阅读说明：本技术 一种基于无人船的信息共享系统 (Unmanned ship-based information sharing system ) 是由 初秀民 马玉鹏 吴勇 郑茂 田国昊 祝添权 于 2021-08-20 设计创作，主要内容包括：本发明涉及一种基于无人船的信息共享系统,包括精准定位模块、小目标识别模块、信息共享模块；精准定位模块采用基于差分定位的GPS定位方法,提供准确的无人船位置信息；小目标识别模块用于采集无人船前端的障碍物相对距离信息、图像信息；并根据采集的障碍物相对距离信息并结合精准定位模块获得的当前无人船位置信息,精准识别目标障碍物的位置信息；并对获取的图像信息进行视频目标的信息提取,包括视频目标识别、像素坐标与摄像机坐标转换；信息共享模块将上述信息通过云服务器进行数据融合,然后共享给航道上其他船舶。本发明能最大限度消除船舶航行盲区,为船舶安全航行创造条件,特别适用于无霾天气或能见度较低情况下的助航服务。(The invention relates to an information sharing system based on an unmanned ship, which comprises an accurate positioning module, a small target identification module and an information sharing module; the precise positioning module provides accurate unmanned ship position information by adopting a GPS positioning method based on differential positioning; the small target identification module is used for acquiring relative distance information and image information of obstacles at the front end of the unmanned ship; the position information of the target barrier is accurately identified according to the collected relative distance information of the barrier and the current unmanned ship position information obtained by combining the accurate positioning module; extracting information of a video target from the acquired image information, wherein the information comprises video target identification and pixel coordinate and camera coordinate conversion; and the information sharing module performs data fusion on the information through the cloud server and then shares the information with other ships on the channel. The invention can eliminate the dead zone of ship navigation to the utmost extent, creates conditions for safe navigation of the ship, and is particularly suitable for navigation aid service in haze-free weather or under the condition of low visibility.)

1. An unmanned ship-based information sharing system is characterized by comprising an accurate positioning module, a small target identification module and an information sharing module, wherein the accurate positioning module, the small target identification module and the information sharing module are installed on an unmanned ship;

the precise positioning module adopts a GPS positioning method based on differential positioning to provide accurate unmanned ship position information;

the small target identification module is used for acquiring relative distance information and image information of obstacles at the front end of the unmanned ship; the position information of the target obstacle is accurately identified according to the collected relative distance information of the obstacle and the current unmanned ship position information obtained by the accurate positioning module; extracting information of a video target from the acquired image information, wherein the information comprises video target identification and pixel coordinate and camera coordinate conversion;

the information sharing module carries out data fusion on the unmanned ship position information, the obstacle position information and the video target information acquired by the unmanned ship through the cloud server, and then shares the fused information to other ships on the channel.

2. The unmanned-vessel-based information sharing system of claim 1, wherein the precise positioning module comprises an electric compass device, a GPS system and a differential GPS system, and the electric compass device, the GPS system and the differential GPS system are all connected with the information sharing module through a wireless network; the GPS positioning method based on differential positioning is characterized in that position parameters obtained by a GPS system and a differential GPS system are subjected to position correction through real-time calculation to obtain the primary position of the unmanned ship, and after the primary position is obtained, correction numbers sent by a onshore reference station at regular time and the primary position of the unmanned ship are subjected to combined calculation to correct a test result, so that the unmanned ship positioning information with higher precision is obtained; the parameters provided by the electric compass device are used for ensuring that the unmanned ship does not have course deviation when navigating and determining that the basic navigation direction is stable and correct.

