Remote sensing image moving ship tracking method based on ship wake identification

Abstract

The invention discloses a remote sensing image moving ship tracking method based on ship wake identification. The method comprises the following steps: acquiring point location information of a to-be-monitored target in the same area; determining an imaging mode and an imaging interval, and performing continuous satellite imaging to obtain multi-scene original data; performing standard product processing on the multi-scene original data to obtain a multi-scene 1A-level standard product; carrying out image registration on the 1A-grade standard product; increasing the image contrast of the 1A-level standard product after registration and enhancing image texture information to obtain a texture discrimination image; marking a point location of the ship target according to the texture information; and calculating navigational speed and course information according to the latitude and longitude coordinates of different positions to finish continuous tracking of the ship target. According to the method, the moving ship target can be accurately distinguished in the cloud-free and broken cloud environment, and meanwhile, long-time continuous motion information of multiple targets in the same area is generated, including position, course, navigational speed and track information.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a remote sensing image moving ship tracking method based on ship wake recognition. Background Art

[0002] Surveillance and monitoring of ships at sea is one of the important tasks of marine monitoring, and is of great significance to economic development, environmental protection, sea area use management, and the maintenance of marine rights and interests. Satellite remote sensing technology is an important means of surveillance and monitoring of ships at sea. Among them, high-resolution geosynchronous orbit remote sensing satellites can conduct high-frequency repeated monitoring of a large area of ​​sea for a long time, and obtain the dynamic changes of ships at sea in near real time. Therefore, compared with low-orbit remote sensing satellites, they have unique advantages and huge potential in the surveillance and monitoring of ships at sea.

[0003] Geostationary orbit optical remote sensing satellites have high orbit altitudes and long imaging distances. In addition, due to atmospheric attenuation and cloud obstruction, ship targets are generally weak and difficult to detect, which also makes the image of ships less detailed. Ships on the sea surface appear very blurry in the image, but the hull of the ship and the wake formed by its tail together form a long shuttle-shaped blurred image in the image, which can show significant texture features after image processing. Therefore, for multiple targets in the same area, achieving long-term continuous monitoring of ships under cloudless or broken cloud conditions has become a research challenge. Summary of the invention

[0004] The technical problem solved by the present invention is: to overcome the deficiencies of the prior art, to provide a remote sensing image moving ship tracking method based on ship wake recognition, and to solve the problem of long-term series monitoring and tracking of ships based on geostationary orbit satellites.

[0005] The technical solution of the present invention is: a remote sensing image moving ship tracking method based on ship wake recognition, comprising:

[0006] Obtain the point information of the target to be monitored in the same area, including time, location and heading;

[0007] Based on the acquired point information, the imaging mode and imaging interval are determined, and continuous satellite imaging is performed to obtain multi-view raw data;

[0008] Perform standard product processing on the original data of multiple scenes to obtain multi-scene 1A level standard products;

[0009] The first 1A-level standard product is used as the basic image, and the remaining 1A-level standard products are used as the images to be registered. The multi-scene 1A-level standard products are image registered, and the registered 1A-level standard products are output;

[0010] Increase the image contrast of the 1A-level standard product after registration and enhance the image texture information to obtain a texture discrimination image;

[0011] According to the texture information in the texture discrimination image, it is judged whether the image texture meets the characteristics of surface waves or turbulent wakes. If so, the point of the ship target is marked, otherwise it is not marked;

[0012] The longitude and latitude coordinates of the ship are obtained according to the marked points, and the positions of the same target marked with different image texture information are connected to form the motion trajectory of the ship target; the speed and heading information are calculated according to the longitude and latitude coordinates of different positions to complete the continuous tracking of the ship target.

[0013] The step of obtaining the location information of the target to be monitored in the same area includes:

[0014] Based on the ship name, call sign and other clear information of the target ship, the time, longitude, latitude and heading of the monitored ship and other point information are obtained through the ship automatic identification system.

