TWI783748B - UAV obstacle avoidance flight control image recognition method, system and application using deep learning - Google Patents

Abstract

The present invention discloses a UAV obstacle avoidance flight control image recognition method, system and application using deep learning. It includes a UAV, a wireless communication unit and an information processing unit. The UAV includes a flight control module and an image capture device. The flight control module controls the flight of the UAV according to the flight path, and the image capture device continuously captures the flight status of the UAV and forms a flight image. The information processing unit receives flight images through the wireless communication unit. The information processing unit includes a deep learning algorithm module and a built-in feature database with object feature samples. Each object feature sample is defined with an object name. The deep learning algorithm module is used for The features of the object are extracted from each flight image and input into the feature database to predict the probability of matching with the object feature sample. When the matching probability is greater than the preset probability, the identification result information of the object matching the object name is output, and it is judged whether If it is an obstacle, if the judgment result is yes, then a correction control signal for correcting the flight path will be generated, so that the UAV can avoid the obstacle, so that the task of identifying obstacles in a small area of the flight path can be realized in real time, and can be accurately located. The actual position of the obstacle can be effectively avoided.

Description

UAV obstacle avoidance flight control image recognition method, system and application using deep learning

The present invention relates to a UAV obstacle avoidance flight control image recognition method, system and application using deep learning, especially a method that can realize real-time identification of obstacles in a small area of the flight path and can accurately locate the actual position of the obstacle UAV obstacle avoidance flight control technology.

By the way, as technologies such as drone manufacturing and flight control become more mature and complete, the scope of application of drones has become more popular. For example, tasks including environmental data detection, patrols in specific areas, land conservation, emergency search and rescue, meteorological observation, and communication relay can all be performed by drones. In addition, some related companies have used drones to collect photogrammetry and remote sensing detection data, etc. In terms of science and people's livelihood, unmanned aerial vehicles can also provide rescue and search, pesticide spraying, atmospheric data collection, and fixed-point shooting, etc. Therefore, in recent years, drones have indeed been widely concerned and favored by various research institutes, business circles and relevant government departments.

As far as we know, the previous patent application of using artificial intelligence technology to guide the autonomous flight of drones is shown in the new No. M558760 "UAV artificial intelligence module". The control module of the patent includes a first wireless communication unit, an artificial intelligence machine learning processing unit electrically connected to the first wireless communication unit, a data storage unit and an unmanned aerial vehicle electrically connected to the artificial intelligence machine learning processing unit. The unmanned aerial vehicle configuration includes an image capture unit, electrically connected to the image capture unit The second wireless communication unit and the power unit electrically connected to the driving component, the image capture unit, and the second wireless communication unit, the image capture unit and the artificial intelligence machine learning through the second wireless communication unit and the first wireless communication unit Processing unit communication connection. Although the artificial intelligence machine learning processing unit of the patent can control the unmanned aerial vehicle based on the machine learning method based on the image information obtained by the image capture unit, and the anti-collision detection unit can confirm the environmental obstacles around the unmanned aerial vehicle and send it back to The machine learning processing unit facilitates the establishment of an actual three-dimensional space concept and avoids obstacles; however, this patent still has the following disadvantages:

1. Its artificial intelligence machine learning processing unit does not have the function of obstacle image recognition, so an additional anti-collision detection unit must be installed, resulting in an increase in the cost of components.

2. The anti-collision detection part is fixed on one side of the UAV, so that it cannot change its detection direction along with the UAV's travel direction. Since the travel direction and detection direction are not synchronized, it is indeed more It is easy to hit obstacles, thus causing property damage caused by drone damage.

It can be known from the above that the patented obstacle avoidance flight technology is indeed not perfect, and there is still a need for further improvement, and based on the urgent needs of related industries, the reason is that the inventors have made continuous efforts in research and development. Finally develop a set of the present invention that is different from above-mentioned prior art.

