阅读说明：本技术 一种自动测量基站天线参数的方法和系统 (Method and system for automatically measuring antenna parameters of base station ) 是由 许理 曹予飞 张磊 曾敬勇 程杰 熊霖峰 雷滔 张肖 于 2020-04-15 设计创作，主要内容包括：一种自动测量基站天线参数的方法和系统,包括：飞行采集装置以基站抱杆为圆点飞行一周,并将采集的视频流传给天线参数测量装置；天线参数测量装置将视频流输入目标跟踪分析模型,输出跟踪到的所有天线目标信息；继续将每个天线目标的天线目标信息输入目标截图筛选模型,输出由每个天线目标在正侧面->正面->正侧面这一过程的所有目标截图所构成的一个目标截图序列；再将每个天线目标的目标截图序列中的每张目标截图输入至目标轮廓拟合模型,并输出每张目标截图中天线目标的轮廓信息；根据每个天线目标的目标截图序列中的所有目标截图及轮廓信息,计算每个天线目标的参数。本发明属于信息技术领域,能有效提高自动测量精度,并避免外部干扰。(A method and system for automatically measuring antenna parameters of a base station includes: the flight acquisition device takes a base station holding pole as a round point to fly for a circle and transmits the acquired video stream to the antenna parameter measurement device; the antenna parameter measuring device inputs the video stream into a target tracking analysis model and outputs all tracked antenna target information; continuously inputting the antenna target information of each antenna target into a target screenshot screening model, and outputting a target screenshot sequence formed by all target screenshots of the process that each antenna target is on the front side- > front side; inputting each target screenshot in the target screenshot sequence of each antenna target into a target contour fitting model, and outputting contour information of the antenna target in each target screenshot; and calculating parameters of each antenna target according to all target screenshots and contour information in the target screenshot sequence of each antenna target. The invention belongs to the technical field of information, and can effectively improve the automatic measurement precision and avoid external interference.)

1. A method for automatically measuring antenna parameters of a base station is characterized by comprising the following steps:

setting a round point, a radius and a flying height of a flying acquisition device flying around a point aiming at the type of an antenna of a base station;

secondly, the flight acquisition device flies for a circle by taking the center of the pole of the base station as a round point according to the preset radius and the preset flight height, and transmits the acquired video stream to the antenna parameter measuring device;

thirdly, the antenna parameter measuring device stores the video stream sent by the flight acquisition device;

fourthly, constructing and training a target tracking analysis model, wherein the target tracking analysis model is used for tracking all antenna targets appearing in the video stream, inputting the video stream sent by the flight acquisition device into the target tracking analysis model, and outputting all antenna target information tracked from the video stream, and the antenna target information comprises coordinate information of each antenna target, a target screenshot, and orientation and longitude and latitude information of the flight acquisition device corresponding to the picture;

step five, constructing and training a target screenshot screening model, wherein the target screenshot screening model is used for screening all target screenshots of each antenna target in the process of front side- > front side for the antenna target, inputting the antenna target information of each antenna target output by the target tracking analysis model into the target screenshot screening model one by one, and outputting a target screenshot sequence formed by all target screenshots of each antenna target in the process of front side- > front side;

step six, constructing and training a target contour fitting model, wherein the target contour fitting model is used for fitting contour information of the antenna target in the target screenshot, continuously inputting each target screenshot in a target screenshot sequence of each antenna target output by the target screenshot screening model into the target contour fitting model, and outputting the contour information of the antenna target in each target screenshot;

and step seven, calculating parameters of each antenna target according to all target screenshots in the target screenshot sequence of each antenna target and the profile information in the target screenshots, wherein the parameters include but are not limited to a direction angle, a mechanical inclination angle and a hanging height.

2. The method of claim 1, wherein step three further comprises:

the antenna parameter measuring device stores the video stream sent by the flight acquisition device, then intercepts each picture in the video stream into a sub-picture in a certain area according to a certain intercepting area, and replaces an original picture in the video stream with the intercepted sub-picture.

3. The method of claim 1, wherein step four further comprises:

counting the number of target screenshots corresponding to each target tracked by the target tracking analysis model and the number num of antennas of the base station, sequencing all tracked targets according to the sequence from the number of target screenshots to the number of target screenshots, finally selecting num targets sequenced in the front as antenna targets of the base station, namely, num targets are antenna targets tracked from a video stream, and finally outputting num antenna target information.

