CN107632615B - Autonomous flight four-rotor tunnel passing method based on visual inspection - Google Patents

Abstract

The invention discloses a method for traveling through a tunnel with an autonomous flying quadrotor based on visual inspection. The autonomous flying quadrotor is used to autonomously find an entrance and enter the tunnel. Inside the tunnel, a bottom camera is used to realize autonomous navigation and flight for the autonomous flying quadrotor. At the same time, the camera is used to detect the cables inside the tunnel and the surrounding conditions in real time, and finally fly out of the tunnel autonomously. The whole process does not require manual intervention. It has good autonomous flight and navigation capabilities, and also has a wide detection range, high speed and high accuracy.

Description

A method of autonomous flying quadrotor traveling through tunnel based on visual inspection

technical field

The invention belongs to the technical field of flight control, and more particularly relates to a method for traveling through a tunnel based on an autonomous flying quadrotor based on visual inspection.

Background technique

In recent years, in the development of large cities, it has become mainstream to replace traditional overhead lines with underground cables. Statistics show that in many modern cities, the proportion of underground transmission lines has exceeded 70%. As a fully enclosed underground structure that accommodates a large number of cables and has passages for installation and inspection, cable tunnels are the best way to carry underground cables. With the popularization of underground cables, how to regularly detect closed tunnels and prevent emergencies has become an urgent problem.

Therefore, it has been proposed to use orbital robots or humans to perform periodic inspections. Among them, the tracked robot mainly relies on the track to move. The track is laid on the top or bottom of the tunnel. The robot travels through the tunnel along the track through the rotation of its own motor or the transmission of the track, and carries out the tunnel through the relevant sensors carried by the robot. For real-time detection, the rail robot usually needs to be deployed in the target tunnel in advance, that is, a tunnel needs to be equipped with a rail robot and supporting rail facilities, and it is still necessary to regularly send people down to the tunnel to check and maintain the rail robot itself, so use And the maintenance cost is high, without the need for advance deployment, the unmanned device that can enter and exit the tunnel autonomously, conduct full autonomous and comprehensive inspection of the tunnel has not yet appeared.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the deficiencies of the prior art, and to provide a method for traveling through a tunnel based on an autonomous flight quadrotor based on visual inspection. The autonomous flight quadrotor is used to autonomously search for an entrance and enter the tunnel, realize autonomous navigation and flight inside the tunnel, and detect in real time. The cables inside the tunnel and the surrounding conditions finally flew out of the tunnel autonomously.

In order to achieve the above-mentioned purpose of the invention, the present invention based on the visual inspection of the autonomous flying quadrotor traveling through the tunnel method, is characterized in that, comprises the following steps:

(1) Add a yellow marking line along the tunnel center line, install a USB camera at the bottom of the autonomous flying quadrotor, and install LED fill light to ensure that the camera works normally inside the tunnel, on the front and rear of the nose and the right side of the fuselage A miniature laser ranging sensor is installed;

(2) Control the autonomous flying quadrotor to drive into the tunnel wellhead

(2.1) Power-on initialization of the autonomous flight quadrotor;

(2.2) Manually take off the autonomous flight quadrotor, make the autonomous flight quadrotor a distance h ₁ from the ground, and then manually control the autonomous flight quadrotor to drive into the tunnel near the wellhead. At this time, switch the autonomous flight quadrotor to fully autonomous flight through the remote control model;

(2.3), the flight control module calculates the real-time attitude angle of the autonomous flying quadrotor;

(2.4) The autonomous flying quadrotor uses the laser ranging sensor on the nose and the right side of the fuselage to locate the real-time position information of the current autonomous flying quadrotor, and fuses it with the accelerometer to estimate the flight speed, combining the real-time attitude angle, position information and flight speed speed, the flight control module controls the autonomous flying quadrotor to fly to the center of the tunnel wellhead;

(3) Control the autonomous flying quadrotor to drive into the tunnel entrance from the center of the tunnel wellhead

The flight control module controls the autonomous flight quadrotor to descend vertically from the center of the tunnel wellhead. When it descends to the top plane of the tunnel, the laser ranging sensors on the tail and the right side of the fuselage are used to locate the real-time position information of the autonomous flight quadrotor. The current position information and real-time attitude angle continue to descend vertically, and when it descends to the set cruising altitude, it will reach the tunnel entrance;