The first purpose of the present invention is to provide a UAV obstacle avoidance flight control image recognition method, system and application using deep learning, mainly to realize the task of instantly identifying obstacles in a small area of the flight path and accurately locate The actual position of the obstacle can be obtained, and the obstacle can be effectively avoided. The technical means adopted to achieve the first purpose of the present invention include unmanned aerial vehicle, wireless communication unit and information processing unit, and unmanned aerial vehicle includes flight control module and image The capture device, the flight control module controls the flight of the drone according to the flight path; the image capture device continuously captures the flight status of the drone and forms a flight image. The information processing unit receives flight images through the wireless communication unit. The information processing unit includes a deep learning algorithm module and a built-in feature database with object feature samples. Each object feature sample is defined with an object name. The deep learning algorithm module is used for The features of the object are extracted from each flight image and input into the feature database to predict the probability of matching with the object feature sample. When the matching probability is greater than the preset probability, the identification result information of the object matching the object name is output, and it is judged whether If it is an obstacle, if the judgment result is yes, then a correction control signal for correcting the flight path is generated, so that the UAV can fly according to the correction control signal to avoid the obstacle.

The second purpose of the present invention is to provide a UAV obstacle avoidance flight control image recognition method, system and application that can change the direction of image acquisition with the application of deep learning, mainly following the direction of UAV travel. When changing the direction of image capture, since the direction of travel of the UAV and the direction of image capture are set to be synchronized, it is possible to detect immediately whether there is an obstacle in the direction of travel, thereby effectively avoiding collisions with obstacles. occur. The technical means adopted to achieve the second objective of the present invention include a drone, a wireless communication unit, and an information processing unit. The drone includes a flight control module and an image capture device. The flight control module controls the flight of the drone according to the flight path; The image capture device continuously captures the flight state of the drone and forms a flight image. The information processing unit receives flight images through the wireless communication unit. The information processing unit includes a deep learning algorithm module and a built-in feature database with object feature samples. Each object feature sample is defined with an object name. The deep learning algorithm module is used for The features of the object are extracted from each flight image and input into the feature database to predict the probability of matching with the object feature sample. When the matching probability is greater than the preset probability, the identification result information of the object matching the object name is output, and it is judged whether If it is an obstacle, if the judgment result is yes, then a correction control signal for correcting the flight path is generated, so that the UAV can fly according to the correction control signal to avoid the obstacle. It further includes an angle adjustment unit installed on the UAV to drive the image capture device to turn and a The data acquisition module installed in the UAV, the data acquisition module is used to acquire the flight status data of the flight status of the flight control module, the flight status is selected from forward flight, climbing flight, turning flight and at least one of the status data for the descent flight.

The third object of the present invention is to provide a UAV obstacle avoidance flight control image recognition method, system and application thereof that combine image recognition and P3P computing technology to realize the landing positioning function and apply deep learning. The technical means adopted to achieve the third objective of the present invention include a drone, a wireless communication unit, and an information processing unit. The drone includes a flight control module and an image capture device. The flight control module controls the flight of the drone according to the flight path; The image capture device continuously captures the flight state of the drone and forms a flight image. The information processing unit receives flight images through the wireless communication unit. The information processing unit includes a deep learning algorithm module and a built-in feature database with object feature samples. Each object feature sample is defined with an object name. The deep learning algorithm module is used for The features of the object are extracted from each flight image and input into the feature database to predict the probability of matching with the object feature sample. When the matching probability is greater than the preset probability, the identification result information of the object matching the object name is output, and it is judged whether If it is an obstacle, if the judgment result is yes, then a correction control signal for correcting the flight path is generated, so that the UAV can fly according to the correction control signal to avoid the obstacle. Wherein, the flight state data is transmitted to the information processing unit through the wireless communication unit, and when the information processing unit interprets it as a descending flight, it is determined whether the object is a landing platform, and when the judgment result is yes, it is determined through the wireless communication unit Transmit an activation signal to the flight control module to activate the image capture device to capture a pattern mark (such as ARUCO MARKER) on the landing platform and image it as a pattern mark image, and the information processing unit passes through the wireless communication unit Receive the image of the pattern mark, and use a recognition technology to perform recognition processing and calculation on the image of the pattern mark to obtain the corresponding identification information and the distance value between the UAV and the landing platform, and check whether the identification information is correct, when When the identification information is correct, it is judged whether the distance value is zero or close to zero. If the judgment result is yes, the UAV can land on the landing platform; The control signal is sent to the flight control module to make the UAV continue to descend and fly until the distance value is equal to zero or close to zero. The recognition technology is an image recognition technology, and the information processing unit includes an image recognition module and a pattern database, when the image recognition module receives the pattern mark image, it extracts the image features of the pattern mark image, and performs feature matching image recognition processing with the feature samples stored in the image database, so as to obtain the Recognition information and the distance value. The image recognition module is based on the P3P (Perspective 3 point) method. Point P is set as the projection center of the image capture device, and points A, B, and C are three-dimensional world points under the world coordinate system. After feature matching, the three points u, v, and w of the two-dimensional pattern mark image in the corresponding image coordinate system are obtained, and then the three points ABC in the world coordinate system and the three points u, v, and w in the image coordinate system are obtained. Point matching, where the lengths of AB, BC, and AC are known, uv, vw, and uw are also known, and the length and distance between PA, PB, and PC are solved by constructing simultaneous equations through the law of cosines, so as to obtain the world coordinate system The following is the camera pose of the distance value; let the length distance from the projection center (P) to each point be x =| PA |, y = | PB |, z = | PC |, α = ∠ BPC , β = ∠ APC , γ = ∠ APB , p = 2cos α , q = 2cos β , r = 2cos γ , a '=| BC |, b '=| AC |, c '=| AB |.