4. The method as claimed in claim 1, wherein the target tracking analysis model in the fourth step is implemented by using a multi-target tracking algorithm yolov3_ depsort, the target screenshot screening model in the fifth step is implemented by using a fine-grained classification algorithm DFL-CNN, and the target contour fitting model in the sixth step is implemented by using a residual error network resnet to fit the target contour.

5. The method of claim 1, wherein when calculating the direction angle of each antenna target, step seven further comprises:

step 7A, calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, then selecting a plurality of target screenshots with large average widths for each antenna target, and calculating the direction angle of each antenna target according to the average width: the direction angle of each antenna target is the average value of the orientation of the flight acquisition device corresponding to the plurality of target screenshots selected for the antenna target,

when calculating the mechanical tilt angle of each antenna target, the seventh step further comprises:

step 7B, calculating the average width of each antenna target in each target screenshot according to the contour information of each antenna target in each target screenshot of the target screenshot sequence, selecting a plurality of target screenshots with small average widths for each antenna target, and calculating the mechanical inclination angle of each antenna target according to the target width: fitting the profile information of the antenna targets in the selected multiple target screenshots to form a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction,

when the hanging height of each antenna target is calculated, the seventh step further comprises:

and 7C, calculating the hanging height of each antenna target, namely calculating the difference between the altitude of the flight acquisition device at the tower base when the flight acquisition device just takes off and the altitude of the flight around the point.

6. The method of claim 5, wherein step 7B further comprises:

step 7B1, selecting a plurality of target screenshots with small average widths for each antenna target according to the average width of each antenna target in each target screenshot of the target screenshot sequence, wherein the selected target screenshots are called original target screenshots;

step 7B2, extracting an original picture corresponding to each original target screenshot from the video stream, then respectively expanding the width and the height of the original target screenshot by a plurality of times, and then taking the original target screenshot frame as the center and intercepting a new target screenshot from the original picture according to the expanded width and height;

and 7B3, inputting a plurality of new target screenshots obtained by intercepting each antenna target into a target contour fitting model, thereby outputting contour information of each antenna target in each new target screenshot, then calculating the average width of each antenna target in each new target screenshot, selecting a plurality of new target screenshots with small average widths, and finally fitting the contour information of the antenna target in the selected plurality of new target screenshots into a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction.

7. The utility model provides a system for automatic measurement base station antenna parameter which characterized in that, including flight collection system, network communication device and antenna parameter measurement device, wherein:

the flight acquisition device is used for flying for a circle by taking the center of the pole of the base station as a circular point according to the preset radius and the preset flight height, and transmitting the acquired video stream to the antenna parameter measuring device;

a network communication device for communication between the flight acquisition device and the antenna parameter measurement device,

the antenna parameter measuring device further comprises a model management unit and a parameter calculation unit, wherein:

the model management unit is used for constructing and training an internal model component, inputting the video stream sent by the flight acquisition device into the internal model component so as to track all the antenna target information appearing in the video stream, screening all target screenshots of each antenna target in the process of front side- > front side for each antenna target from all the antenna target information, and simultaneously fitting to obtain the contour information of the antenna targets in all the target screenshots;

and the parameter calculation unit is used for calculating parameters of each antenna target according to all target screenshots and profile information in the target screenshots of the process that each antenna target is on the front side surface- > front side surface, which are output by the model management unit, wherein the parameters include but are not limited to a direction angle, a mechanical inclination angle and a hanging height.

8. The system of claim 7, wherein the flight collection device further comprises an aircraft and a remote control device, the aircraft is installed with a high-definition camera and a navigation recording unit, the high-definition camera is used for collecting video stream data of an antenna in the base station, the navigation recording unit is used for recording the flight height and the latitude and longitude information of the aircraft, and the remote control device is used for controlling the flight of the aircraft.