3. The unmanned-vessel-based information sharing system according to claim 1, wherein the GPS positioning method based on differential positioning specifically comprises the steps of:

1) the positions of an onshore reference station and a mobile station (unmanned ship) are observed by adopting a carrier phase positioning method;

the pseudorange equation of the shore reference station is

The pseudorange equation of the mobile station (unmanned ship) is

Wherein λ is a carrier wavelength;

the pseudorange observed values of a j th satellite of an onshore reference station and a mobile station (unmanned ship) are respectively obtained;the integer ambiguity of the j th satellite carrier when the onshore reference station and the mobile station (unmanned ship) receive respectively;

respectively are the carrier phase observed quantity of the jth satellite of the onshore reference station and the mobile station (unmanned ship);

the true distance from the onshore reference station to the satellite is known data;

is thatAndthe difference between them, called the carrier phase correction;

2) the mobile station (unmanned ship) carries out differential positioning through carrier phase correction, and the correction equation is as follows:

wherein the content of the first and second substances,the true distance from the mobile station (unmanned ship) to the satellite is to be evaluated in the positioning operation; dp is the residual error due to mobile and reference station clock differences;

3) by the formulae (1) to (3), it can be obtained

From equation (4), the position coordinates of the mobile station (unmanned ship), i.e., the position coordinates of the mobile station (unmanned ship) can be obtained by resolving the carrier phase ambiguity

4. The unmanned ship-based information sharing system of claim 3, wherein after the position coordinates of the mobile station (unmanned ship) are obtained, the obtained position coordinates are obtainedAssociationAnd performing combined calculation, setting multi-level neural network calculation, and finally obtaining accurate position coordinates.

5. The unmanned-vessel-based information sharing system of claim 1, wherein the small-target recognition module comprises a radar, a camera, an ultrasonic ranging device installed on the unmanned vessel; the radar and the camera are used for actively detecting the surrounding environment and detecting the small target obstacle under the condition of poor sight distance; the ultrasonic distance measuring device is used for measuring the distance between the obstacle and the unmanned ship; the radar, the camera and the ultrasonic distance measuring device are connected with the cloud server through a wireless network.

6. The unmanned-vessel-based information sharing system according to claim 5, wherein after the video information is collected by the camera, the image fusion analysis is performed by the information sharing module to further determine the type and risk of the obstacle.

7. The unmanned-vessel-based information sharing system of claim 1, wherein the small target recognition module performs video target recognition by adopting a deep learning framework-based Darknet network model and combining with a YOLOv3 algorithm to realize type recognition of a ship; the method specifically comprises the following steps:

1) image input: by designing a capture area, capturing the target mass center after entering the area, and numbering the captured target;

2) feature extraction: extracting and capturing target features including perimeter, area, gray value and the like, and carrying out fine adjustment according to target characteristics;

3) model learning: taking input image data as a learning sample, and performing template matching model relearning;

4) matching and positioning: rotating and translating the target area of the captured image to a standard position according to a standard template;

5) classification and identification: and after matching the captured barrier targets, classifying and labeling different categories.

6) And outputting a result: and finishing the obstacle type judgment and outputting a result.

8. The unmanned ship-based information sharing system of claim 1, wherein the small target recognition module performs pixel coordinate and camera coordinate conversion, which means that coordinate system conversion of obstacle coordinates in video target recognition is completed, and the specific method is as follows:

adopting a relatively ideal pinhole imaging model, assuming that light rays from a three-dimensional space are inversely mapped to a two-dimensional plane by a pinhole, a camera is horizontally installed, a pitch angle, a roll angle and a yaw angle are all 0 degrees, the installation height of the camera from a water surface is h, therefore, the coordinate system of the camera imaging model needs to be subjected to space transformation, and the pixel coordinate is converted into an image coordinate system with the following formula:

wherein u is the abscissa of the pixel coordinate system, v is the ordinate of the pixel coordinate system, and u₀Is the origin of the abscissa, v, of the pixel coordinate system₀Is the origin of the ordinate of the pixel coordinate system, x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, dx is the physical size of a single pixel point on the x axis, and dy is the physical size of a single pixel point on the y axis; the formula for converting the image coordinate system into the camera coordinate system is as follows:

wherein x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, f is the focal length of the camera, and x_cAs the abscissa, y, of the camera coordinate system_cAs ordinate, z, of the camera coordinate system_cIs the horizontal coordinate of the camera coordinate system;

according to the formulae (5) to (6), it can be obtained

Let y_cH, get:

9. the unmanned-vessel-based information sharing system of claim 1, wherein the information sharing module comprises an onboard STM32F407 development board, a peanut shell server, an ali cloud server; board year STM32F407 development board, peanut shell server, Ali cloud server all pass through wireless network connection, and the various information of gathering unmanned ship is through STM32F407 development board simple processing back, carries out high in the clouds data fusion by the cloud ware to data transmission to other boats and ships of channel that will obtain through wireless network accomplishes the information interaction task.