10: Drone

11: Flight control module

12: Image capture device

13: Angle adjustment unit

130: pivot seat

131: with pivot seat

132: Rotary drive mechanism

14: Flight drive unit

15: Power supply unit

20: Wireless communication unit

30: Information processing unit

31: Deep Learning Calculation Module

310:Deep Learning Models

311:Image recognition module

312: Pattern database

32: Feature database

40: Landing platform

41:Pattern mark

A: Obstacles

D1: Direction ahead

D2: Downward direction

D3: Direction above

Original flight path: dn

Corrected flight path: dn-1

FIG. 1 is a functional block schematic diagram of the basic implementation of the present invention.

Fig. 2 is a functional block schematic diagram of the embodiment of the present invention.

Fig. 3 is a schematic diagram of implementing the correction of the flight path when the UAV encounters an obstacle in the present invention.

FIG. 4 is a schematic diagram of the process control implementation of the deep learning calculation module in the training phase of the present invention.

FIG. 5 is a schematic diagram of the process control implementation of the deep learning calculation module in the prediction stage of the present invention.

Fig. 6 is an implementation schematic diagram of the drone of the present invention preparing to land on the landing platform.

FIG. 7 is a schematic diagram of an implementation of the image capture device being driven to change the image capture direction of the present invention.

FIG. 8 is a schematic diagram of the implementation of projection using the P3P method in the present invention.

In order to allow your review committee to further understand the overall technical characteristics of the present invention and the technical means to achieve the purpose of the present invention, specific embodiments and accompanying drawings are hereby described in detail:

Please refer to FIGS. 1 and 3 . In order to achieve the first purpose of the present invention, the first embodiment includes a drone 10 , a wireless communication unit 20 and an information processing unit 30 . Each drone 10 includes a flight control module 11 and an image capture device 12 . The flight control module 11 controls and drives the flight of the UAV 10 according to a flight path. The image capture device 12 is used to continuously capture the flight state of the UAV 10 and image it into a plurality of flight images. The information processing unit 30 (such as a computer or server located at a ground station) receives a plurality of flight images through the wireless communication unit 20 (such as a 4G, 5G mobile communication system; or a UHF, VHF radio frequency communication system). The information processing unit 30 includes a deep learning algorithm module 31 and a built-in feature database 32 with multiple object feature samples, each object feature sample defines an object name, and the deep learning algorithm module 31 is used for each At least one object feature is extracted from the flying image, and the at least one object feature is input into the feature database 32 in order to predict the coincidence probability of the input at least one object feature and the object feature sample. When the coincidence probability is greater than the preset probability, then Read out the corresponding object name to output the recognition result information of the object that matches the object name, and judge whether the object is an obstacle A that hinders the flight path. When the judgment result is yes, a correction control signal for correcting the flight path is generated. , the correction control signal is wirelessly transmitted to the flight control module 11 through the wireless communication unit 20, so that the UAV 10 flies according to the correction control signal and avoids the obstacle A. Wherein, as shown in FIG. 3 , dn is the original flight path, and dn-1 is the corrected flight path.