(4) The flight control module controls the autonomous flying quadrotor to travel through the tunnel

(4.1) The USB camera periodically collects the environmental images in the tunnel with the cooperation of the LED fill light;

(4.2), open the line inspection thread of the autonomous flight quadrotor, and obtain the actual distance d of the autonomous flight quadrotor deviating from the yellow marking line and the deviation angle θ between the flight direction of the autonomous flight quadrotor and the tunnel direction;

(4.2.1) Convert the environmental image to the HSV color space, and then perform a binarization process on the environmental image according to the yellow HSV range value to obtain a binarized image;

(4.2.2), use the open operation to filter out the white noise in the binarized image, and then use the closed operation to connect the discrete white areas in the binarized image to obtain an ideal binarized image;

(4.2.3), take the first n rows of images above the ideal binarized image, for the first row, start from the leftmost left of the ideal binarized image to the right to find the left edge of the white area, and start from the rightmost to the left For the right edge of the white area, get the white area of the line of the image, and then find the center of the white area; for the remaining n-1 lines of images, find the left and right edges in the previous line within m pixels near the position of the left and right edges to find the line of this line. Left and right edges, get the white area of the image in this row, and then find the center of the white area corresponding to the image in this row;

(4.2.4) Make the difference between the center of the white area of the n-line image and the absolute center of the corresponding line in the ideal binarized image, then accumulate and average, and use the average difference as the pixel of the autonomous flying quadrotor that deviates from the white area quantity;

(4.2.5) According to the pixel amount of the autonomous flight quadrotor deviating from the white area, the current real-time altitude information of the autonomous flight quadrotor, and the focal length information of the USB camera, the pinhole imaging model is used to calculate the deviation of the autonomous flight quadrotor from the yellow marking line the actual distance d;

(4.2.6) In the n-line image, select the center of the white area extracted from the n/3th line and the 2n/3th line, and calculate the difference e between the white area centers of these two lines by arctan(e/ (n/3)) Calculate the arctangent value to obtain the deflection angle θ between the flight direction of the autonomous flying quadrotor and the direction of the tunnel;

(4.3), open the optical flow thread of the autonomous flying quadrotor, and obtain the real-time flight speed v of the autonomous flying quadrotor;

(4.3.1) At the initial time t ₀ when the autonomous flying quadrotor is powered on, run the corner detection algorithm on the environmental image collected by the USB camera at the initial time to obtain k Shi-Tomasi corners, and then use the k Shi-Tomasi corners The Tomasi corner point is used as the Shi-Tomasi corner point set a(t ₀ ) of the environment image at this moment;

(4.3.2) In the rest of the following moments, periodically collect the environmental image, and use the LK optical flow motion estimation method to process the environmental image at the current moment to find the Shi-Tomasi corner point set a of the environmental image at the previous moment ( The matching points of the Shi-Tomasi corner points in t-1) are formed into a point set b(t), and then the Shi-Tomasi corner points that fail to find matching points are eliminated from the point set a(t-1), so that Get the Shi-Tomasi corner set a(t) at the current moment;

(4.3.3), determine whether the number of Shi-Tomasi corner points in a(t) is greater than the threshold M, if so, go to step (4.3.4), otherwise return to step (4.3.1);

(4.3.4) Accumulate and average the coordinate deviations between the corresponding points in a(t) and b(t) to obtain the flight speed information v ₁ in the pixel sense;

(4.3.5), use the real-time altitude and real-time attitude angle of the current flight of the autonomous flying _quadrotor to compensate v1, and obtain the flight speed information v2 _;

(4.3.6), use the accelerometer and gyroscope to estimate the inertial navigation speed information v ₃ , and then perform Kalman filter fusion with v ₃ and v ₂ to obtain the real-time flight speed v of the autonomous flying quadrotor;

(4.4) According to the actual distance d of the autonomous flight quadrotor deviating from the yellow marking line, the real-time flight speed v and the deviation angle θ between the flight direction of the autonomous flight quadrotor and the tunnel direction, the flight control module controls the autonomous flight quadrotor to fly, so that d and θ approach 0, and cruise forward, and reach the tunnel exit when the distance measured by the nose laser ranging sensor is half the width of the tunnel;

(5) Control the autonomous flying quadrotor to leave the tunnel wellhead from the center of the tunnel exit

Using the laser ranging sensor on the nose and the right side of the fuselage to locate the real-time position information of the autonomous flying quadrotor, the flight control module controls the autonomous flying quadrotor to rise vertically from the tunnel exit. When it rises to the top plane of the tunnel, the tail is used instead. And the laser ranging sensor on the right side of the fuselage locates the real-time position information of the autonomous flying quadrotor. The flight control module controls the autonomous flying quadrotor to continue to rise vertically from the center of the tunnel exit. When it rises above the tunnel wellhead, it leaves the tunnel wellhead.