10. The unmanned-vessel-based information sharing system of claim 1, wherein the Ariiyun server is configured to transmit radar images and radar target tracking data; and the peanut shell server transmits video images.

Technical Field

The invention belongs to the technical field of ship information sharing, and particularly relates to an unmanned ship information sharing device under haze weather or low visibility.

Background

With the development of global economy integration, the ocean becomes a link for transportation and trade development of countries in the world. The development of the shipping industry brings about the remarkable increase of the traffic flow of ships, meanwhile, the large-scale ships also put higher requirements on the navigation conditions of water areas, and the intense competition of the shipping industry prompts the shipping to pursue higher operation efficiency. On the premise of ensuring the transportation safety, the method for improving the navigation efficiency of the ship, reducing the accidental loss and ensuring the navigation efficiency is always a key focus of the ship transportation industry. However, in recent years, the haze problem caused by air pollution is frequent, so that the visible distance of the ship in the sailing process is insufficient, great difficulty is brought to avoidance, navigation and positioning of the ship, and hidden danger is brought to safe sailing of the ship. In the past, under the conditions of heavy fog weather or low visibility, a ship carries out position determination through radio communication with other ships, but due to the problems of limited ship communication, difficult danger avoidance of small targets, rough positioning and the like, the processing of accurate information by a sailing ship falls into a performance bottleneck, and accurate navigation aid service cannot be provided.

Disclosure of Invention

The invention aims to solve the technical problem of providing an information sharing system based on an unmanned ship, which can automatically acquire information of a front sailing water area, including position information, image information, obstacle distance information and the like, and share monitoring information to other ships on a channel after data fusion is carried out through a cloud server, so that a ship sailing blind area is eliminated to the maximum extent, and conditions are created for safe sailing of the ship.

The technical scheme adopted by the invention for solving the technical problems is as follows:

an information sharing system based on an unmanned ship comprises an accurate positioning module, a small target identification module and an information sharing module, wherein the accurate positioning module, the small target identification module and the information sharing module are arranged on the unmanned ship;

the precise positioning module adopts a GPS positioning method based on differential positioning to provide accurate unmanned ship position information;

the small target identification module is used for acquiring relative distance information and image information of obstacles at the front end of the unmanned ship; the position information of the target obstacle is accurately identified according to the collected relative distance information of the obstacle and the current unmanned ship position information obtained by the accurate positioning module; extracting information of a video target from the acquired image information, wherein the information comprises video target identification and pixel coordinate and camera coordinate conversion;

the information sharing module carries out data fusion on the unmanned ship position information, the obstacle position information and the video target information acquired by the unmanned ship through the cloud server, and then shares the fused information to other ships on the channel.

In the above scheme, the accurate positioning module includes an electronic compass device, a GPS system, and a differential GPS system, and the electronic compass device, the GPS system, and the differential GPS system are all connected to the information sharing module through a wireless network; the GPS positioning method based on differential positioning is characterized in that position parameters obtained by a GPS system and a differential GPS system are subjected to position correction through real-time calculation to obtain the primary position of the unmanned ship, and after the primary position is obtained, correction numbers sent by a onshore reference station at regular time and the primary position of the unmanned ship are subjected to combined calculation to correct a test result, so that the unmanned ship positioning information with higher precision is obtained; the parameters provided by the electric compass device are used for ensuring that the unmanned ship does not have course deviation when navigating and determining that the basic navigation direction is stable and correct.