Specifically, each object name (such as other unmanned aerial vehicles 10 that break into the flight path; or aircraft; or other utility poles, wires, obstacles, towers, birds, etc. that block the flight path) defines a preset outline size, so that The distance between the UAV 10 and the object is determined according to the contour size of the image, and the flight path is corrected based on the contour size of the image.

Please refer to Figures 4 to 5, when the deep learning calculation module 31 is executed, the Contains the following steps:

(a) The step of the training stage is to establish at least one deep learning model 310, and input a huge amount of object feature samples, obstacle A recognition parameters and image recognition parameters into the at least one deep learning model 310, and the deep learning model 310 Test the correct rate of image recognition, and then judge whether the correct rate of image recognition is sufficient. If the judgment result is yes, then output and store the recognition result;

(b) The step of running the prediction stage is to input the real-time captured flight image into the deep learning model 310, and calculate the corresponding object features by the deep learning model 310, so as to predict the name of the object and whether it is an obstacle A. The identification result information of .

More specifically, the deep learning model 310 is a training prediction model based on the YOLO deep learning algorithm.

In addition, the present invention mainly proposes a training model based on You Only Look Once (YOLO) deep learning algorithm, which is used to estimate the actual distance of a specific target (such as an obstacle or a landing platform) in continuous images. In the ideal distance detection method, the correctness of geometric information is usually relied on, and the geometric characteristics will be lost when the object is rotated or the camera lens is distorted. In our proposed method, the deep learning algorithm of You Only Look Once (YOLO) is used to extract the spatial information between the targets as training data, and the training model is used to explore the relationship between the specific target shape and the actual world coordinates after being projected by the camera. Relative relationship, and restore the geometric error caused by different viewing angles of the target, and correct the actual distance of the target in the camera image. The feature of the present invention is that based on the deep learning method of YOLO, the task of instantly identifying the target in a small area can be realized, and the position of the target can also be accurately located. The research method of the present invention is based on the YOLO network framework, supplemented by the method of transfer learning to train the underwater-YOLO network to identify targets in time. The experimental results confirm that the proposed method is better than YOLO for the target small target and Overlapping targets have better recognition performance.

Please refer to Fig. 1~3 and Fig. 7, in order to achieve the second embodiment of the second purpose of the present invention, this embodiment not only includes the overall technical content of the above-mentioned first embodiment, but also includes a 10 is used to drive the angle adjustment unit 13 of the image capture device 12 to turn, and a flight drive unit 14 that can drive the UAV 10 to make a corresponding flight state by receiving multiple control instructions from the flight control module 11, and a flight drive unit 14 for supplying Power supply unit 15 for the required power supply. The flight state may be forward flight, climbing flight, turning flight; or at least one flight state of descending flight.

Specifically, the angle adjustment unit 13 as shown in FIG. 7 comprises a pivot seat 130 located at the bottom of the drone 10, a pivot seat 131 and a rotation drive mechanism 132. One end of the pivot seat 131 is provided with a pivot portion. The pivot part and the pivot part of the pivot base 130 form a rotatable pivot connection, and the end of the pivot base 131 can be used to fix the image capture device 12, and the rotation drive mechanism 132 is located on one side of the pivot base 130; When flying forward, the flight control module 11 triggers the rotary drive mechanism 132 to drive the lens of the image capture device 12 to rotate and locate in the forward direction D1 to capture the flying image in front of the UAV 10; When the flight state is climbing flight, the flight control module 11 triggers the rotary drive mechanism 132 to drive the lens of the image capture device 12 to rotate and position in the upward direction D3 to capture the flight image above the UAV 10; When the flight state is descending flight, the flight control module 11 triggers the rotation drive mechanism 132 to drive the lens of the image capture device 12 to rotate and position in the downward direction D2 to capture the flight image below the UAV 10 .

