WO2023173330A1

WO2023173330A1 - Flight control method and apparatus for unmanned aerial vehicle, unmanned aerial vehicle, and storage medium

Info

Publication number: WO2023173330A1
Application number: PCT/CN2022/081230
Authority: WO
Inventors: 颜江; 李罗川; 高成强
Original assignee: 深圳市大疆创新科技有限公司
Priority date: 2022-03-16
Filing date: 2022-03-16
Publication date: 2023-09-21

Abstract

A flight control method and apparatus for an unmanned aerial vehicle, an unmanned aerial vehicle, and a storage medium. The method comprises: obtaining description information of a target object in the space; obtaining a target flight direction of an unmanned aerial vehicle; detecting, on the basis of a sensor carried by the unmanned aerial vehicle, description information of a detected object in the space indicated by the target flight direction; if the description information of the detected object matches the description information of the target object, controlling the unmanned aerial vehicle to perform a decelerating motion along the target flight direction to approach the detected object; and if the description information of the detected object does not match the description information of the target object, controlling the unmanned aerial vehicle to change the flight direction to bypass the detected object. Thus, whether the description information of the detected object matches the description information of the target object is automatically detected, and the unmanned aerial vehicle is automatically controlled, on the basis of a detection result, to avoid or approach the detected object, so that to a certain extent, the control efficiency can be improved and the labor cost can be reduced.

Description

UAV flight control method, device, UAV and storage medium

Technical field

The present application relates to the technical field of drones, and in particular to a flight control method and device for a drone, a drone and a storage medium.

Background technique

At present, with the continuous development of drone technology, drones are being used more and more. In the process of using drones, how to control drones to fly safely has become an issue of increasing concern.

In the prior art, when controlling the flight of a drone, the user often manually controls the drone to fly based on his or her own judgment of the surrounding environment to ensure flight safety. In this way, efficiency is lower and labor costs are higher.

Contents of the invention

This application provides a UAV flight control method, device, UAV and storage medium, which can improve control efficiency and reduce labor costs to a certain extent.

In a first aspect, embodiments of the present application provide a flight control method for a UAV, which method includes:

Obtain description information of target objects in space;

Obtain the target flight direction of the drone;

Based on the sensor mounted on the drone, detect the description information of the detected object in the space indicated by the target flight direction;

If the description information of the detected object matches the description information of the target object, control the drone to decelerate along the target flight direction to approach the detected object;

If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object.

In a second aspect, embodiments of the present application provide a flight control device for a drone, the device including a memory and a processor;

The memory is used to store program code;

The processor calls the program code, and when the program code is executed, is used to perform the following operations:

Obtain description information of target objects in space;

Obtain the target flight direction of the drone;

In a third aspect, embodiments of the present application provide a drone, which is used to implement the above method.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, including instructions that, when run on a computer, cause the computer to perform the above method.

In a fifth aspect, embodiments of the present application provide a computer program product containing instructions, which when the instructions are run on a computer, cause the computer to perform the above method.

In the embodiment of the present application, the description information of the target object in the space is obtained, and the target flight direction of the UAV is obtained. Based on the sensor mounted on the drone, the description information of the detected object in the space indicated by the target flight direction is detected. If the description information of the detected object matches the description information of the target object, the drone is controlled to decelerate along the target flight direction until it approaches the detected object. If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object. In this way, by automatically detecting whether the description information of the detected object matches the description information of the target object, and automatically controlling the drone to avoid the detected object or approach the detected object based on the detection results, control efficiency can be improved to a certain extent and labor costs can be reduced. . At the same time, it can avoid collisions between the drone and the detected object to ensure the safety of the drone's flight, while ensuring that the drone can get close to the target object in space, thus taking into account flight safety and scene requirements.

Moreover, the manual control of drone flight is limited by personal experience, and sometimes it is impossible to reasonably control the movement of the drone. In the embodiment of this application, the drone is automatically controlled to avoid the detected object or approach the detected object based on the detection results. To a certain extent, the movement of the drone can be more reasonably controlled for the detected object in space.