In the above solution, the GPS positioning method based on differential positioning specifically includes the following steps:

1) the positions of an onshore reference station and a mobile station (unmanned ship) are observed by adopting a carrier phase positioning method; the pseudorange equation of the shore reference station is

The pseudorange equation of the mobile station (unmanned ship) is

Wherein λ is a carrier wavelength;

the pseudorange observed values of a j th satellite of an onshore reference station and a mobile station (unmanned ship) are respectively obtained;the integer ambiguity of the j th satellite carrier when the onshore reference station and the mobile station (unmanned ship) receive respectively;

respectively are the carrier phase observed quantity of the jth satellite of the onshore reference station and the mobile station (unmanned ship);the true distance from the onshore reference station to the satellite is known data;

is thatAndthe difference between them, called the carrier phase correction;

2) the mobile station (unmanned ship) carries out differential positioning through carrier phase correction, and the correction equation is as follows:

wherein the content of the first and second substances,the true distance from the mobile station (unmanned ship) to the satellite is to be evaluated in the positioning operation; dp is the residual error due to mobile and reference station clock differences;

3) by the formulae (1) to (3), it can be obtained

From equation (4), the position coordinates of the mobile station (unmanned ship), i.e., the position coordinates of the mobile station (unmanned ship) can be obtained by resolving the carrier phase ambiguity

In the above scheme, after the position coordinates of the mobile station (unmanned ship) are obtained, the obtained position coordinates are obtainedAssociationAnd performing combined calculation, setting multi-level neural network calculation, and finally obtaining accurate position coordinates.

In the scheme, the small target identification module comprises a radar, a camera and an ultrasonic distance measuring device which are arranged on the unmanned ship; the radar and the camera are used for actively detecting the surrounding environment and detecting the small target obstacle under the condition of poor sight distance; the ultrasonic distance measuring device is used for measuring the distance between the obstacle and the unmanned ship; the radar, the camera and the ultrasonic distance measuring device are connected with the cloud server through a wireless network.

In the above scheme, after the video information is collected by the camera, the image fusion analysis is performed by the information sharing module, so as to further judge the type and the risk of the obstacle.

In the scheme, the small target identification module is used for identifying the video target, and the type identification of the ship is realized by adopting a Darknet network model based on a deep learning frame and combining a YOLOv3 algorithm; the method specifically comprises the following steps:

1) image input: by designing a capture area, capturing the target mass center after entering the area, and numbering the captured target;

2) feature extraction: extracting and capturing target features including perimeter, area, gray value and the like, and carrying out fine adjustment according to target characteristics;

3) model learning: taking input image data as a learning sample, and performing template matching model relearning;

4) matching and positioning: rotating and translating the target area of the captured image to a standard position according to a standard template;

5) classification and identification: and after matching the captured barrier targets, classifying and labeling different categories.

6) And outputting a result: and finishing the obstacle type judgment and outputting a result.

In the above solution, the small target recognition module performs conversion between pixel coordinates and camera coordinates, which means that conversion of a coordinate system of coordinates of an obstacle in video target recognition is completed, and the specific method includes:

adopting a relatively ideal pinhole imaging model, assuming that light rays from a three-dimensional space are inversely mapped to a two-dimensional plane by a pinhole, a camera is horizontally installed, a pitch angle, a roll angle and a yaw angle are all 0 degrees, the installation height of the camera from a water surface is h, therefore, the coordinate system of the camera imaging model needs to be subjected to space transformation, and the pixel coordinate is converted into an image coordinate system with the following formula:

wherein u is the abscissa of the pixel coordinate system, v is the ordinate of the pixel coordinate system, and u₀Is the origin of the abscissa, v, of the pixel coordinate system₀Is the origin of the ordinate of the pixel coordinate system, x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, dx is the physical size of a single pixel point on the x axis, and dy is the physical size of a single pixel point on the y axis; the formula for converting the image coordinate system into the camera coordinate system is as follows:

wherein x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, f is the focal length of the camera, and x_cAs the abscissa, y, of the camera coordinate system_cAs ordinate, z, of the camera coordinate system_cIs the horizontal coordinate of the camera coordinate system;

according to the formulae (5) to (6), it can be obtained

Let y_cH, get:

in the scheme, the information sharing module comprises an onboard STM32F407 development board, a peanut shell server and an Ali cloud server; board year STM32F407 development board, peanut shell server, Ali cloud server all pass through wireless network connection, and the various information of gathering unmanned ship is through STM32F407 development board simple processing back, carries out high in the clouds data fusion by the cloud ware to data transmission to other boats and ships of channel that will obtain through wireless network accomplishes the information interaction task.