Please refer to Figures 1 to 3 and Figure 6, in order to achieve the third embodiment of the third purpose of the present invention, this embodiment includes the overall technical content of the above-mentioned first and second embodiments, the flight control module The control command of the group is transmitted to the information processing unit 30 through the wireless communication unit 20. When the information processing unit 30 interprets it as a landing flight, the deep learning calculation module 31 determines whether the object is a landing platform 40 after image recognition. If the result is yes, then The start signal is transmitted to the flight control module 11 through the wireless communication unit 20, so as to start the image capture device 12 to capture the image mark 41 (such as ARUCO MARKER; or QR CODE) on the landing platform 40 and form an image of the image mark 41 The information processing unit 30 receives the image of the pattern mark 41 through the wireless communication unit 20, and the deep learning calculation module 31 uses a recognition technology to perform recognition processing and calculation on the image of the pattern mark 41 to obtain corresponding recognition information (in each The UAV presets an identification code as the authority evaluation basis for whether the UAV can land on the landing platform) and the distance value between the UAV 10 and the landing platform 40, and checks whether the identification information is correct. When the identification information is correct, Then judge whether the distance value is zero or close to zero, if the judgment result is yes, the UAV 10 can land on the landing platform 40; Module 11, so that the UAV 10 continues to descend and fly until the distance value is equal to zero or close to zero. Wherein, the recognition technology is an image recognition technology, and the deep learning calculation module 31 further includes an image recognition module 311 and a pattern database 312. When the image recognition module 33 receives the image of the pattern mark 41, it retrieves The pattern mark 41 is an image feature of the image, and performs feature matching image recognition processing with the feature samples stored in the image database 34 to obtain identification information and distance values.

As mentioned above, as shown in FIG. 8 , the image recognition module is based on the P3P (Perspective 3 point) method, and point P is set as the projection center of the image capture device 12, and points A, B, and C are points in the world coordinate system. After the feature matching, the three points u, v, and w of the two-dimensional pattern mark 41 image in the corresponding image coordinate system are obtained, and then the ABC three points in the world coordinate system are compared with the u in the image coordinate system , v, and w are matched at three points, where the lengths of AB, BC, and AC are known, and uv, vw, and uw are also known. Through the cosine theorem, the simultaneous equations are constructed to solve the length distance between PA, PB, and PC, and then Obtain the camera pose as the distance value in the world coordinate system; let the length and distance from the projection center (P) to each point be x = | PA |, y = | PB |, z = | PC |, α = ∠BPC , β = ∠APC , γ = ∠APB , p =2cos α , q =2cos β , r =2cos γ , a '=| BC |, b '=| AC |, c '=| AB |, According to the law of cosines, the simultaneous equations can be obtained as follows:

The present invention is applied to a spraying glue supply system of an UAV obstacle avoidance flight control image recognition system using deep learning, which includes a glue supply device (not shown in this figure) and a spray control module arranged on the UAV 10 Group (not shown in this figure example), the glue supply device includes a glue supply cabin (not shown in this figure example) for holding the gelatinous plant growth substrate and a spraying machine for spraying the gelatinous plant growth substrate Nozzle (not shown in this drawing example); the flight control module 11 of the UAV 10 is preset with a glue supply flight path for flying to a glue-spraying area, so that the UAV 10 can follow the glue supply flight path And fly to the sky above the area to be sprayed; the spray control module is used to control the glue supply device to actuate and make the nozzle spray the timing and spraying amount of the gel-like plant substrate toward the area to be sprayed, so that the surface of the area to be sprayed is coated At least one layer of gelatinous vegetation substrate.

Through the description of the above specific embodiments, the present invention does have the following characteristics:

1. The present invention can indeed realize the task of instantly identifying obstacles in a small area of the flight path, and can accurately locate the actual position of the obstacle, thereby effectively avoiding the obstacle.

2. The present invention can indeed change the direction of image capture along with the direction of travel of the UAV. Since the direction of travel of the UAV and the direction of image capture are set to be synchronized, it can be detected in real time whether there is any difference in the direction of travel. obstacles, thus effectively avoiding collisions with obstacles.

3. The present invention does have the function of combining image recognition and P3P computing technology to realize landing positioning.

The above descriptions are only feasible embodiments of the present invention, and are not intended to limit the present invention All equivalent implementations of other changes based on the content, features and spirit of the following claims shall be included in the patent scope of the present invention. The structural features of the invention specifically defined in the claims are not found in similar items, and are practical and progressive, and have met the requirements of an invention patent. I file an application in accordance with the law. I would like to ask the Jun Bureau to approve the patent in accordance with the law to maintain this invention. The legitimate rights and interests of the applicant.

