CN113989659A

CN113989659A - Ship target rapid detection method facing GEO remote sensing satellite

Info

Publication number: CN113989659A
Application number: CN202111192705.4A
Authority: CN
Inventors: 王俐云; 和涛; 方峰; 孙亚楠; 伊成俊; 陈绍龙
Original assignee: China Academy of Space Technology CAST
Current assignee: China Academy of Space Technology CAST
Priority date: 2021-10-13
Filing date: 2021-10-13
Publication date: 2022-01-28

Abstract

A rapid ship target detection method for a GEO remote sensing satellite. The application provides a ship target detection and tracking method, which comprises the following steps: the ground station receives imaging data downloaded in a real-time code stream mode in real time, and the received imaging data are transmitted to a user side in a real-time code stream data mode; in a preprocessing module at a user end, preprocessing imaging data to generate a sequence image product, wherein the preprocessing of the imaging data comprises screening of imaging spectral segments; the preprocessing module carries out geographic position registration on a frame of image product every time the frame of image product is generated, and simultaneously carries out ship target extraction on the frame of image product in parallel, wherein the ship target extraction comprises the following steps: slicing the frame of image product according to a preset position point to obtain a frame of image slice; detecting and extracting coordinates of the ship target in the frame image slice; and correcting the coordinates of the ship target in the frame of image product by using the geographic position registration result of the frame of image product.

Description

Ship target rapid detection method facing GEO remote sensing satellite

Technical Field

The application relates to rapid processing, detection and tracking of a ship target by using a GEO remote sensing satellite, in particular to detection and tracking of a large ship target in the middle and high sea by using a medium-low resolution sequence optical remote sensing image.

Background

The geostationary orbit remote sensing satellite (GEO remote sensing satellite) can reside above a fixed area for a long time due to being positioned in a geosynchronous orbit, and has the characteristics of observation width of hundreds of kilometers, continuous monitoring of targets in an observation range, real-time data downloading to a ground station and the like. Compared with a low-orbit observation system in the aspects of time resolution, observation range and the like, the system has the capability of quick task response and the capability of continuous monitoring of multiple targets in a large range. The method mainly guarantees weather and environment and reduces disasters in the current application, and has wide application prospect in ship target detection and tracking.

The task response speed, the data processing timeliness, the target detection accuracy and the target correlation correctness are important technical indexes influencing the detection and tracking of the ship target. The research aiming at ship target detection and tracking is mostly focused on the research of a detection and tracking algorithm, because the number of orbit of the GEO remote sensing satellite is small at present, only one high-resolution four-size satellite is used for civil use in China, the conventional task flow is not designed aiming at the ship target detection and tracking requirement, in the practical application, the time consumption is long due to task response, data transmission, processing and application ring, the timeliness is greatly reduced, and the time from the acquisition of imaging data to the generation of image products of a user is prolonged to several hours. Due to the lack of an accurate and efficient processing application system, the accuracy of target detection and tracking is difficult to guarantee. The practical application effect of ship target detection and tracking is severely restricted.

Fig. 1 shows a conventional task flow. After the full-spectrum satellite data (real-time code stream data) of a single task falls to the ground, a satellite control department performs batch preprocessing in a centralized manner to generate a standard 2-level image product (2-level full-spectrum sequence image). The generated image product is packaged for delivery to the user. At a user side, the whole batch of full spectral sequence images are utilized to perform ship target detection and target tracking aiming at the panoramic images of a plurality of spectral bands so as to obtain target tracking information.

In the conventional task flow of the target, the following problems exist:

1. the code stream data processing flow and the target detection and tracking flow are independent, and the serial connection of the two flows increases the time delay. After the production of the image products of the single task is finished, the satellite management and control department packs and transmits the image products to the user, the user carries out target detection and tracking after receiving all the image products of the single task, and the task flow time delay is long.

2. Because the single-task imaging data needs to be batch processed in a single-task imaging data set in a satellite control department to generate a standard 2-level image product and then distributed to users, a task queuing processing situation exists, and the processing delay is generally small at present.

3. Because the 2-level image product is generated by using the code stream data of the full spectrum, the geometric fine correction loop in the preprocessing program saves time.

4. The time consumption for panoramic image detection by utilizing multi-spectral-segment data is long.

Therefore, the conventional task flow of the target has the problem of poor task flow timeliness.