In the scheme, the Ali cloud server is used for transmitting radar images and radar target tracking data; and the peanut shell server transmits video images.

The invention has the beneficial effects that:

1. the information sharing system of the unmanned ship can automatically acquire information of a front navigation water area, including position information, image information, barrier distance information and the like, and share monitoring information to other ships on a channel after data fusion is carried out through the cloud server, so that the navigation blind area of the ships is eliminated to the maximum extent, conditions are created for safe navigation of the ships, and the information sharing system is particularly suitable for navigation aid service under the conditions of haze-free weather or low visibility.

2. The small target identification technology can estimate the motion track of the small target, realize high-precision real-time tracking of the target, identify and classify the small targets such as yachts, fishing boats, navigation marks, obstructive objects and the like by adopting a deep learning method, and automatically associate, correlate, estimate and fuse small target tracking information detected by a radar and a camera through the fusion criterion of a cloud server, so that relatively more complete and accurate real-time dynamic information of the large ship field can be acquired, the method is superior to the detection based on the large ship radar in the past, and a new solution is provided for ship collision avoidance.

3. The video target recognition of the invention adopts a Darknet network model based on a deep learning frame, combines with a YOLOv3 algorithm, improves a basic classification network structure and a target binary classification prediction method in the traditional deep learning, and realizes the tracking of ships and the recognition of ship types. Based on the Darknet network, the real-time ship tracking and identifying algorithm is constructed by combining the YOLOv3 algorithm, so that the speed and precision are balanced, and the reliability of unmanned ship detection under the condition of low visibility is guaranteed.

4. The invention provides an offshore real-time regulation and control platform for designers, navigation assistance personnel can adjust unmanned ship accessories based on different water areas/weather conditions, and error correction is carried out on unmanned ship acquired data by adopting a multi-data fusion technology through a fusion principle of a cloud server, so that the accuracy is improved. The unmanned ship system can still provide accurate navigation service under the condition of a complex sea area, and navigation service in the complex water area is realized in a real sense.

5. The system adopts the unmanned ship to detect, has a relatively comprehensive overall risk evaluation method, and has the advantages of comprehensive perception and quick response to the traditional detection method based on the large ship.

Drawings

The invention will be further described with reference to the accompanying drawings and examples, in which:

FIG. 1 is a simplified block diagram of an unmanned ship based information sharing system of the present invention;

FIG. 2 is a diagram of a small object recognition module video object recognition process of the present invention;

fig. 3 is a diagram of an information sharing network of the unmanned ship based information sharing system of the present invention.

Detailed Description

For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

As shown in fig. 1, an information sharing system based on an unmanned ship provided by an embodiment of the present invention includes an accurate positioning module, a small target identification module, and an information sharing module installed on the unmanned ship. The precise positioning module adopts a GPS positioning method based on differential positioning to provide accurate unmanned ship position information. The small target identification module is used for acquiring relative distance information and image information of an obstacle at the front end of the unmanned ship, accurately identifying the position information of a target obstacle according to the acquired relative distance information of the obstacle and the current unmanned ship position information acquired by combining the accurate positioning module, and extracting information of a video target from the acquired image information, wherein the information comprises video target identification and pixel coordinate and camera coordinate conversion. The information sharing module carries out data fusion on the unmanned ship position information, the obstacle position information and the video target information acquired by the unmanned ship through the cloud server, and then shares the fused information to other ships on the channel, so that the ship navigation blind area is eliminated to the maximum extent, and conditions are created for safe navigation of the ships.