9. The system of claim 7, wherein the model management unit of the antenna parameter measuring device further comprises:

the label training component is used for labeling the training data of the model component in the mode management unit, and constructing and training the labeled training data to generate a corresponding model component;

the target tracking analysis model component is used for reading the stored video stream, tracking all antenna targets appearing in the video stream, and then outputting all tracked antenna target information, wherein the antenna target information comprises coordinate information of each antenna target, a target screenshot, and orientation and longitude and latitude information of a flight acquisition device corresponding to a picture, and the component is realized by adopting a multi-target tracking algorithm yolov3_ depsort;

the target screenshot screening model component is used for receiving antenna target information of each antenna target output by the target tracking analysis model component, screening all target screenshots of the process that each antenna target is on the front side- > front side for each antenna target, and then outputting a target screenshot sequence of each antenna target, wherein the target screenshot sequence is formed by all target screenshots of the process that the antenna target is on the front side- > front side, and the component is realized by adopting a fine-grained classification algorithm DFL-CNN;

and the target contour fitting model component is used for receiving each target screenshot in the target screenshot sequence of each antenna target output by the target screenshot screening model component, fitting to obtain contour information of the antenna target in each target screenshot, and then outputting the contour information of the antenna target in the target screenshot, and the component adopts a residual error network resnet algorithm to fit the target contour.

10. The system of claim 9, wherein the model management unit further comprises:

and the video stream processing component is used for storing the video stream sent by the flight acquisition device, then intercepting each picture in the video stream into a sub-picture in a certain area according to a certain intercepting area, and replacing an original picture in the video stream with the intercepted sub-picture.

11. The system of claim 9, wherein the target tracking analysis model component further comprises:

the antenna target selection component is used for counting the number of target screenshots corresponding to each target tracked by the target tracking analysis model and the number num of antennas of the base station, then sequencing all tracked targets according to the sequence of the number of target screenshots from multiple to few, finally selecting num targets sequenced in the front as antenna targets of the base station, namely the num targets are antenna targets tracked from a video stream, and finally outputting the num antenna target information.

12. The system of claim 7, wherein the parameter calculating unit further comprises:

the direction angle measuring component is used for calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, then selecting a plurality of target screenshots with large average widths for each antenna target, and calculating the direction angle of each antenna target according to the average width: the direction angle of each antenna target is the average value of the orientation of the flight acquisition device corresponding to the plurality of target screenshots selected for the antenna target; or the like, or, alternatively,

the mechanical inclination angle measuring component is used for calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, selecting a plurality of target screenshots with small average widths for each antenna target, and calculating the mechanical inclination angle of each antenna target according to the target width: fitting the contour information of the antenna targets in the picked multiple target screenshots to form a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction; or the like, or, alternatively,

and the hanging height measuring component is used for calculating the hanging height of each antenna target, namely the difference value of the altitude of the flight acquisition device at the tower base when the flight acquisition device just takes off and the altitude of the flight around the point.

13. The system of claim 12, wherein the mechanical tilt measurement assembly further comprises:

the mechanical inclination angle calculation component is used for selecting a plurality of target screenshots with small average widths for each antenna target according to the average width of each antenna target in each target screenshot of the target screenshot sequence, and the selected target screenshots are called original target screenshots; then, extracting an original picture corresponding to each original target screenshot from the video stream, respectively expanding the width and the height of the original target screenshot by a plurality of times, and taking the original target screenshot frame as the center and intercepting a new target screenshot from the original picture according to the expanded width and the expanded height; and finally, inputting a plurality of new target screenshots obtained by intercepting each antenna target into a target contour fitting model component, thereby obtaining contour information of each antenna target in each new target screenshot, calculating the average width of each antenna target in each new target screenshot, selecting a plurality of new target screenshots with small average widths, fitting the contour information of the antenna target in the selected plurality of new target screenshots into a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction.

Technical Field

The invention relates to a method and a system for automatically measuring antenna parameters of a base station, belonging to the technical field of information.

Background

In the process of putting the base station into use, because the base station is influenced by factors such as severe weather, project quality, improper maintenance and the like, the coverage area of the base station is easy to change, a communication blind area and a call drop rate in a certain range are increased, and the communication quality is directly influenced.

The main parameters influencing the coverage range of the base station comprise the direction angle, the mechanical inclination angle and the hanging height of the antenna, and for the inspection of the parameters of the antenna, the inspection by a tower worker is mainly relied on at present, but the inspection mode has low working efficiency, untimely response to emergency conditions and higher cost.