[0015] The method includes determining the imaging mode and imaging interval according to the acquired point information, and performing continuous satellite imaging to obtain multi-view raw data; including:

[0016] The imaging modes include staring imaging mode, area imaging mode and mobile patrol mode; when imaging known targets, the staring mode is adopted, that is, the satellite attitude is maintained and pointing is performed for a short time and continuous imaging of the known position; single-segment imaging of the visible spectrum is adopted for imaging, with a total of 5 spectral bands; the working time of visible light imaging does not exceed 8 hours, and the number of power on and off times does not exceed 12 times; the imaging interval is set to 5 to 10 minutes.

[0017] The five spectral bands include 0.45 μm to 0.90 μm, 0.45 μm to 0.52 μm, 0.52 μm to 0.60 μm, 0.63 μm to 0.69 μm, and 0.76 μm to 0.90 μm.

[0018] The standard product processing of the multi-scene raw data to obtain the multi-scene 1A-level standard products includes: the satellite ground station receives the raw data transmitted by the satellite and transmits it to the ground data processing center, and the ground processing system sequentially performs input, scene division, auxiliary data processing, radiation correction, and system geometry correction processing on the received raw data to generate the 1A-level standard products.

[0019] The image registration and the output of the registration result include: selecting the same land, island, reef and other features in the first scene 1A-level standard product and the 1A-level standard product to be registered as a feature point pair set, using a polynomial transformation to transform the image to be registered, and obtaining the registered 1A-level standard product after the transformation.

[0020] The method of obtaining a visual discrimination image by increasing image contrast and enhancing image texture information includes:

[0021] The pixel grayscale value of the image after contrast enhancement is calculated based on the pixel grayscale value of the 1A standard product:

[0022] N new =N average +(NN average )*(1+k)

[0023] Where N new represents the pixel gray value after contrast enhancement, N represents the pixel gray value of the original image, and N average represents the average pixel grayscale value of the entire image, and k represents the adjustment range of image brightness, with a value range of [-1,1];

[0024] According to the gray value N of the image pixel after contrast enhancement new , the visual discrimination image after texture enhancement is calculated:

[0025]

[0026] g(i,j) represents the pixel grayscale value of the visual discrimination image after texture enhancement, a and b represent the number of selected surrounding neighborhood pixels, corresponding to a filter of size (2a+1)×(2b+1); w(s,t) represents the filter coefficient, f(i+s,j+t) represents N new The grayscale value of the image pixel in the i-th row and j-th column.

[0027] The step of judging whether the texture of the image meets the characteristics of the surface wave or the turbulent wake according to the texture information in the texture discrimination image includes:

[0028] Typical wakes of moving ships include surface waves and turbulent or eddy wakes. Surface waves are divided into Bragg waves and Kelvin wakes.

[0029] Feature 1: Bragg waves appear as narrow "V"-shaped bright lines on texture images. The line length is greater than 20 pixels, the angle is less than 10°, and the apex of the V is a bright spot with a length and width between 1 and 10 pixels.

[0030] Feature 2: Kelvin trails also appear as "V"-shaped bright lines on texture images. The line length is greater than 20 pixels, the angle is 39°, and the apex of the V is a bright spot with a length and width between 1 and 10 pixels.

[0031] Feature 3: Turbulence and eddy wakes appear as single straight lines on the texture image. The line length is greater than 20 pixels, and one of the endpoints is a bright spot with a length and width between 1 and 10 pixels.

[0032] If it meets one of the above three texture features, it is determined to be a ship target, and the center of gravity of the bright spot is marked as the point of the ship target.