The precise positioning module comprises an electric compass device, a GPS system and a differential GPS system, wherein the electric compass device, the GPS system and the differential GPS system are all connected with the information sharing module through a wireless network. The parameters provided by the electric compass device are used for ensuring that the unmanned ship does not have course deviation when navigating and determining that the basic navigation direction is stable and correct. The unmanned ship accurate positioning adopts a GPS positioning method based on differential positioning, the GPS positioning method based on differential positioning corrects the position of the unmanned ship by real-time calculation according to the position parameters obtained by a GPS system and the differential GPS system to obtain the initial position of the unmanned ship, and after the initial position is obtained, the correction number sent by a onshore reference station at regular time and the initial position of the unmanned ship are subjected to combined calculation to correct the test result, so that the unmanned ship positioning information with higher precision is obtained. The GPS positioning method based on differential positioning comprises the following specific steps:

the onshore reference station transmits the carrier phase correction number to the mobile station (unmanned ship), and the mobile station (unmanned ship) corrects the carrier phase observation value by using the correction number, thereby solving the coordinate of the mobile station (unmanned ship). The observation model comprises the following steps:

1) the positions of the onshore reference station and the mobile station (unmanned ship) are observed by adopting a carrier phase positioning method.

The pseudorange equation of the shore reference station is

The pseudorange equation of the mobile station (unmanned ship) is

Where λ is the carrier wavelength.

The pseudorange observations are made to the jth satellite by the onshore reference station and the mobile station (unmanned ship), respectively.

The integer ambiguity of the j th satellite carrier is received by the onshore reference station and the mobile station (unmanned ship).

The j-th satellite carrier phase observed quantity of the onshore reference station and the mobile station (unmanned ship) are respectively.

Is the true distance from the onshore reference station to the satellite, which is known data.

Is thatAndthe difference between them is called the carrier phase correction.

2) The mobile station (unmanned ship) carries out differential positioning through carrier phase correction, and the correction equation is as follows:

wherein the content of the first and second substances,for the true distance of the mobile station (unmanned vessel) to the satelliteTo be evaluated in the positioning operation; dp is the residual error caused by the mobile station and reference station clock differences.

3) By the formulae (1) to (3), it can be obtained

From equation (4), the position coordinates of the mobile station (unmanned ship), i.e., the position coordinates of the mobile station (unmanned ship) can be obtained by resolving the carrier phase ambiguity

After the position coordinates of the mobile station (unmanned ship) are obtained, the obtained position coordinates are obtainedAssociationAnd performing combined calculation, setting multi-level neural network calculation, and finally obtaining a relatively accurate position coordinate.

The small target identification module comprises a radar, a camera and an ultrasonic distance measuring device which are arranged on the unmanned ship. The radar and the camera are used for actively detecting the surrounding environment and detecting small target obstacles under the condition of poor sight distance. The ultrasonic distance measuring device is used for measuring the distance between the obstacle and the unmanned ship. The radar, the camera and the ultrasonic distance measuring device are connected with the cloud server through a wireless network. After the video information is collected by the camera, image fusion analysis is carried out through the information sharing module, and the type and the danger of the barrier are further judged.

The small target identification module extracts the information of the video target by analyzing the video image acquired by the camera, the current GPS coordinate position and the compass information, and mainly comprises the functions of video target identification, pixel coordinate and camera coordinate conversion and the like.

The small target identification module is used for identifying the video target, and realizes the type identification of the ship by adopting a Darknet network model based on a deep learning frame and combining a YOLOv3 algorithm. As shown in fig. 2, the method specifically includes the following steps:

1) image input: by designing a capture area, capturing the target mass center after entering the area, and numbering the captured target;

2) feature extraction: extracting and capturing target features including perimeter, area, gray value and the like, and carrying out fine adjustment according to target characteristics;

3) model learning: taking input image data as a learning sample, and performing template matching model relearning;

4) matching and positioning: rotating and translating the target area of the captured image to a standard position according to a standard template;

5) classification and identification: after matching the captured barrier targets, classifying the captured barrier targets, and labeling different categories;

6) and outputting a result: and finishing the obstacle type judgment and outputting a result.

The video target identification method improves the basic classification network structure and the target classification prediction method in the traditional deep learning, and constructs a real-time identification algorithm by combining the YOLOv3 algorithm on the basis of the Darknet network, thereby achieving the equalization of speed and precision.