Patent application CN 201811126647.3 (application name: a method for measuring base station antenna engineering parameters, filing date: 2018.09.26, applicant: zhou ri communication planning and design limited) discloses a method for measuring base station antenna engineering parameters, which comprises the following steps: controlling the unmanned aerial vehicle to fly to the horizontal plane of the base station antenna, performing circular flight by taking the antenna holding pole as the circle center, recording a video, and recording GPS position information in real time; selecting an antenna holding pole as a target tracking area, tracking the antenna holding pole by using a meanShift algorithm for each frame of image, and setting areas on two sides of the holding pole as antenna detection areas; carrying out image contour detection processing on the antenna detection area to obtain the front and side contours of the antenna board; clustering the contours of the antenna boards by using a Kmeans algorithm, and carrying out image matching on the difference of the direction angles of the front and the side of the antenna boards to obtain the front and corresponding side contours of different antenna boards; and calculating all engineering parameters of the base station antenna plate by combining the recorded GPS position data with the obtained front surface and corresponding side profile of the antenna plate. The technical scheme has the following two problems: 1) extracting the front profile and the side profile of the antenna in the image by adopting a gray scale image and binarization operation, and tracking the area where the antenna is located by using a meanShift algorithm, wherein the area still has background information of a lot of interference, wherein the interference closer to the antenna has larger influence on the extraction of the profile, so that the measurement accuracy is low; 2) the RGB pixel information on the outline picture is used as clustering data of a Kmeans algorithm for clustering, and the technical scheme has poor feasibility because the RGB information on the picture is unreliable due to the influence of light and weather.

Therefore, how to effectively improve the automatic measurement accuracy of the antenna parameters of the base station and avoid external interference such as light, weather and the like has become a technical problem generally concerned by technicians.

Disclosure of Invention

In view of the above, an object of the present invention is to provide a method and a system for automatically measuring parameters of a base station antenna, which can effectively improve the accuracy of automatically measuring the parameters of the base station antenna and avoid external interference such as light, weather, etc.

In order to achieve the above object, the present invention provides a method for automatically measuring parameters of a base station antenna, comprising:

setting a round point, a radius and a flying height of a flying acquisition device flying around a point aiming at the type of an antenna of a base station;

secondly, the flight acquisition device flies for a circle by taking the center of the pole of the base station as a round point according to the preset radius and the preset flight height, and transmits the acquired video stream to the antenna parameter measuring device;

thirdly, the antenna parameter measuring device stores the video stream sent by the flight acquisition device;

fourthly, constructing and training a target tracking analysis model, wherein the target tracking analysis model is used for tracking all antenna targets appearing in the video stream, inputting the video stream sent by the flight acquisition device into the target tracking analysis model, and outputting all antenna target information tracked from the video stream, and the antenna target information comprises coordinate information of each antenna target, a target screenshot, and orientation and longitude and latitude information of the flight acquisition device corresponding to the picture;

step five, constructing and training a target screenshot screening model, wherein the target screenshot screening model is used for screening all target screenshots of each antenna target in the process of front side- > front side for the antenna target, inputting the antenna target information of each antenna target output by the target tracking analysis model into the target screenshot screening model one by one, and outputting a target screenshot sequence formed by all target screenshots of each antenna target in the process of front side- > front side;

step six, constructing and training a target contour fitting model, wherein the target contour fitting model is used for fitting contour information of the antenna target in the target screenshot, continuously inputting each target screenshot in a target screenshot sequence of each antenna target output by the target screenshot screening model into the target contour fitting model, and outputting the contour information of the antenna target in each target screenshot;

and step seven, calculating parameters of each antenna target according to all target screenshots in the target screenshot sequence of each antenna target and the profile information in the target screenshots, wherein the parameters include but are not limited to a direction angle, a mechanical inclination angle and a hanging height.

In order to achieve the above object, the present invention further provides a system for automatically measuring parameters of a base station antenna, which includes a flight acquisition device, a network communication device and an antenna parameter measuring device, wherein:

the flight acquisition device is used for flying for a circle by taking the center of the pole of the base station as a circular point according to the preset radius and the preset flight height, and transmitting the acquired video stream to the antenna parameter measuring device;

a network communication device for communication between the flight acquisition device and the antenna parameter measurement device,

the antenna parameter measuring device further comprises a model management unit and a parameter calculation unit, wherein:

the model management unit is used for constructing and training an internal model component, inputting the video stream sent by the flight acquisition device into the internal model component so as to track all the antenna target information appearing in the video stream, screening all target screenshots of each antenna target in the process of front side- > front side for each antenna target from all the antenna target information, and simultaneously fitting to obtain the contour information of the antenna targets in all the target screenshots;

and the parameter calculation unit is used for calculating parameters of each antenna target according to all target screenshots and profile information in the target screenshots of the process that each antenna target is on the front side surface- > front side surface, which are output by the model management unit, wherein the parameters include but are not limited to a direction angle, a mechanical inclination angle and a hanging height.