Disclosure of Invention

At least one object of the present application is to provide a ship target detection and tracking method, including:

the satellite control department injects the imaging task instruction to the stationary orbit remote sensing satellite;

the ground station receives imaging data downloaded in a real-time code stream mode in real time, and the received imaging data are transmitted to a user side in a real-time code stream data mode;

in a pre-processing module at a user end, pre-processing imaging data to generate a sequence image product, the sequence image product comprising a plurality of frames of image products, wherein the pre-processing of the imaging data comprises a screening of imaging spectra selected from at least one of the visible spectrum and the near infrared spectrum;

the preprocessing module carries out geographic position registration on a frame of image product every time the frame of image product is generated, and simultaneously carries out ship target extraction on the frame of image product in parallel, wherein the ship target extraction comprises the following steps: slicing the frame of image product according to a preset position point to obtain a frame of image slice; detecting the frame image slice and extracting the coordinates of the ship target in the frame image product;

correcting the coordinates of the ship target in the frame of image product by using the geographic position registration result of the frame of image product;

using a target tracking module to correlate the time and space information of the target from the second frame of image slices, and after the correlation of the continuous three frames of image slices is completed, confirming that the target tracking is successful;

and updating and outputting the target tracking information to realize the rolling output of the target tracking information.

According to one embodiment of the invention, wherein each time a new frame of image production is generated by the preprocessing module, the frame of image production is detected by the object detection module,

the detecting of the frame of image products comprises: carrying out geographic position registration on the frame of image product, and simultaneously carrying out ship target extraction on the frame of image product in parallel; and correcting the coordinates of the ship target in the frame image slice of the frame image product by using the geographic position registration result of the frame image product.

According to one embodiment of the invention, the control points are selected through a loaded control point library, so that the geographic position registration is carried out by using a standard map.

According to an embodiment of the invention, wherein the sequential image product is a level 1A sequential image product.

According to an embodiment of the present invention, the process of obtaining a frame image slice includes:

determining the coordinate of the ship target at the starting moment through prior information, extrapolating the radius of an active area at the current moment according to the movement speed of the ship, wherein the size of the image slice is a circumscribed rectangle of a circle which takes the coordinate at the starting moment as the center of the circle and the radius of the active area as the radius.

According to one embodiment of the invention, wherein the radius of the active area is 20km to 50 km.

According to one embodiment of the invention, the preprocessing further comprises data formatting, attitude and orbit determination, radiation correction, RPC orthorectification calculation.

According to an embodiment of the present invention, the target tracking information includes a target point position, a target heading, and a target speed.

The present invention also provides a computer readable storage medium having stored thereon software instructions which, when executed, implement the above-described method.

The invention also provides a ship target detection and tracking system, which is used for executing the method, and the system comprises a user side, wherein the user side comprises:

a preprocessing module;

a target detection module;

and a target tracking module.

The method provided by the application can greatly improve the processing timeliness, shorten the processing time delay and realize the real-time detection and tracking of the moving target. In addition, the method can also carry out rapid detection and identification, reduce the false alarm rate and improve the accuracy of target tracking.

Drawings

The above features, technical features, advantages and modes of realisation of the present application will be further described in the following detailed description of preferred embodiments in a clearly understandable manner, in conjunction with the accompanying drawings. The drawings are only for purposes of illustrating and explaining the present application and are not to be construed as limiting the scope of the present application. Wherein:

fig. 1 shows a conventional task flow for detection and tracking of a ship target.

Fig. 2 shows a flow of a ship target detection tracking method according to an embodiment of the present application.

Detailed Description

In order to more clearly understand the technical features, objects, and effects of the present application, embodiments of the present application will now be described with reference to the accompanying drawings.

The application provides a ship target rapid processing, detecting and tracking method based on a GEO remote sensing satellite, which can improve data processing speed, reduce processing time delay, has strong timeliness, and can give full play to the advantages of the GEO remote sensing satellite in rapid response and continuous tracking.

The flow of the ship target detection and tracking method provided by the application is shown in fig. 2. The ship target detection and tracking method provided by the invention comprises data preprocessing, target detection and target tracking.

The data preprocessing and the target detection are performed in a multi-thread parallel processing mode, and the received data are used as signals to start a data preprocessing thread and a target detection thread.

The target tracking operation needs to be initiated after at least the target detection operation for three frames of image data is completed. After the target tracking operation is started, the target motion information is correspondingly updated every time a new target detection result is received, so that the continuous rolling output of the target position and the navigational course motion information is realized.

