgment threshold, ts4 is a set fourth judgment threshold, ts5 is a set fifth judgment threshold, ts6 is a set sixth judgment threshold, and c4 is a fourth correction constant obtained by training historical data; qx j is the abscissa of the upper left corner fixed point of the water splash area, qy j is the ordinate of the upper left corner fixed point of the water splash area, qw j is the width of the water splash area, qh j is the height of the water splash area, and j is the number of the water splash area; wx i is the abscissa of the upper left fixed point of the water splash area, wy i is the ordinate of the upper left fixed point of the water splash area, ww i is the width of the water splash area, wh i is the height of the water splash area, and i is the frame number of the monitoring screen; The detecting whether the vessel is a motorboat according to the trajectory information and the vessel splash information of the vessel includes: Based on the trajectory information of the vessel {(lx i , ly i )} and the water splash information of the vessel {(wx i , wy i , ww i , wh i ), count the number of trajectory points N and calculate the vessel speed score g v ; When the vessel speed score g v is greater than the set seventh judgment threshold ts7, it is determined that the detected vessel is a speedboat; Wherein, g v = g s + g w c5 is a fifth correction constant obtained by training historical data, and c6 is a sixth correction constant obtained by training historical data.

6. An intelligent detection device for a motorboat, characterized in that, Including a memory, a processor, and a user interface; The memory is used to store a computer program; The user interface is used to interact with the user; The processor is used to read the computer program in the memory. When the processor executes the computer program, the intelligent motorboat detection method according to any one of claims 1 to 4 is implemented.

7. A processor-readable storage medium, characterized in that, The processor-readable storage medium stores a computer program. When the processor executes the computer program, the intelligent motorboat detection method according to any one of claims 1 to 4 is implemented.

Images