Description of the drawings

Figure 1 is a step flow chart of a UAV flight control method provided by an embodiment of the present application;

Figure 2 is a schematic diagram of speed change provided by the embodiment of the present application;

Figure 3 is a schematic diagram of a query provided by an embodiment of the present application;

Figure 4 is a schematic diagram of a function curve provided by an embodiment of the present application;

Figure 5 is a schematic diagram of a collision query channel provided by an embodiment of the present application;

Figure 6 is a schematic diagram of a preset range provided by an embodiment of the present application;

Figure 7 is a system flow diagram involved in a flight control method provided by an embodiment of the present application;

Figure 8 is a block diagram of a flight control device for a drone provided by an embodiment of the present application;

Figure 9 is a block diagram of a computing processing device provided by an embodiment of the present application;

Figure 10 is a block diagram of a portable or fixed storage unit provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

First, an exemplary application scenario involved in the embodiment of the present application is described. In this application scenario, the drone needs to inspect specific objects in space to perform specified operations. Correspondingly, during flight, the drone needs to be close to specific objects in space to complete specified operations. There are often many objects distributed in space, and sometimes the distribution of objects is more complex. In order to ensure the flight safety of the drone, the drone needs to avoid objects in the space. This leads to the need to take into account safety in flight control while ensuring that the drone can get close to designated objects.

To this end, embodiments of the present application provide a flight control method for a drone.

Figure 1 is a step flow chart of a UAV flight control method provided by an embodiment of the present application. As shown in Figure 1, the method may include:

Step 101: Obtain description information of the target object in the space.

Step 102: Obtain the target flight direction of the drone.

In the embodiment of the present application, the space may refer to the space where the drone is located, the target object may be specified in advance, and the target object may be a designated object in the above application scenario. The description information of the target object may be information that can be used to indicate the target object, such as image information, position distribution information, etc. of the target object. By obtaining the description information of the target object, the drone can accurately learn the objects that need to be approached during this flight. Further, the target flight direction may refer to the current flight direction of the drone. For example, the current flight direction of the drone can be detected based on the direction sensor mounted on the drone, thereby obtaining the target flight direction. Of course, other methods can also be used to obtain the target flight direction, and this application does not limit this.

Step 103: Based on the sensor mounted on the drone, detect the description information of the detected object in the space indicated by the target flight direction.

In the embodiment of the present application, the space indicated by the target flight direction may be a space within a preset range in the target flight direction. Among them, the preset range can be set according to actual needs. The description information of the detected object may be information that can be used to indicate the detected object, such as image information, location information, etc. of the detected object. The detected object may be an object falling into the space indicated by the target's flight direction. For example, the space indicated by the target flight direction can be detected based on sensors capable of detecting objects such as infrared sensors and ultrasonic sensors mounted on the drone to detect whether objects appear in the space indicated by the target flight direction. For objects that appear in the space indicated by the target flight direction, the object can be determined as a detected object. Further, the image of the detected object can be collected based on the mounted visual sensor as description information of the detected object. Alternatively, the position information of the detected object can be collected as description information of the detected object.

Step 104: If the description information of the detected object matches the description information of the target object, control the drone to decelerate along the target flight direction to approach the detected object.

Step 105: If the description information of the detected object does not match the description information of the target object, control the drone to change its flight direction to bypass the detected object.

Among them, the description information of the detected object matches the description information of the target object, which can be understood to mean that the currently detected object is the target object. If the description information of the detected object does not match the description information of the target object, it can be understood that the currently detected object being detected is not the target object. Specifically, taking the description information as image information as an example, after obtaining the image information of the target object, it can be combined with the collected image information of the detected object based on the Convolutional Neural Networks (CNN) carried by the drone. The target detection and tracking module detects whether the detected object is a target. During specific implementation, the position information of the detected object can be further determined, and the position information of the detected object can be associated with the object position distribution information in space, so that the drone can locate the detected object based on the associated object position distribution information. , thereby approaching or bypassing the detected object. Among them, the object position distribution information can be a map of the current space, and the position information of the detected object can be the position of the detected object in the three-dimensional navigation coordinate system.

If the currently detected object is a target, it means that there is a target in the current flight direction. Therefore, there is no need to change the flight direction. The drone can be controlled to decelerate along the target flight direction until it is close to the detected object to facilitate detection. Operate on the target. At the same time, slowing down the movement until it is close to the detected object can avoid excessive speed to a certain extent, thereby ensuring the stability of the drone during the approach. If the currently detected object is not a target, the drone can be directly controlled to change its flight direction to bypass the detected object, thus ensuring flight safety while avoiding unnecessary deceleration and approach operations. It should be noted that the flight control method provided by the embodiment of the present application can be applied to a drone, that is, the drone automatically controls itself to approach or bypass the detected object. Alternatively, it can also be applied to the control device of a drone, that is, the control device automatically controls the controlled drone to approach or bypass the detected object. The embodiments of this application do not limit this.