The small target identification module performs pixel coordinate and camera coordinate conversion, namely completes coordinate system conversion of obstacle coordinates in video target identification, and the specific method comprises the following steps:

adopting a relatively ideal pinhole imaging model, assuming that light rays from a three-dimensional space are inversely mapped to a two-dimensional plane by a pinhole, a camera is horizontally installed, a pitch angle, a roll angle and a yaw angle are all 0 degrees, the installation height of the camera from a water surface is h, therefore, the coordinate system of the camera imaging model needs to be subjected to space transformation, and the pixel coordinate is converted into an image coordinate system with the following formula:

wherein u is the abscissa of the pixel coordinate system, v is the ordinate of the pixel coordinate system, and u₀As the origin of the abscissa of the pixel coordinate system，v₀Is the origin of the ordinate of the pixel coordinate system, x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, dx is the physical size of a single pixel point on the x-axis, and dy is the physical size of a single pixel point on the y-axis.

The formula for converting the image coordinate system into the visible light camera coordinate system is as follows:

wherein x is the abscissa of the image coordinate system, y is the ordinate of the image coordinate system, f is the focal length of the camera, and x_cAs the abscissa, y, of the camera coordinate system_cAs ordinate, z, of the camera coordinate system_cThe camera coordinate system abscissa.

According to the formulae (5) to (6), it can be obtained

Let y_cH, get:

by the deduction of the formula, the coordinate system conversion of the obstacle coordinates in the video target recognition is completed, and a wider prospect is provided for the information utilization of the small target recognition module.

As shown in fig. 3, the information sharing module includes an onboard STM32F407 development board, a peanut shell server, and an ali cloud server. Onboard STM32F407 development board, peanut shell server, Ali cloud server all pass through wireless network connection, and the various information of gathering unmanned ship is through STM32F407 development board simple processing back, carries out high in the clouds data fusion by the cloud ware to data transmission to the channel other boats and ships that will obtain through wireless network accomplishes the information interaction task.

The specific implementation process of the invention is as follows: the accurate positioning module and the small target identification module collect environmental information through various sensors, and transmit radar data, GPS data, video data and compass data to the Ali cloud server and the peanut shell server through a wireless network which is internally arranged in the information sharing module. The Ali cloud server is used for transmitting radar images and radar target tracking data; the peanut shell server transmits video images, can play a double-channel role in information sharing, and can provide relatively accurate information sharing service even if the server is damaged. After the information reaches the Aliyun server and the peanut shell server, the obtained data are shared to other ships of the channel through cloud data fusion, and the information interaction task is completed. In the information interaction process, the requirements of transmission of high-definition videos, images and real-time ship acquisition data on communication bandwidth are high, so that the application of the 4G/5G technology in ship-shore information interaction can qualitatively improve the monitoring overall perception capability.

Further optimize, in this embodiment, the ultrasonic ranging function in the small target recognition module is used for perceiving the distance of the obstacle in the ship perception system, and for the selection of this module, in order to reliably guarantee the distance between this ship and other obstacles, at least two probes should be set up, are connected with the self-navigation mode controller of unmanned ship through the RS485 interface, and can return various data values for reference and use. So KS106 waterproof probe ultrasound module was selected. Ranging by the ultrasonic module is accomplished by SCL line polling. The detection command delay of about 40us exists before actual distance measurement, in the stage of sending the detection command, the SCL line is always arranged at a lower position, and when the SCL line is arranged at a higher position, the control driver can receive distance or time data in the distance measurement module through the RS485 bus to further determine the position of the obstacle. This module can direct output distance, and the unit is mm, also can export sound wave transmission time, can be used to later stage data comprehensive investigation or AIS data and exchange.

Further preferably, in this embodiment, the electric compass device adopts BW-H200 compass for acquiring the azimuth information of the unmanned ship, and the electric compass device adopts 485 communication with the self-navigation mode controller of the unmanned ship.

Further optimization, in this embodiment, the camera employs a visible light camera.

While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.