The purpose of the invention of the present invention is achieved in this way:

The present invention is based on a visual inspection-based method for autonomously flying quadrotors to travel through tunnels. The autonomous flight quadrotors are used to autonomously find the entrance and enter the tunnel. Inside the tunnel, the autonomous flight quadrotors are autonomously navigated through the bottom camera, and the camera is used while flying. It conducts real-time detection of the cables inside the tunnel and the surrounding conditions, and finally flies out of the tunnel autonomously. The whole process does not require manual intervention. It has good autonomous flight and navigation capabilities, and also has the characteristics of wide detection range, high speed and high accuracy.

At the same time, the present invention based on the visual inspection of the autonomous flying quadrotor traveling through the tunnel method also has the following beneficial effects:

(1) At present, most of the inspection robots used in tunnels are mainly track type, and this method needs to lay tracks on the top of the tunnel or on the ground, and the robot needs to be deployed in the tunnel in advance, that is, each tunnel needs to be equipped with a robot. The cost is high, the inspection range is limited, and the present invention does not need to lay tracks, and has the advantage of low cost;

(2) The present invention can fully autonomously realize the entry and exit of the tunnel, without being deployed in the target tunnel in advance, one robot can complete the inspection task of all the tunnels in an area, and the maintenance work of the robot itself is also more difficult than the rails that need to go down for maintenance. Robots are much more convenient;

(3), the present invention has better passability in the tunnel, as long as a certain space is left, the quadrotor can carry out detection and inspection tasks in the tunnel, and is less affected by the environment in the tunnel;

(4) The present invention can replace the manual downhole to complete the inspection and inspection task, and the use and maintenance process does not require workers to go downhole, greatly shortens the inspection time, ensures the safety of personnel, reduces the use and maintenance costs, and is safer and more reliable.

Description of drawings

Fig. 1 is the sectional structure schematic diagram of the tunnel;

Fig. 2 is the top view structure schematic diagram of tunnel;

Fig. 3 is the flow chart of the present invention's self-flying quadrotor traveling through tunnel method based on visual inspection;

Detailed ways

The specific embodiments of the present invention are described below with reference to the accompanying drawings, so that those skilled in the art can better understand the present invention. It should be noted that, in the following description, when the detailed description of known functions and designs may dilute the main content of the present invention, these descriptions will be omitted here.

Example

In this embodiment, the tunnel can be abstracted into the structure shown in FIG. 1 and FIG. 2 , and the entrance and exit of the tunnel are closed on three sides. The present invention will be described in detail below with reference to FIGS. 1 and 2 . .

As shown in FIG. 3 , a method of autonomously flying quadrotors traveling through tunnels based on visual inspection of the present invention specifically includes the following steps:

S1. Add a yellow marking line along the center line of the tunnel, install a USB camera at the bottom of the autonomous flying quadrotor, and install LED fill light to ensure that the camera works normally inside the tunnel. Install miniature laser ranging sensor;

S2. Control the autonomous flying quadrotor to drive into the tunnel wellhead

S2.1. Power-on initialization of the autonomous flight quadrotor;

S2.2. Manual take-off of the autonomous flying quadrotor, so that when the autonomous flying quadrotor is at a certain height from the ground, there is no strict requirement for this height, and then manually remote control the autonomous flying quadrotor and drive it into the vicinity of the tunnel wellhead. Generally, the autonomous flying quadrotor flies to When the nose is about 1 meter away from the opposite side wall, switch the autonomous flight quadrotor to fully autonomous flight mode through the remote control;

S2.3. The flight control module calculates the real-time attitude angle of the autonomous flying quadrotor;