As shown in fig. 2, the ship target detection and tracking method provided by the present application includes:

s1: reporting a ship target detection requirement by a user: the user determines that a suspicious ship target needs to be continuously tracked, and after a certain time position of the suspicious ship target is obtained through prior information (such as information obtained by detection means such as a shore-based radar and an AIS), task requirements (including target discovery time, target point position, tracking start time and continuous tracking time) are generated and directly reported to a satellite control department.

S2: the satellite control department carries out task planning: and generating an imaging task plan after the satellite control department performs task planning, and generating an imaging task instruction by the ground station and injecting the imaging task instruction to the geostationary orbit remote sensing satellite.

S3: imaging and downloading real-time code stream data by a stationary orbit remote sensing satellite: the satellite starts to shoot the area where the target point is located at the appointed time according to the imaging task instruction, and the satellite downloads imaging data in a real-time code stream mode while shooting.

S4: the ground station receives imaging data in real time and transmits the imaging data to a user: the ground station receives the original real-time code stream data downloaded by the satellite in real time and synchronously transmits the data to the satellite control department, and transmits the imaging data in the form of the original real-time code stream data downloaded by the satellite to a user side through a ground lead special line.

S5: in a pre-processing module at the user end, the imaging data is pre-processed to generate a sequence image product: and preprocessing the received imaging data by utilizing a preprocessing module. The pretreatment comprises the following steps: screening an imaging spectrum; generation of a level 1A image product. In the imaging spectrum screening, two imaging spectrum ranges can be selected (one or two of B1 visible spectrum range and B5 near infrared spectrum range). And in the step of generating the 1A-level image product, data formatting, attitude and orbit determination, radiation correction and RPC (remote position control) orthorectification calculation are included, and finally the 1A-level sequence image product (a standard product with an RPC file) is generated.

S6: detecting a ship target: and when the preprocessing module generates a frame of 1A-level image product (the frame of 1A-level image product is a panoramic image), carrying out geographic position registration on the frame of 1A-level image product, and simultaneously carrying out ship target extraction on the frame of 1A-level image product. Wherein the two steps of geographic position registration and ship target extraction are performed simultaneously and in parallel.

In the step of geographic position registration, a control point is selected through a loaded control point library, so that the geographic position registration of the frame of the 1A-level image product is carried out by using a standard map.

Wherein, in the naval vessel target extraction step, include: slicing the 1A-level image product of the frame according to a preset position point to obtain an image slice; detecting the image slice and extracting the coordinates of the ship target in the image product. Specifically, in the slicing process, a possible distribution area of the ship target is determined through prior information, and the frame of 1A-level image product (panoramic image) is sliced, so that the detection range is reduced. Through prior information, longitude and latitude coordinates of the ship target at the starting time T0 can be obtained (the coordinates can be a preset position point), and the radius of the moving area at the current time T1 can be extrapolated according to the ship moving speed V (R ═ V x (T1-T0)). The slice size is a circumscribed rectangle of a circle having the coordinates of start time T0 as the center and R as the radius. The image slice radius R is adjustable between 20-50km (e.g., 30km), determined based on the duration of the continuous tracking and the ship's speed.

S7: carrying out accurate positioning on a ship target: and correcting the coordinates of the ship target in the image slice of the frame of the 1A-level image product by using the geographic position registration result of the frame of the 1A-level image product, thereby improving the positioning accuracy. For example, the frame level 1A image product has an offset (Δ x, Δ y) with respect to the standard map according to the geographic position registration result, and the offset (Δ x, Δ y) is applied to the coordinates of the ship target on the frame level 1A image product to correct the coordinates of the ship target in the frame level 1A image product with respect to the standard map.

In the prior art, the two steps of geographic position registration and ship target extraction are performed in series. Namely, a standard map is used for carrying out geographic position registration on a certain frame of 1A-level image product, and then ship target extraction is carried out on the registered frame of 1A-level image product. The two steps are time-consuming and serial, so that the task flow is delayed greatly.

In the application, the two steps of geographic position registration and ship target extraction are performed in parallel, so that the time consumption can be greatly reduced.