10: Drone

11: Flight control module

12: Image capture device

14: Flight drive unit

15: Power supply unit

20: Wireless communication unit

30: Information processing unit

31: Deep Learning Calculation Module

32: Feature database

Claims (9)

A UAV obstacle avoidance flight control image recognition method using deep learning, which includes: providing at least one UAV, a wireless communication unit and an information processing unit; wherein, each UAV includes a flight control module and an image Capture device; make the flight control module drive the at least one drone to fly according to a flight path control; use the image capture device to continuously capture the flight status of the drone and image it into a plurality of flight images; and make the information The processing unit receives the plurality of flight images through the wireless communication unit, the information processing unit includes a deep learning algorithm module and a built-in feature database with a plurality of object feature samples, each of the object feature samples defines an object name, The deep learning algorithm module is used to extract at least one object feature for each of the flying images, and input the at least one object feature into the feature database in order to predict the input of the at least one object feature and the object feature The matching probability of the sample, when the matching probability is greater than a preset probability, read out the corresponding object name to output the identification result information of the object matching the object name, and judge whether the object is obstructing the flight path If the judgment result is yes, a correction control signal for correcting the flight path will be generated, and the correction control signal will be wirelessly transmitted to the flight control module through the wireless communication unit, so that the UAV will fly according to the correction control signal. Avoiding the obstacle; wherein, the control command of the flight control module is transmitted to the information processing unit through the wireless communication unit, and when the information processing unit interprets the control command as descending flight, the deep learning calculation module passes After image recognition, it is determined whether the object is a landing platform, and when the judgment result is yes, an activation signal is transmitted to the flight control module through the wireless communication unit to activate the image capture device to capture images on the landing platform. A pattern mark is imaged as a pattern mark image, the information processing unit receives the pattern mark image through the wireless communication unit, and the deep learning calculation module uses a recognition technology to recognize, process and calculate the pattern mark image to obtain a corresponding identification information and the distance value between the UAV and the landing platform, and check whether the identification information is correct, when the When the identification information is correct, it is judged whether the distance value is zero or close to zero. If the judgment result is yes, the UAV can land on the landing platform; The signal is sent to the flight control module, so that the UAV continues to descend and fly until the distance value is equal to zero or close to zero. The UAV obstacle avoidance flight control image recognition method applying deep learning as described in claim 1, wherein the deep learning calculation module includes the following steps when executed: (a) the training phase step is to establish at least one deep learning model, and input a huge number of object feature samples, obstacle recognition parameters, and image recognition parameters into the at least one deep learning model, and test the accuracy of image recognition by the deep learning model, and then determine whether the accuracy of image recognition is sufficient, When the judgment result is yes, the identification result is output and stored; when the judgment result is no, the deep learning model is made to self-correct and study; and (b) the step of running the prediction stage is obtained in real time as the input of the deep learning model The flight image, and the corresponding feature of the object is calculated by the deep learning model, so as to predict the identification result information of the object name and whether the object is the obstacle. The UAV obstacle avoidance flight control image recognition method applying deep learning as described in claim 2, wherein the at least one deep learning model is a training prediction model based on the YOLO deep learning algorithm. An image recognition system for UAV obstacle avoidance flight control using deep learning, which includes: at least one UAV, each UAV includes a flight control module and an image capture device, and the flight control module is based on a flight Path control drives the flight of the at least one unmanned aerial vehicle; the image capture device is used to continuously capture the flight state of the unmanned aerial vehicle and image it into a plurality of flight images; a wireless communication unit; and an information processing unit, which through the wireless communication The unit receives the plurality of flight images; the information processing unit includes a deep learning algorithm module and a built-in characteristic data of the plurality of object characteristic samples a database, each of the object feature samples defines an object name, and the deep learning algorithm module is used to extract at least one object feature for each of the flight images, and input the at least one object feature into the feature database in sequence, To predict the coincidence probability of the at least one input object characteristic and the object characteristic sample, when the coincidence probability is greater than a preset probability, then read out the corresponding object name, so as to output the object name matching the object name Identify the result information and judge whether the object is an obstacle blocking the flight