[0033] The method of obtaining the longitude and latitude coordinates of the ship according to the marked points, connecting the positions of the same target marked with different image texture information to form a motion trajectory of the ship target, calculating the speed and heading information according to the longitude and latitude coordinates of different positions, and completing the continuous tracking of the ship target includes:

[0034] The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship speed information is calculated from the longitude and latitude, as follows:

[0035]

[0036] Where t0 and t1 represent any two different times of satellite imaging, d represents the distance the ship moves from t0 to t1, r represents the radius of the earth, and lat t0 ,lat t1 Indicates the latitude of the ship at time t0 and t1, lon t1 ,lon t1 Indicates the longitude of the ship at time t0 and t1;

[0037] The speed of the ship during the time period t0 and t1 is calculated from the imaging interval and the sailing distance. The calculation formula is as follows:

[0038]

[0039] In the formula, V represents the speed of the ship, d represents the distance the ship moves from t0 to t1, and the time difference between t0 and t1 is the imaging interval;

[0040] The longitude and latitude of the ship's position in remote sensing images of different phases are obtained according to the marked points, and the ship's heading information is calculated from the longitude and latitude.

[0041] The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship's heading information is calculated from the longitude and latitude, including:

[0042] Convert longitude and latitude from degrees to radians using the following formula:

[0043]

[0044] In the formula, Lat t0 , Lat t1 , Lon t1 , Lon t1 Indicates longitude and latitude lat t0 ,lat t1 ,lont1 ,lon t1 The corresponding radian value;

[0045] y=sin(Lon t1 -Lon t0 )·cos(Lat t1 )

[0046] x=cos(Lat t0 )·sin(Lat t1 )-sin(Lat t0 )·cos(Lat t1 )·cos(Lon t1 -Lon t0 )

[0047]

[0048] Where θ represents the heading angle of the ship, x and y represent the components of the change in latitude and longitude of the ship's position at t0 and t1 respectively, and their ratio determines the size and direction of the heading angle.

[0049] The advantages of the present invention compared with the prior art are:

[0050] Compared with low-orbit remote sensing satellite ship monitoring, the present invention provides a method for monitoring ships based on high-resolution geosynchronous orbit remote sensing satellites, which can perform high-frequency repeated monitoring of a large range of sea areas over a long period of time. Compared with low-orbit remote sensing satellite monitoring, it has the advantages of long monitoring time and high monitoring frequency.

[0051] In view of the problem that ship targets are generally weak and difficult to detect in high-orbit remote sensing satellite images, the present invention provides a method for tracking ship targets through the wake formed by the hull and the stern of the ship. By increasing the image contrast and enhancing the image texture information, the texture features of the ship target can be significantly enhanced, and the discrimination accuracy and efficiency of the ship target are significantly improved.

[0052] According to the above method, in cloudless and broken cloud environments, moving ship targets can be accurately identified, and continuous motion information of multiple targets in the same area can be generated, including position, heading, speed, track and other information. The single-band imaging mode can achieve continuous tracking and monitoring for up to 8 hours. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of the steps of a method for tracking a moving ship in remote sensing images based on ship wake recognition provided by an embodiment of the present invention;

[0054] Figure 2 Comparison of satellite single-band image ship wake imaging provided by an embodiment of the present invention;

[0055] Figure 3 A schematic diagram of filters of different sizes for remote sensing image texture enhancement provided by an embodiment of the present invention;

[0056] Figure 4 Effect diagram of texture enhancement of filters of different sizes for remote sensing images provided by an embodiment of the present invention;

[0057] Figure 5 A schematic diagram of a typical trail based on remote sensing images provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0058] Reference Figure 1 , which shows a flowchart of the steps of a remote sensing image moving ship tracking method based on ship wake recognition provided by an embodiment of the present invention, such as Figure 1 As shown, the method may include the following steps:

[0059] Step S1: Obtaining the point information of the target to be monitored in the same area, including time, location and heading;

[0060] First, obtain the location information of the targets to be monitored in the same area and clarify the needs of satellite imaging. The location information is obtained through the ship automatic identification system, and the static information such as the ship's name, call sign, MMSI, IMO, ship type, length, and ship width are obtained, and the dynamic information such as the longitude, latitude, bow direction, track direction, and speed of the ship target is obtained.