S8: target association tracking, calculating course speed: and correlating the time and space information of the target so as to track the target, correlating from the second frame of image slices, and confirming that the target is tracked successfully after detecting and correlating the continuous three frames of image slices. And calculating the mean value and variance of the course and the speed, wherein the default variance is the final track with the minimum variance.

S9: continuously tracking the target: continuously tracking the target: and detecting the image slice of the image product and associating the image slice with a detected target every time a new frame of image product is generated in the preprocessing step. And if the target association is successful, updating and outputting the target tracking information, and realizing the rolling output of the target tracking information. The target tracking information comprises a target point position, a course and a navigation speed. And if the target association fails, indicating that the target is lost.

The ship target detection and tracking method provided by the application has the following technical characteristics:

(1) and integrating a data preprocessing function to a user side, optimizing a flow according to task requirements, and realizing real-time and rapid processing of data at the user side. The B1 or B5 single spectral band imaging data processing is selected to generate a 1A class image product.

(2) The target detection process improves timeliness and simultaneously considers positioning precision improvement.

By utilizing the characteristic of large single imaging width of the area array imaging of the stationary orbit remote sensing satellite, the geographic position registration and correction are carried out by selecting the control point under the condition that the shot image contains the land, so that the positioning precision is improved. The control point registration (geographic position registration) process and the ship target extraction are carried out in parallel, and only the position of the detected target is corrected, so that the target positioning precision is improved, and the detection timeliness is improved.

(3) In the existing target detection algorithm, a sea-land segmentation algorithm is generally required to be firstly utilized to identify and distinguish sea-land features. In the application, the scheme of image slices is adopted, and in the ship target detection of the sparse sea area in the middle and far seas, no land area exists in the image slices, so that the problem of sea and land segmentation does not need to be considered, and the target detection can be carried out only by utilizing the single-spectral-band image characteristics of visible light.

(4) And data preprocessing and target detection tracking are performed in parallel. The method comprises the steps that a pre-processed and generated 1A-level image product is transmitted to a target detection module and a target tracking module in real time, the target detection module starts target detection, after three continuous frames of image products are received, the target tracking module correlates the three frames of images to generate continuous motion information of a target, wherein the continuous motion information comprises information such as position, course, navigational speed and track. And then, when one frame of image product is received, the product is detected and the current information is updated, so that real-time and rapid detection and tracking are realized.

The method provided by the application has the following beneficial effects:

(1) target detection tracking is carried out only by using B1 or B5 single spectral band sequence image products, and processing efficiency is improved.

(2) The image is subjected to target slice windowing detection based on the prior information, so that rapid detection and identification can be performed, the false alarm rate is reduced, and the target tracking accuracy is improved.

(3) The pretreatment, the target detection and the target tracking are integrally designed, the processing flow is optimized, and all modules run in parallel according to the set service logic, so that the processing application efficiency is greatly improved. The time is shortened from several hours to less than 10 minutes, and the real-time detection and tracking of the moving target are realized.

It should be understood that although the present description has been described in terms of various embodiments, not every embodiment includes only a single embodiment, and such description is for clarity purposes only, and those skilled in the art will recognize that the embodiments described herein may be combined as suitable to form other embodiments, as will be appreciated by those skilled in the art.

The above description is only illustrative of the present invention and is not intended to limit the scope of the present invention. Any equivalent alterations, modifications and combinations that may be made by those skilled in the art without departing from the spirit and principles of this application shall fall within the scope of this application.

Claims

1. A ship target detection and tracking method comprises the following steps:

2. The method of claim 1, wherein,

the preprocessing module detects a frame of image product by the target detection module every time the preprocessing module generates a new frame of image product,

3. The method of claim 1, wherein the control points are selected by a loaded library of control points for geographic location registration using a standard map.

4. The method of claim 1, wherein the sequential image product is a level 1A sequential image product.

5. The method of claim 1, wherein obtaining a frame of image slices comprises:

6. The method of claim 5, wherein the active area radius is 20km to 50 km.

7. The method of claim 1, wherein the preprocessing further comprises data formatting, pose and orbit determination, radiation correction, RPC orthorectification calculations.

8. The method of claim 1, wherein the target tracking information includes a target point location, a target heading, a target speed.

9. A computer readable storage medium having stored thereon software instructions that, when executed, implement the method of any of claims 1-8.

10. A vessel target detection tracking system for performing the method of any of claims 1-8, the system comprising a user terminal comprising:

a preprocessing module;

a target detection module;

and a target tracking module.