path. If the judgment result is yes, then generate a correction control signal to correct the flight path, and the correction control signal is wirelessly transmitted to the flight control through the wireless communication unit module, so that the UAV can fly according to the corrected control signal to avoid the obstacle; wherein, the control command of the flight control module is transmitted to the information processing unit through the wireless communication unit, and the information processing unit interprets the When the control command is to descend and fly, the deep learning algorithm module judges whether the object is a landing platform after image recognition, and when the judgment result is yes, then transmits an activation signal to the flight control module through the wireless communication unit to Start the image capture device to capture a pattern mark on the landing platform and image it as a pattern mark image, the information processing unit receives the pattern mark image through the wireless communication unit, and the deep learning calculation module uses a recognition The technology performs recognition processing and calculation on the image of the pattern mark to obtain the corresponding recognition information and the distance value between the UAV and the landing platform, and check whether the recognition information is correct. When the recognition information is correct, then judge the distance If the value is zero or close to zero, if the judgment result is yes, the UAV can land on the landing platform; Let the UAV continue to descend until the distance value is equal to zero or close to zero. The UAV obstacle avoidance flight control image recognition system applying deep learning as described in claim 4, which further includes an angle adjustment unit installed on the UAV to drive the image capture device to turn and a device that can be controlled by the flight Multiple control commands of the module drive the flight drive unit of the UAV to make a corresponding flight state. The flight state is selected from one of forward flight, climbing flight, turning flight and descending flight. The UAV obstacle avoidance flight control image recognition system applying deep learning as described in claim 5, wherein the angle adjustment unit includes a pivot seat arranged at the bottom of the UAV, a supporting pivot seat and a rotating drive mechanism, the fitting One end of the pivot base is provided with a pivot part, which forms a rotatable pivot connection with one of the pivot parts of the pivot base. The end of the pivot base can be used to fix the image capture device. One side of the pivot seat; when the flight state is forward flight, the flight control module triggers the rotary drive mechanism to drive a lens of the image capture device to rotate and position it towards the front to capture the unmanned The flight image in front of the aircraft; when the flight state is climbing flight, the flight control module triggers the rotary drive mechanism to drive the lens of the image capture device to rotate and position upward to capture the The flight image above the UAV; when the flight state is descending flight, the flight control module triggers the rotary drive mechanism to drive the lens of the image capture device to rotate and position downward to capture This flight image below the drone. The UAV obstacle avoidance flight control image recognition system applying deep learning as described in claim 4, wherein the recognition technology is an image recognition technology, and the deep learning calculation module further includes an image recognition module and a pattern database , when the image recognition module receives the pattern mark image, it extracts the image features of the pattern mark image, and performs feature matching image recognition processing with the feature samples stored in the image database to obtain the recognition information and the distance value. The UAV obstacle avoidance flight control image recognition system applying deep learning as described in claim 7, wherein the image recognition module is based on the P3P (Perspective 3 point) method, setting point P as the projection center of the image capture device, Points A, B, and C are three-dimensional world points in the world coordinate system. After feature matching, three points u, v, and w of the two-dimensional pattern mark image in the corresponding image coordinate system are obtained, and then the points in the world coordinate system are The three points ABC match the three points u, v, and w in the image coordinate system, where the lengths of AB, BC, and AC are known, and uv, vw, and uw are also known, and PA is solved by constructing a simultaneous equation through the law of cosines. PB, the length distance between PC to obtain the camera pose as the distance value in the world coordinate system; let the length distance from the projection center (P) to each point be x = | PA |, y = | PB |, z = | PC |, α = ∠ BPC , β = ∠ APC , γ = ∠ APB , p = 2cosα, q = 2cosβ, r = 2cosγ, a '=| BC |, b '=| AC |, c '=| AB |, according to the law of cosines, the simultaneous equation can be obtained as follows:

A spraying glue supply system that applies the UAV obstacle avoidance flight control image recognition system using deep learning such as claim 4, which includes a glue supply device, a nozzle and a spray control module arranged on the UAV. The glue supply device includes a glue supply cabin for accommodating the gelatinous planting substrate and a nozzle for spraying the gelatinous planting substrate; the flight control module of the drone is preset with one for flying to A glue supply flight path over the area to be sprayed allows the UAV to fly to the sky above the area to be sprayed according to the glue supply flight path; the spray control module is used to control the glue supply device to operate, so that The nozzle sprays the timing and spraying amount of the gel-like vegetation substrate toward the area to be sprayed, and makes the surface of the area to be sprayed coated with at least one layer of the gel-like vegetation substrate.

Images