Compared with the prior art, the invention has the beneficial effects that: according to the invention, through automatic measurement of the antenna parameters of the base station, the intelligent level of intelligent inspection of the base station can be effectively improved, and the inspection cost is reduced; the invention uses the deep learning method, has good robustness to external interference such as light, weather and the like, and has high measurement precision and strong feasibility.

Drawings

Fig. 1 is a flow chart of a method for automatically measuring antenna parameters of a base station according to the present invention.

Fig. 2 is a flowchart illustrating the detailed steps of step 7B.

Fig. 3 is a schematic structural diagram of a system for automatically measuring parameters of a base station antenna according to the present invention.

Fig. 4 is a schematic diagram of the structure of the model management unit of the antenna parameter measuring apparatus.

Fig. 5 is a schematic diagram of the composition structure of the parameter calculating unit of the antenna parameter measuring device.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the accompanying drawings.

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the accompanying drawings.

As shown in fig. 1, the method for automatically measuring antenna parameters of a base station of the present invention includes:

setting a round point, a radius and a flying height of a flying acquisition device flying around a point aiming at the type of an antenna of a base station;

secondly, the flight acquisition device flies for a circle by taking the center of the pole of the base station as a round point according to the preset radius and the preset flight height, and transmits the acquired video stream to the antenna parameter measuring device;

thirdly, the antenna parameter measuring device stores the video stream sent by the flight acquisition device;

fourthly, constructing and training a target tracking analysis model, wherein the target tracking analysis model is used for tracking all antenna targets appearing in the video stream, inputting the video stream sent by the flight acquisition device into the target tracking analysis model, and outputting all antenna target information tracked from the video stream, and the antenna target information comprises coordinate information of each antenna target, a target screenshot, and orientation and longitude and latitude information of the flight acquisition device corresponding to the picture;

step five, constructing and training a target screenshot screening model, wherein the target screenshot screening model is used for screening all target screenshots of each antenna target in the process of front side- > front side for the antenna target, inputting the antenna target information of each antenna target output by the target tracking analysis model into the target screenshot screening model one by one, and outputting a target screenshot sequence formed by all target screenshots of each antenna target in the process of front side- > front side; screenshots of the process of screening the front side surface- > front side surface for each antenna target can be eliminated, and screenshots and related information of the back surface of the antenna can be eliminated;

step six, constructing and training a target contour fitting model, wherein the target contour fitting model is used for fitting contour information of the antenna target in the target screenshot, continuously inputting each target screenshot in a target screenshot sequence of each antenna target output by the target screenshot screening model into the target contour fitting model, and outputting the contour information of the antenna target in each target screenshot;

and step seven, calculating parameters of each antenna target according to all target screenshots in the target screenshot sequence of each antenna target and the profile information in the target screenshots, wherein the parameters include but are not limited to a direction angle, a mechanical inclination angle and a hanging height.

The third step can further comprise:

the antenna parameter measuring device stores the video stream sent by the flight acquisition device, then intercepts each picture in the video stream into a sub-picture in a certain area according to a certain intercepting area, and replaces an original picture in the video stream with the intercepted sub-picture; the size of the intercepting area can be determined according to the radius of the flying acquisition device flying around the point, so that the influence of interference factors outside the antenna area in the video stream can be eliminated.

The fourth step can also comprise:

counting the number of target screenshots corresponding to each target tracked by the target tracking analysis model and the number num of antennas of the base station, sequencing all tracked targets according to the sequence from the number of target screenshots to the number of target screenshots, finally selecting num targets sequenced in the front as antenna targets of the base station, namely, num targets are antenna targets tracked from a video stream, and finally outputting num antenna target information.

The target tracking analysis model in the fourth step can be realized by adopting a multi-target tracking algorithm yolov3_ depsort, the target screenshot screening model in the fifth step can be realized by adopting a fine-grained classification algorithm DFL-CNN, and the target contour fitting model in the sixth step can be realized by adopting a residual error network resnet to fit the target contour.