[0061] Step S2: Determine the imaging mode and imaging interval according to the acquired point information, and perform continuous satellite imaging to obtain multi-view raw data;

[0062] The (GF-4) satellite is my country's first geosynchronous orbit high-resolution remote sensing satellite, equipped with a staring camera with a resolution of 50 meters in visible light / 400 meters in medium-wave infrared and a width of more than 400 kilometers. It uses a staring array imaging method and has visible light, multi-spectral and infrared imaging capabilities. A stationary ship is a target of no more than 6 pixels in a visible light near-infrared image, and it is impossible to identify the ship target. Therefore, the present invention is mainly aimed at moving ships.

[0063] Imaging modes include staring imaging mode, regional imaging mode and mobile patrol mode. When imaging a known target, the staring mode is generally used, that is, the satellite attitude is kept pointing and the known position is imaged for a short period of time. In order to improve the efficiency of data transmission and processing, satellite imaging uses single-spectrum imaging in the visible spectrum. The specific spectral information is shown in the following table.

[0064] Table 1 Satellite visible light and near infrared payload spectrum information

[0065]

[0066]

[0067] like Figure 2 As shown in the figure, the ship wake is not obvious in the remote sensing image of the B2 blue spectrum band, so it cannot be selected as the imaging band; other bands have good imaging effects, and the B5 near-infrared band has the best imaging effect.

[0068] The design of satellite imaging intervals needs to consider satellite usage constraints and the motion characteristics of the target. First, during on-orbit operation, the satellite needs to meet usage constraints such as imaging time and number of power on and off times. Taking the average daily working time of visible light as an example, the daily visible light working time shall not exceed 8 hours; the number of power on and off times shall not exceed 12 times. In addition, the target motion characteristics need to be considered. If the interval is too short, the amount of data will be large and the target motion characteristics will not be obvious. The interval for transmitting data on the satellite should be greater than 5 minutes. Therefore, the imaging interval for long-term continuous monitoring should be greater than 5 minutes. On the other hand, the longer the time, the greater the difficulty of target association, and the sparser the track points, so it is necessary to consider the size and texture characteristics of the ship's wake on the remote sensing image. Based on comprehensive consideration of satellite usage constraints and motion characteristics, the imaging interval is generally set to 5 to 10 minutes.

[0069] Step S3: Perform standard product processing on the multi-scene original data to obtain multi-scene 1A level standard products;

[0070] Standard product processing refers to the satellite ground station receiving the original data transmitted by the satellite and transmitting it to the ground data processing center. The ground processing system will input, split the scene, perform auxiliary data processing, radiation correction, system geometry correction, etc. on the received original data to generate a Class 1A standard product.

[0071] Step S4: using the first scene 1A-level standard product as the basic image and the subsequent 1A-level standard products as the images to be registered, performing image registration on the obtained multiple scene 1A-level standard products, and outputting the registered 1A-level standard products;

[0072] Affected by factors such as satellite positioning accuracy and attitude control accuracy, multi-temporal remote sensing images have positioning errors, so image registration is required. This method uses the image registration function in ARCGIS and ENVI processing software to select the same land, islands, reefs and other features in the first scene 1A standard product and the 1A standard product to be registered as the feature point pair set, and uses polynomial transformation to transform the image to be registered, and then obtains the registered 1A standard product.

[0073] Step S5: increasing the image contrast of the 1A-level standard product after registration and enhancing the image texture information to obtain a texture discrimination image;

[0074] Since the grayscale values ​​of the ship wake and the background pixels are similar, as well as the fluid properties of the ship wake and the changes in the external environment, the shape and texture of the wake will be affected by many factors, such as wind speed, ocean current, temperature, etc., resulting in the ship wake in the infrared image often showing low contrast, blur and other problems, which affects the recognition and analysis of the wake. Therefore, it is necessary to improve the texture characteristics of the ship wake by enhancing the image contrast and spatial filtering algorithms.