S2.4. The autonomous flying quadrotor uses the laser ranging sensor on the nose and the right side of the fuselage to locate the real-time position information of the current autonomous flying quadrotor, and fuses it with the accelerometer to estimate the flight speed, combining the real-time attitude angle, position information and flight speed speed, the flight control module controls the autonomous flying quadrotor to fly to the center of the tunnel wellhead;

S3. Control the autonomous flying quadrotor to drive into the tunnel entrance from the center of the tunnel wellhead

The flight control module controls the autonomous flying quadrotor to descend vertically from the center of the tunnel wellhead. When descending to the top plane of the tunnel, due to the disappearance of the forward wall, the distance measured by the forward laser ranging sensor will jump greatly or directly. At the same time, the backward laser ranging sensor can start to obtain valid data. Therefore, the control algorithm immediately activates the laser ranging sensor on the tail and the right side of the fuselage by sensing the jump in the forward ranging. In order to obtain the real-time position information of the autonomous flying quadrotor, it continues to descend vertically according to the current position information and real-time attitude angle, and when it descends to the set cruising altitude, it reaches the tunnel entrance;

S4. The flight control module controls the autonomous flying quadrotor to travel through the tunnel

S4.1, the USB camera periodically collects the environmental images in the tunnel with the cooperation of the LED fill light;

S4.2. Open the line inspection thread of the autonomous flight quadrotor, and obtain the actual distance d of the autonomous flight quadrotor deviating from the yellow marking line and the deviation angle θ between the flight direction of the autonomous flight quadrotor and the tunnel direction;

S4.2.1. Convert the environmental image to the HSV color space, and then perform binarization processing on the environmental image according to the HSV range value of yellow to obtain a binarized image, where the HSV range value of yellow is: h:26-34; s: 43-255; v: 46-255;

S4.2.2. Use the open operation to filter out the white noise in the binarized image, and then use the closed operation to connect the discrete white areas in the binarized image to obtain an ideal binarized image;

S4.2.3. Take the first 100 lines of images above the ideal binarized image. For the first line, start from the leftmost edge of the ideal binarized image to the right to find the left edge of the white area, and start from the rightmost to the left to search for the white area. The right edge, get the white area of the line image, and then find the center of the white area; for the remaining 99 lines of images, search for the left and right edges of the current line within 30 pixels near the left and right edges of the previous line, and get The white area of the image in this line, and then find the center of the white area corresponding to the image in this line;

S4.2.4. Make the difference between the center of the white area of the 100-line image and the absolute center of the corresponding line in the ideal binarized image, and then accumulate and average, and use the average difference as the amount of pixels that the autonomous flying quadrotor deviates from the white area;

S4.2.5. According to the pixel amount of the autonomous flight quadrotor deviating from the white area, the current real-time altitude information of the autonomous flight quadrotor, and the focal length information of the USB camera, use the pinhole imaging model to calculate the actual deviation of the autonomous flight quadrotor from the yellow marking line. distance d;

S4.2.6. In the 100-line image, select the center of the white area extracted from the 33rd line and the 66th line, and calculate the arc tangent value by arctan(e/33) for the difference e between the white area centers of the two lines , get the deflection angle θ between the flight direction of the autonomous flying quadrotor and the direction of the tunnel;

S4.3. Open the optical flow thread of the autonomous flying quadrotor to obtain the real-time flight speed v of the autonomous flying quadrotor;

S4.3.1. At the initial time t ₀ when the autonomous flying quadrotor is powered on, run the corner detection algorithm on the environmental image collected by the USB camera at the initial time to obtain 200 Shi-Tomasi corners, and then use these 200 Shi-Tomasi corners point as the Shi-Tomasi corner point set a(t ₀ ) of the environment image at this moment;

S4.3.2. At the rest of the following moments, periodically collect environmental images, and use the LK optical flow motion estimation method to process the environmental images at the current moment to find the Shi-Tomasi corner point set a(t- The matching points of the Shi-Tomasi corner points in 1) are formed into a point set b(t), and then the Shi-Tomasi corner points that fail to find a matching point are removed from the point set a(t-1), so as to obtain the current The Shi-Tomasi corner set a(t) at time;

S4.3.3. Determine whether the number of Shi-Tomasi corner points in a(t) is greater than the threshold 30, if so, go to step S4.3.4, otherwise return to step S4.3.1;