In step seven, the parameters of each antenna target may include, but are not limited to, a direction angle, a mechanical tilt angle, a hang height, etc., where:

1. when calculating the direction angle of each antenna target, the seventh step may further include:

step 7A, calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, then selecting a plurality of target screenshots with large average widths for each antenna target, and calculating the direction angle of each antenna target according to the average width: the direction angle of each antenna target is the average value of the orientation of the flight acquisition device corresponding to the plurality of target screenshots selected for the antenna target,

2. when calculating the mechanical tilt angle of each antenna target, the seventh step may further include:

step 7B, calculating the average width of each antenna target in each target screenshot according to the contour information of each antenna target in each target screenshot of the target screenshot sequence, selecting a plurality of target screenshots with small average widths for each antenna target, and calculating the mechanical inclination angle of each antenna target according to the target width: fitting the profile information of the antenna targets in the selected multiple target screenshots to form a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction,

3. when calculating the hanging height of each antenna target, the seventh step may further include:

and 7C, calculating the hanging height of each antenna target, namely calculating the difference between the altitude of the flight acquisition device at the tower base when the flight acquisition device just takes off and the altitude of the flight around the point.

As shown in fig. 2, in order to perform an accurate fitting on the antenna profile, so as to improve the measurement accuracy of the mechanical tilt angle, step 7B may further include:

step 7B1, selecting a plurality of target screenshots with small average widths for each antenna target according to the average width of each antenna target in each target screenshot of the target screenshot sequence, wherein the selected target screenshots are called original target screenshots;

step 7B2, extracting an original image corresponding to each original target screenshot from the video stream, then expanding the width and height of the original target screenshot by a plurality of times (for example, expanding the width to 3 times and expanding the height to 2 times), and then taking the original target screenshot frame as a center and intercepting a new target screenshot from the original image according to the expanded width and height;

and 7B3, inputting a plurality of new target screenshots obtained by intercepting each antenna target into a target contour fitting model, thereby outputting contour information of each antenna target in each new target screenshot, then calculating the average width of each antenna target in each new target screenshot, selecting a plurality of new target screenshots with small average widths, and finally fitting the contour information of the antenna target in the selected plurality of new target screenshots into a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction.

When the measured direction angle, mechanical inclination angle or hanging height of each antenna target exceeds a set threshold value, an alarm signal is sent out, and the threshold value can be set according to the value of the antenna when being installed.

As shown in fig. 3, the system for automatically measuring antenna parameters of a base station of the present invention includes a flight acquisition device, a network communication device and an antenna parameter measuring device, wherein:

the flight acquisition device is used for flying for a circle by taking the center of the pole of the base station as a circular point according to the preset radius and the preset flight height, and transmitting the acquired video stream to the antenna parameter measuring device; the method comprises the following steps that (1) the round points, the radius and the flying height of a flying acquisition device flying around points can be preset according to the type of an antenna of a base station;

a network communication device for communication between the flight acquisition device and the antenna parameter measurement device (such as a 5G network),

the antenna parameter measuring device may further include a model management unit and a parameter calculation unit, wherein:

the model management unit is used for constructing and training an internal model component, inputting the video stream sent by the flight acquisition device into the internal model component so as to track all the antenna target information appearing in the video stream, screening all target screenshots of each antenna target in the process of front side- > front side for each antenna target from all the antenna target information, and simultaneously fitting to obtain the contour information of the antenna targets in all the target screenshots;

and the parameter calculation unit is used for calculating parameters of each antenna target according to all target screenshots and profile information in the target screenshots of the process that each antenna target is on the front side surface- > front side surface, which are output by the model management unit, wherein the parameters can include but are not limited to a direction angle, a mechanical inclination angle, a hanging angle and the like.

The flight acquisition device further comprises an aircraft (such as an unmanned aerial vehicle) and remote control equipment (such as a smart phone). The high-definition camera and the navigation recording unit are mounted on the aircraft, the high-definition camera is used for collecting video stream data of the antenna in the base station, the navigation recording unit is used for recording the flight height and the longitude and latitude information of the aircraft, and the remote control equipment is used for controlling the flight of the aircraft.