To sum up, the flight control method of a UAV provided by the embodiment of the present application obtains the description information of the target object in space and obtains the target flight direction of the UAV. Based on the sensor mounted on the drone, the description information of the detected object in the space indicated by the target flight direction is detected. If the description information of the detected object matches the description information of the target object, the drone is controlled to decelerate along the target flight direction until it approaches the detected object. If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object. In this way, by automatically detecting whether the description information of the detected object matches the description information of the target object, and automatically controlling the drone to avoid the detected object or approach the detected object based on the detection results, control efficiency can be improved to a certain extent and labor costs can be reduced. . At the same time, it can avoid collisions between the drone and the detected object to ensure the safety of the drone's flight, while ensuring that the drone can get close to the target object in space, thus taking into account flight safety and scene requirements.

Optionally, in this embodiment of the present application, when the drone is controlled to decelerate along the target flight direction until it is close to the detected object, that is, when the currently detected detected object is the target object, the following operations may be further performed. : Control the drone to stop flying when the distance between the drone and the detected object is not greater than the specified stopping distance. That is to say, in the process of controlling the UAV to decelerate along the target flight direction to approach the detected object, the current distance between the UAV and the detected object can be continuously detected, and the current distance and the specified stopping distance can be compared. Comparison. If the current distance is equal to the specified stopping distance, the drone can be controlled to stop flying, thereby controlling the drone to decelerate along the target flight direction to the specified stopping distance from the detected object. In this way, there is no need to manually adjust the distance between the drone and the target. By automatically detecting the distance and stopping the flight at the specified stopping distance, it is possible to achieve automatic and precise control of the distance to the target to a certain extent. Further save labor costs. Among them, the designated stopping distance may be pre-specified according to actual needs, and controlling the drone to stop flying may control the flying speed of the drone to reduce to 0 at the designated stopping distance from the detected object. In the embodiment of the present application, by controlling the deceleration of the drone to a designated stopping distance from the detected object, it is possible to ensure that the drone can operate on the target object while preventing the drone from getting too close to the target object. , thereby ensuring the safety of drone operations.

Furthermore, in the embodiment of the present application, after the drone stops flying, the drone can be controlled to perform a target operation on the detected object, thereby meeting the operation requirements for the target object. Among them, the target operation can be specified in advance according to actual needs. For example, performing the target operation may be to observe the target object to determine whether the target object is in a preset state, and transmit the observation results back to the user equipment of the monitoring personnel. For example, in a wind power generation scenario, it is possible to observe whether the wind power generation equipment is in normal operation, and the observation results indicating whether the wind power generation equipment is in normal operation are sent back to the monitoring personnel's equipment, thus facilitating the monitoring personnel when the wind power generation equipment is not in normal operation. If the equipment is in good condition, perform maintenance on the equipment in a timely manner. Alternatively, the target operation can also be to collect images of the target object, spray maintenance fluid on the target object, and so on.

In one implementation, the designated stopping distance can be set based on one or more of the following factors: the density of objects in the space; the description information of the target object; and the operation type of the drone. Among them, the specified stopping distance can be preset before the UAV performs this mission, or set according to these factors during the flight. In the embodiment of this application, the designated stopping distance is set based on factors such as the density of objects in the space, the description information of the target object, and the operation type of the drone, which can make the set designated stopping distance more suitable for this flight scenario to a certain extent. , thereby ensuring the reasonableness of the specified stopping distance set. In the actual setting, the user can be provided with a selectable range, and the value selected by the user from the selectable range is set to the specified stopping distance. Among them, the end value of the selectable range can be automatically adjusted based on one or more factors including the density of objects in the space, description information of the target object, and the type of operation of the drone. Specifically, the greater the density of the object, that is, the denser the object in the space, the more complex the shape of the target represented by the description information of the target, and the higher the difficulty of the drone's operation type, the range of options can be The larger the end value can be, the larger the designated stopping distance set by the user can be, thereby ensuring the safety of the UAV when performing target operations on detected objects detected in space. It should be noted that the minimum value of the selectable range limit can be no less than the preset safe distance. In this way, the problem of the set designated stopping distance being too small and the UAV being too close to the target can be avoided.

Optionally, the above-mentioned operation of controlling the drone to decelerate along the target flight direction to approach the detected object may specifically include the following steps:

Step S21: Determine the target safe speed based on local map information and/or single-frame depth image; the target safe speed is positively related to the current distance between the drone and the detected object.

Among them, the local map information can be used to characterize the local map of the surrounding environment of the drone, and the depth image can be a depth image of the current surrounding environment collected by the drone. The depth image can be obtained from the image collector of the drone to the surrounding environment. An image of the depth of each point in the environment as pixel values.

Step S22: Control the drone to move along the target flight direction and decelerate at the target safe speed until it approaches the detected object.