S4.3.4. Accumulate and average the coordinate deviations between the corresponding points in a(t) and b(t) to obtain the flight speed information v ₁ in the pixel sense;

S4.3.5. Compensate v1 by using the real-time altitude and real-time attitude angle of the current flight of the autonomous flying _quadrotor to obtain the flight speed information v2 _;

S4.3.6. Use the accelerometer and gyroscope to estimate the inertial navigation speed information v ₃ , and then perform Kalman filter fusion with v ₃ and v ₂ to obtain the real-time flight speed v of the autonomous flying quadrotor;

S4.4. According to the actual distance d of the autonomous flight quadrotor deviating from the yellow marking line, the real-time flight speed v, and the deviation angle θ between the flight direction of the autonomous flight quadrotor and the tunnel direction, the autonomous flight quadrotor is controlled by the flight control module to fly, so that d and θ approach 0, and cruise forward, and reach the center of the tunnel exit when the distance measured by the laser ranging sensor on the nose is half the width of the tunnel;

S5. Control the autonomous flying quadrotor to leave the tunnel wellhead from the center of the tunnel exit

Using the laser ranging sensor on the nose and the right side of the fuselage to locate the real-time position information of the autonomous flying quadrotor, the flight control module controls the autonomous flying quadrotor to rise vertically from the center of the tunnel exit. When it rises to the top plane of the tunnel, use The laser ranging sensors on the tail and the right side of the fuselage locate the real-time position information of the autonomous flying quadrotor. The flight control module controls the autonomous flying quadrotor to continue to rise vertically from the center of the tunnel exit. When it rises above the tunnel wellhead, it leaves the tunnel. wellhead.

Although the illustrative specific embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, As long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the inventive concept are included in the protection list.

Claims (1)

1. The method for the autonomous flight four-rotor-wing tunnel passing based on the visual inspection is characterized by comprising the following steps of:

(1) adding a yellow identification line along the central line of the tunnel, mounting a USB camera at the bottom of the four self-flying rotors, additionally mounting L ED light supplement lamps to ensure that the camera normally works in the tunnel, and additionally mounting miniature laser ranging sensors at the front and rear parts of the camera and the right side of the machine body;

(2) four rotors of controlling independent flight are sailed into tunnel well head

(2.1) carrying out power-on initialization on four autonomous flight rotors;

(2.2) manually taking off and autonomously flying four rotors to make autonomous flyingDistance h between four rotors and ground₁Then manually remotely controlling the four self-flying rotors to drive into the vicinity of the tunnel wellhead, and switching the four self-flying rotors into a full self-flying mode through a remote controller;

(2.3) calculating the real-time attitude angle of the four autonomically flying rotors by the flight control module;

(2.4) positioning real-time position information of the current autonomous flight four-rotor by using laser ranging sensors on the right sides of the machine head and the machine body through the autonomous flight four-rotor, fusing the real-time position information and the current autonomous flight four-rotor with an accelerometer to estimate flight speed, and controlling the autonomous flight four-rotor to fly to the central position of a tunnel wellhead by a flight control module by combining a real-time attitude angle, the position information and the flight speed;

(3) the four rotors which are controlled to fly autonomously are driven into the tunnel entrance from the center of the tunnel wellhead

The flight control module controls the autonomous flight quadrotors to vertically descend from the center of a tunnel wellhead, when the autonomous flight quadrotors descend to a tunnel top plane, the laser ranging sensors on the tail and the right side of the machine body are used for positioning real-time position information of the autonomous flight quadrotors, the autonomous flight quadrotors continue to vertically descend according to the current position information and the real-time attitude angle, and when the autonomous flight quadrotors descend to a set cruising height, the autonomous flight quadrotors reach a tunnel entrance;

(4) four-rotor-wing passing tunnel controlled by flight control module to fly autonomously

(4.1) periodically collecting an environment image in the tunnel by the USB camera under the coordination of an L ED supplementary lighting lamp;

(4.2) starting a line patrol thread of the four self-flying rotors to obtain the actual distance d of the four self-flying rotors deviating from the yellow identification line and the deflection angle theta of the four self-flying rotors in the flying direction and the tunnel trend;