As shown in fig. 4, the model management unit of the antenna parameter measuring apparatus may further include:

the system comprises a label training component, a model contour fitting component and a model tracking component, wherein the label training component is used for labeling the training data of the model component inside the mode management unit, and constructing and training to generate a corresponding model component according to the labeled training data, and the model component comprises a target tracking analysis model component, a target screenshot screening model component and a target contour fitting model component;

the video stream processing component is used for storing the video stream sent by the flight acquisition device, then intercepting each picture in the video stream into a sub-picture in a certain area according to a certain intercepting area, and replacing an original picture in the video stream with the intercepted sub-picture; the size of the intercepting area can be determined according to the radius of the flying acquisition device flying around the point, so that the influence of interference factors outside the antenna area in the video stream can be eliminated;

the target tracking analysis model component is used for reading the stored video stream, tracking all antenna targets appearing in the video stream and then outputting all tracked antenna target information, wherein the antenna target information comprises coordinate information of each antenna target, a target screenshot, and orientation and longitude and latitude information of a flight acquisition device corresponding to a picture; the component can be realized by adopting a multi-target tracking algorithm yolov3_ deepsort;

the target screenshot screening model component is used for receiving antenna target information of each antenna target output by the target tracking analysis model component, screening all target screenshots of the process that each antenna target is on the front side- > front side for each antenna target, and then outputting a target screenshot sequence of each antenna target, wherein the target screenshot sequence is formed by all target screenshots of the process that the antenna target is on the front side- > front side; the component can be realized by adopting a fine-grained classification algorithm DFL-CNN algorithm, and screenshots of the process of front side- > front side are screened out for each antenna target, so that screenshots and related information of the back side of the antenna can be eliminated;

the target contour fitting model component is used for receiving each target screenshot in the target screenshot sequence of each antenna target output by the target screenshot screening model component, fitting to obtain contour information of the antenna target in each target screenshot, and then outputting the contour information of the antenna target in the target screenshot; this component may employ a residual network resnet algorithm to achieve a fit to the target profile.

The target tracking analysis model component may further comprise:

the antenna target selection component is used for counting the number of target screenshots corresponding to each target tracked by the target tracking analysis model and the number num of antennas of the base station, then sequencing all tracked targets according to the sequence of the number of target screenshots from multiple to few, finally selecting num targets sequenced in the front as antenna targets of the base station, namely the num targets are antenna targets tracked from a video stream, and finally outputting the num antenna target information.

As shown in fig. 5, the parameter calculating unit may further include:

the direction angle measuring component is used for calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, then selecting a plurality of target screenshots with large average widths for each antenna target, and calculating the direction angle of each antenna target according to the average width: the direction angle of each antenna target is the average value of the orientation of the flight acquisition device corresponding to the plurality of target screenshots selected for the antenna target;

the mechanical inclination angle measuring component is used for calculating the average width of each antenna target in each target screenshot according to the profile information of each antenna target in each target screenshot of the target screenshot sequence, selecting a plurality of target screenshots with small average widths for each antenna target, and calculating the mechanical inclination angle of each antenna target according to the target width: fitting the contour information of the antenna targets in the picked multiple target screenshots to form a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction;

the hanging height measuring component is used for calculating the hanging height of each antenna target, namely the difference value of the altitude of the flight acquisition device at the tower base during the takeoff and the altitude of the winding point flight;

and the abnormity warning component is used for judging that when the measured parameters (including the direction angle, the mechanical inclination angle, the hanging angle and the like) of each antenna target exceed a set threshold value, a warning signal is sent out, and the threshold value can be set according to the value of the antenna when the antenna is installed.

In order to accurately fit the antenna profile and thereby improve the measurement accuracy of the mechanical tilt angle, the mechanical tilt angle measurement unit may further include:

the mechanical inclination angle calculation component is used for selecting a plurality of target screenshots with small average widths for each antenna target according to the average width of each antenna target in each target screenshot of the target screenshot sequence, and the selected target screenshots are called original target screenshots; then, extracting an original image corresponding to each original target screenshot from the video stream, respectively expanding the width and the height of the original target screenshot by a plurality of times (for example, expanding the width to 3 times and expanding the height to 2 times), and taking the original target screenshot frame as a center and intercepting a new target screenshot from the original image according to the expanded width and height; and finally, inputting a plurality of new target screenshots obtained by intercepting each antenna target into a target contour fitting model component, thereby obtaining contour information of each antenna target in each new target screenshot, calculating the average width of each antenna target in each new target screenshot, selecting a plurality of new target screenshots with small average widths, fitting the contour information of the antenna target in the selected plurality of new target screenshots into a straight line, wherein the mechanical inclination angle of each antenna target is the included angle between the fitted straight line and the vertical direction.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.