In the embodiment of this application, the target safety speed is positively related to the current distance between the UAV and the detected object. That is to say, when the current distance between the UAV and the detected object is smaller, the determined target The safe speed can be smaller. As the drone continues to get closer to the detected object, the target safety speed can be continuously updated and become smaller and smaller. In this way, in the process of controlling the UAV to decelerate at the target safe speed until it approaches the detected object, it can gradually decelerate and approach the detected object, ensuring that the speed changes smoothly during the process of approaching the detected object and ensuring the flight safety of the UAV.

Optionally, in one implementation, the above-mentioned operation of determining the target safe speed based on local map information and/or single-frame depth images may specifically include:

Step S31: Calculate the first safe speed of the drone based on the local map information.

In one implementation, the local map information may be generated based on multi-frame depth images of the current surrounding environment collected by the drone. In this way, local map information is generated based on multi-frame depth images of the current surrounding environment, so that the local map information can more accurately represent the local map of the current surrounding environment of the drone, and to a certain extent, the determined safe speed can be more suitable for the current surrounding environment. , thereby improving the speed planning effect.

Step S32: Calculate the second safe speed of the drone based on the single-frame depth image.

Step S33: Determine the target safety speed based on the first safety speed and the second safety speed.

In this step, the target safe speed may be positively correlated with the first safe speed and the second safe speed. For example, the average value of the first safe speed and the second safe speed may be used as the target safe speed. Alternatively, the minimum value of the first safety speed and the second safety speed may be determined as the target safety speed. In this way, by selecting the minimum value as the target safe speed, the problem of excessive target safe speed can be avoided as much as possible while setting a reasonable safe speed.

In the embodiment of the present application, speed planning is performed based on local map information and a single-frame depth image, that is, the first safe speed and the second safe speed are determined based on the local map information and the single-frame depth image, and the first safe speed and the second safe speed are determined based on the local map information and the single-frame depth image. 2. Safe speed determines the target safe speed. Since the local map has strong anti-noise ability and has the ability to "memory" the observed area, it has a certain degree of blind spot perception. The single-frame depth image can represent the distance while making up for the shortcomings of the small range of the local map. Therefore, speed planning based on local map information and single-frame depth images can make the determined target safety speed more reasonable to a certain extent. It should be noted that speed planning can also be performed based only on local map information, or speed planning can be performed only on a single frame of depth image. In this way, the calculation amount of speed planning can be reduced to a certain extent.

And compared with the direct braking method, in the embodiment of the present application, when automatically switching to the braking mode, that is, when controlling the drone to stop at a designated stopping distance from the detected object, by combining the local map information And single-frame depth images are used for smooth speed planning, which to a certain extent can enable the drone to smoothly decelerate to a position close to the detected object, thereby accurately controlling the distance between the drone and the detected object after it stops.

Further, the above-mentioned local map information may be specifically generated based on the multi-frame depth image when the distance between the drone and the detected object reaches a preset mapping distance, and the preset Mapping distance is directly related to the specified stopping distance of the drone. Specifically, in the process of controlling the UAV to decelerate along the target flight direction to approach the detected object, the current distance between the UAV and the detected object can be continuously detected, and the current distance and the preset mapping distance can be continuously detected. Make a comparison. If the current distance is equal to the preset mapping distance, local map information can be generated based on the collected multi-frame depth images. For example, fusion can be performed based on the collected multi-frame depth images to generate local map information. Among them, the preset mapping distance can be greater than the specified stopping distance. The set preset mapping distance can be positively related to the specified stopping distance, thereby ensuring that there is sufficient distance between the position where the drone starts mapping and the specified stopping distance from the detected object, and to avoid being at the specified stopping distance from the detected object. When the vehicle is at a relatively close location, it starts to perform speed planning based on the established local map information, resulting in a problem of poor speed planning effect.