[0075] In a specific implementation of the embodiment of the present invention, step S5 may include:

[0076] S5-1: Contrast enhancement. This is a method of changing the contrast of image pixels by changing the brightness value distribution of image pixels, expanding the overly concentrated pixel distribution area (brightness value distribution range) in the image, expanding the contrast of the image, and enhancing the layering of the image. The calculation formula is as follows:

[0077] N new =N average +(NN average )*(1+k) (1)

[0078] Where N new represents the pixel gray value after contrast enhancement, N represents the pixel gray value of the original image, and N average represents the average pixel grayscale value of the entire image, and k represents the adjustment range of the image brightness, which ranges from [-1,1].

[0079] S5-2: Spatial filtering.

[0080] In order to enhance the texture features of the ship wake, the difference between the wake contour and the edge is obtained through spatial filtering, thereby outlining the three-dimensional contour of the wake. Spatial filtering is a basic technology in the field of image processing. It involves direct operations on each pixel in the image and its surrounding neighborhood pixels. The core of spatial filtering is to use a small window called a filter or kernel, slide this window on the image, and multiply the grayscale value of the pixel in the window by the weight of the filter and then sum it to update the grayscale value of the center pixel. The calculation formula is as follows:

[0081]

[0082] g(i,j) represents the gray value of the image after filtering, a and b represent the number of selected surrounding neighborhood pixels, corresponding to a filter of size (2a+1)×(2b+1), such as Figure 3 As shown, they represent filters of size 3×3, 5×5, 7×7, and 9×9, respectively. w(s, t) represents the filter coefficient, and f(i+s, j+t) represents N new The grayscale value of the image pixel in the i-th row and j-th column.

[0083] Figure 4 The texture enhancement effects of different filters. It can be seen from the figure that the 5×5 filter can enhance the texture characteristics of the ship wake and effectively filter the sea clutter, with a better processing effect.

[0084] Step S6: judging whether the image texture meets the characteristics of surface waves or turbulent wakes according to the texture information in the texture discrimination image, if so, marking the point of the ship target, otherwise not marking;

[0085] The trace left by a moving ship or other moving object in the water is called a wake. Figure 5 As shown in the figure, the typical wake of a moving ship includes two forms: surface waves and turbulent or eddy wakes. Surface waves are divided into Bragg waves and Kelvin wakes. Bragg waves appear as narrow "V"-shaped bright lines on the texture image, with a line length greater than 20 pixels, an angle less than 10°, and the apex of the V is a bright spot with a length and width between 1-10 pixels; Kelvin wakes also appear as "V"-shaped bright lines on the texture image, with a line length greater than 20 pixels, an angle of about 39°, and the apex of the V is a bright spot with a length and width between 1-10 pixels. Turbulent and eddy wakes appear as a single straight line on the texture image, with a line length greater than 20 pixels, and one of the endpoints is a bright spot with a length and width between 1-10 pixels. If the above texture features are met, it is determined to be a ship target, and the center of gravity of the "bright spot" is marked as the target point.

[0086] Step S7: acquiring the longitude and latitude coordinates of the ship according to the marked points, and connecting the positions of the same target marked by different image texture information to form a motion trajectory of the ship target; calculating the heading and speed information according to the longitude and latitude coordinates of different positions to complete the continuous tracking of the ship target;

[0087] The longitude and latitude coordinates of the ship are obtained according to the marked points, and the positions of the same target marked with different image texture information are connected to form the motion trajectory of the ship target. The speed and heading information are calculated according to the longitude and latitude coordinates of different positions to complete the continuous tracking of the ship target, including:

[0088] The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship speed information is calculated from the longitude and latitude, as follows:

[0089]

[0090] Where t0 and t1 represent the time of satellite imaging, d represents the distance the ship moves from t0 to t1, r represents the radius of the earth, and lat t0 ,lat t1Indicates the latitude of the ship at time t0 and t1, lon t1 ,lon t1 Indicates the longitude of the ship at time t0 and t1.