(4.2.1) converting the environment image into an HSV color space, and then carrying out binarization processing on the environment image according to the yellow HSV range value to obtain a binarization image;

(4.2.2) filtering white noise points in the binary image by utilizing open operation, and connecting discrete white areas in the binary image by utilizing closed operation to obtain an ideal binary image;

(4.2.3) taking the front n rows of images above the ideal binary image, searching the left edge of a white area from the leftmost side of the ideal binary image to the right and searching the right edge of the white area from the rightmost side to the left for the first row to obtain the white area of the row of images, and then finding out the center of the white area; for the rest n-1 lines of images, searching the left edge and the right edge of the line in the range of m pixels near the left edge and the right edge searched in the previous line of the line to obtain the white area of the line of images, and then finding out the center of the white area corresponding to the line of images;

(4.2.4) making a difference between the center of the white area of the found n rows of images and the absolute center of the corresponding row in the ideal binary image, accumulating and averaging, and taking the average difference value as the pixel quantity of the autonomous flight quadrotors deviating from the white area;

(4.2.5) calculating the actual distance d of the autonomous flight four rotors deviating from the yellow identification line by using a pinhole imaging model according to the pixel quantity of the autonomous flight four rotors deviating from the white area, the current real-time height information of the autonomous flight four rotors and the focal length information of the USB camera;

(4.2.6) selecting the white area centers extracted from the n/3 th line and the 2n/3 th line from the n-line images, and calculating the arctan value of the difference e of the white area centers of the two lines by arctan (e/(n/3)) to obtain the drift angle theta of the flight direction of the autonomous flight four rotors and the tunnel trend;

(4.3) starting an optical flow thread of the four independent flying rotors to obtain the real-time flying speed v of the four independent flying rotors;

(4.3.1) at the initial moment t of the powering-on of the four autonomy flight rotors₀Operating a corner detection algorithm on an environment image acquired at the initial moment of the USB camera to obtain k Shi-Tomasi corners, and taking the k Shi-Tomasi corners as a Shi-Tomasi corner set a (t) of the environment image at the moment₀)；

(4.3.2) periodically acquiring the environment image at the rest of subsequent moments, processing the environment image at the current moment by using an L K optical flow motion estimation method, searching matching points of Shi-Tomasi angular points in a Shi-Tomasi angular point set a (t-1) of the environment image at the previous moment, forming a point set b (t), and removing Shi-Tomasi angular points, at which the matching points cannot be found, from the point set a (t-1) so as to obtain a Shi-Tomasi angular point set a (t) at the current moment;

(4.3.3), judging whether the number of the Shi-Tomasi angular points in a (t) is larger than a threshold value M, if so, entering a step (4.3.4), otherwise, returning to the step (4.3.1);

(4.3.4) accumulating and averaging coordinate deviations between corresponding points in a (t) and b (t) to obtain pixel-wise flight speed information v₁；

(4.3.5) utilizing the real-time altitude and the real-time attitude angle of the current flight of the four independent flight rotors to v₁Compensating to obtain flight speed information v₂；

(4.3.6) estimating inertial navigation speed information v by using accelerometer and gyroscope₃Then v is further determined₃And v₂Performing Kalman filtering fusion to obtain the real-time flight speed v of the autonomous flight quadrotors;

(4.4) according to the actual distance d of the autonomous flight four rotors deviating from the yellow identification line, the real-time flight speed v and the deflection angle theta between the flight direction of the autonomous flight four rotors and the tunnel trend, controlling the autonomous flight four rotors to fly through the flight control module, enabling d and theta to approach 0, cruising and flying forwards, and enabling the distance measured by the locomotive laser ranging sensor to reach the tunnel outlet when the distance is half the width of the tunnel;

(5) the four rotors of the autonomous flight are controlled to drive away from the well head of the tunnel from the center position of the tunnel outlet

Utilize the laser range finding sensor location on aircraft nose and fuselage right side to independently fly the real-time positional information of four rotors, flight control module control independently flies four rotors and rises perpendicularly from tunnel exit central point department, when rising to tunnel top plane, change into the laser range finding sensor location on aircraft tail and fuselage right side and independently flies four rotors's of real-time positional information, flight control module control independently flies four rotors and continues to rise perpendicularly from tunnel exit central point, when rising to being higher than tunnel well head after driving away from tunnel well head.

Images