[0091] The speed of the ship during this period is calculated from the imaging interval and the sailing distance. The calculation formula is as follows:

[0092]

[0093] Where V represents the speed of the ship, d represents the distance the ship moves from t0 to t1, and the time difference between t0 and t1 is the imaging interval.

[0094] The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship's heading information is calculated from the longitude and latitude, as follows:

[0095] Convert longitude and latitude from degrees to radians using the following formula:

[0096]

[0097] In the formula, Lat t0 , Lat t1 , Lon t1 , Lon t1 Indicates longitude and latitude lat t0 ,lat t1 ,lon t1 ,lon t1 The corresponding radian value.

[0098] y=sin(Lon t1 -Lon t0 )·cos(Lat t1 ) (6)

[0099] x=cos(Lat t0 )·sin(Lat t1 )-sin(Lat t0 )·cos(Lat t1 )·cos(Lon t1 -Lon t0 ) (7)

[0100]

[0101] Where θ represents the heading angle of the ship, x and y represent the components of the change in latitude and longitude of the ship's position at t0 and t1 respectively, and their ratio determines the size and direction of the heading angle.

[0102] Although the present invention has been disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solution of the present invention by using the technical contents disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for tracking moving ships in remote sensing images based on ship wake recognition, characterized in that: include: Obtain the point information of the target to be monitored in the same area, including time, location and heading; Based on the acquired point information, the imaging mode and imaging interval are determined, and continuous satellite imaging is performed to obtain multi-view raw data; Perform standard product processing on the original data of multiple scenes to obtain multi-scene 1A level standard products; The first 1A-level standard product is used as the basic image, and the remaining 1A-level standard products are used as the images to be registered. The multi-scene 1A-level standard products are image registered, and the registered 1A-level standard products are output; Increase the image contrast of the 1A-level standard product after registration and enhance the image texture information to obtain a texture discrimination image; According to the texture information in the texture discrimination image, it is judged whether the image texture meets the characteristics of surface waves or turbulent wakes. If so, the point of the ship target is marked, otherwise it is not marked; The longitude and latitude coordinates of the ship are obtained according to the marked points, and the positions of the same target marked with different image texture information are connected to form the motion trajectory of the ship target; the speed and heading information are calculated according to the longitude and latitude coordinates of different positions to complete the continuous tracking of the ship target.

2. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The step of obtaining the location information of the target to be monitored in the same area includes: Based on the ship name, call sign and other clear information of the target ship, the time, longitude, latitude and heading of the monitored ship and other point information are obtained through the ship automatic identification system.

3. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The method includes determining the imaging mode and imaging interval according to the acquired point information, and performing continuous satellite imaging to obtain multi-view raw data; including: The imaging modes include staring imaging mode, area imaging mode and mobile patrol mode; when imaging known targets, the staring mode is adopted, that is, the satellite attitude is maintained and pointing is performed for a short time and continuous imaging of the known position; single-segment imaging of the visible spectrum is adopted for imaging, with a total of 5 spectral bands; the working time of visible light imaging does not exceed 8 hours, and the number of power on and off times does not exceed 12 times; the imaging interval is set to 5 to 10 minutes.

4. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 3, characterized in that: The five spectral bands include 0.45 μm to 0.90 μm, 0.45 μm to 0.52 μm, 0.52 μm to 0.60 μm, 0.63 μm to 0.69 μm, and 0.76 μm to 0.90 μm.

5. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The standard product processing of the multi-scene raw data to obtain the multi-scene 1A-level standard products includes: the satellite ground station receives the raw data transmitted by the satellite and transmits it to the ground data processing center, and the ground processing system sequentially performs input, scene division, auxiliary data processing, radiation correction, and system geometry correction processing on the received raw data to generate the 1A-level standard products.

6. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The image registration and the output of the registration result include: selecting the same land, island, reef and other features in the first scene 1A-level standard product and the 1A-level standard product to be registered as a feature point pair set, using a polynomial transformation to transform the image to be registered, and obtaining the registered 1A-level standard product after the transformation.

7. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The method of obtaining a visual discrimination image by increasing image contrast and enhancing image texture information includes: The pixel grayscale value of the image after contrast enhancement is calculated based on the pixel grayscale value of the 1A standard product: N new =N average +(N-N average )*(1+k) Where N new represents the pixel gray value after contrast enhancement, N represents the pixel gray value of the original image, and N average represents the average pixel grayscale value of the entire image, and k represents the adjustment range of image brightness, with a value range of [-1,1]; According to the gray value N of the image pixel after contrast enhancement new , the visual discrimination image after texture enhancement is calculated: g(i,j) represents the pixel grayscale value of the visual discrimination image after texture enhancement, a and b represent the number of selected surrounding neighborhood pixels, corresponding to a filter of size (2a+1)×(2b+1); w(s,t) represents the filter coefficient, f(i+s,j+t) represents N new The grayscale value of the image pixel in the i-th row and j-th column.

8. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The step of judging whether the texture of the image meets the characteristics of the surface wave or the turbulent wake according to the texture information in the texture discrimination image includes: Typical wakes of moving ships include surface waves and turbulent or eddy wakes. Surface waves are divided into Bragg waves and Kelvin wakes. Feature 1: Bragg waves appear as narrow "V"-shaped bright lines on texture images. The line length is greater than 20 pixels, the angle is less than 10°, and the apex of the V is a bright spot with a length and width between 1 and 10 pixels. Feature 2: Kelvin trails also appear as "V"-shaped bright lines on texture images. The line length is greater than 20 pixels, the angle is 39°, and the apex of the V is a bright spot with a length and width between 1 and 10 pixels. Feature 3: Turbulence and eddy wakes appear as single straight lines on the texture image. The line length is greater than 20 pixels, and one of the endpoints is a bright spot with a length and width between 1 and 10 pixels. If it meets one of the above three texture features, it is determined to be a ship target, and the center of gravity of the bright spot is marked as the point of the ship target.

9. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 1, characterized in that: The method of obtaining the longitude and latitude coordinates of the ship according to the marked points, connecting the positions of the same target marked with different image texture information to form a motion trajectory of the ship target, calculating the speed and heading information according to the longitude and latitude coordinates of different positions, and completing the continuous tracking of the ship target includes: The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship speed information is calculated from the longitude and latitude, as follows: Where t0 and t1 represent any two different times of satellite imaging, d represents the distance the ship moves from t0 to t1, r represents the radius of the earth, and lat t0 ,lat t1 Indicates the latitude of the ship at time t0 and t1, lon t1 ,lon t1 Indicates the longitude of the ship at time t0 and t1; The speed of the ship during the time period t0 and t1 is calculated from the imaging interval and the sailing distance. The calculation formula is as follows: In the formula, V represents the speed of the ship, d represents the distance the ship moves from t0 to t1, and the time difference between t0 and t1 is the imaging interval; The longitude and latitude of the ship's position in remote sensing images of different phases are obtained according to the marked points, and the ship's heading information is calculated from the longitude and latitude.

10. The method for tracking a moving ship in remote sensing images based on ship wake recognition according to claim 9, characterized in that: The longitude and latitude of the ship position in remote sensing images of different time phases are obtained according to the marked points, and the ship's heading information is calculated from the longitude and latitude, including: Convert longitude and latitude from degrees to radians using the following formula: In the formula, Lat t0 , Lat t1 , Lon t1 , Lon t1 Indicates longitude and latitude lat t0 ,lat t1 ,lon t1 ,lon t1 The corresponding radian value; y=sin(Lon t1 -Lon t0 )·cos(Lat t1 ) x=cos(Lat t0 )·sin(Lat t1 )-sin(Lat t0 )·bas(Lat t1 )·cos(Lon t1 -Lon t0 ) Where θ represents the heading angle of the ship, x and y represent the components of the change in latitude and longitude of the ship's position at t0 and t1 respectively, and their ratio determines the size and direction of the heading angle.

Images