It should be noted that in the embodiment of the present application, local map information can be used for short-range fine-grained speed planning, and a single frame depth image can be used for long-range coarse-grained speed planning. When the preset mapping distance is not reached, the default value can be used as the first safe speed. Further, after the local map information is obtained through mapping, the first safe speed is determined based on the local map information. Because the first safe speed determined based on local map information is often smaller than the second safe speed. Therefore, after reaching the preset mapping distance and establishing local map information, the drone will be controlled to fly at a smaller first safe speed as the target safe speed. In other words, after reaching the preset mapping distance, the drone's speed will drop suddenly. Therefore, in the embodiment of the present application, by controlling the positive correlation between the preset mapping distance and the specified stopping distance, it is also possible to avoid a sudden drop in speed at a position closer to the specified stopping distance of the target, thereby enabling more precise control of the unmanned vehicle. The stop position of the machine. Exemplarily, Figure 2 is a schematic diagram of speed changes provided by an embodiment of the present application. As shown in Figure 2, the horizontal axis represents the distance to the detected object, and the vertical axis represents the speed value. Line 1 represents the first safe speed, line 2 represents the second safe speed, and line 3 represents the flight direction. It can be seen that when the preset mapping distance is not reached, the first safe speed is the default value and is greater than the second safe speed. When the preset mapping distance is reached, the first safe speed determined based on the local map information is smaller than the second safe speed. It should be noted that after the drone performs the target operation, it can continue to fly away from the currently detected object. At this time, as the distance becomes farther and farther, the speed can become higher and higher. And because the distance between the drone and the detected object will not meet the preset mapping distance after moving away, the first safe speed can be returned to the default value.

Optionally, the multi-frame depth images used to establish local map information are obtained by sensors mounted on the drone in different directions. In this way, multiple frames of depth images acquired in different directions are used to establish local map information, so that the established local map information can represent the map information in each direction, so that the drone can omnidirectionally perceive and detect objects based on the local map information. the distance between. For example, a visual sensor may be provided in the first direction of the drone. When collecting multiple frames of depth images of the current surrounding environment, the depth image in the first direction of the current surrounding environment may be collected based on the visual sensor. Control the drone to flip in the second direction to obtain a depth image in a direction where no visual sensor is provided. Wherein, the second direction is different from the first direction, so that a drone with non-omnidirectional sensing can also obtain multi-frame depth images in different directions, thereby omnidirectionally sensing the detected object. For example, the drone can only be equipped with visual sensors in the front-to-back direction. After taking off, the drone can be controlled to rotate a certain yaw angle (Yaw) with the drone as the center to collect depth images in the second direction. Of course, the attitude of the drone will change during flight, and the drone can be controlled to collect depth images based on the visual sensor after the attitude occurs. In this way, depth images in different orientations are obtained through continuous accumulation.

Optionally, the above-mentioned operation of calculating the first safe speed of the drone based on the local map information may specifically include:

Step S41: Determine the current distance between the drone and the detected object based on the local map information.

In this step, trajectory generation can be performed based on the current flight direction and local map information to generate a query trajectory. A collision query is performed based on the query trajectory to determine the current distance between the drone and the detected object. In this way, direct point cloud manipulation can be avoided, thus saving computing resources. And the current flight direction is used to provide a priori information. In which direction the drone flies, the query is performed in that direction, and other irrelevant directions are involved in the calculation. In this way, computing resource consumption can be further reduced. Specifically, the method of generating query trajectories can be selected according to actual needs, and this application does not limit this. For example, it can be generated based on a differential flat model or a method based on convex optimization.

When performing a collision query based on the query trajectory, the current location of the drone's center point can be used as the starting query point, based on the distance between the query point represented in the local map and the nearest object and the current collision query radius. Calculate the next step of trajectory sampling. Among them, the local map used in collision query can be any environment map available for query, such as occupation map, esdf map, etc.

Determine the next query point based on the trajectory sampling step, and calculate the next trajectory sampling step based on the distance between the query point and the nearest object and the current collision query radius, until the drone reaches the distance specified by the detected object and stops. distance. Further, based on the accumulated trajectory sampling step, the current distance between the drone and the detected object is determined. For example, the sum of the trajectory sampling steps is determined as the current distance.

For example, esdf represents the distance between the current query point and the nearest object, r represents the current collision query radius, step represents the next trajectory sampling step,

In this way, the calculation steps can be made more reasonable. For example, Figure 3 is a query schematic diagram provided by an embodiment of the present application. As shown in Figure 3, if a uniform step size is used and esdf is directly used as the next trajectory sampling step, the next query point will be point 2. However, in fact the next query point should be point 1. Therefore, uneven steps are used in the embodiments of the present application, and calculating steps in the above manner can make the calculated steps more reasonable, thus avoiding the problem that the distance found is longer than the actual distance. Furthermore, the designated stopping distance from the detected object is used as the target position, and the target position is regarded as an obstacle. Due to the uneven step, the situation in front of the target position may be found in one step, causing the drone to exceed the target. Location. Among them, the part covered by the diagonal lines in Figure 3 can represent the part of the drone that exceeds the specified stopping distance. Therefore, corrections can be made based on detecting the current distance. For example, real_dist represents the corrected current distance, checked_dist represents the detected current distance, and checked _esdf represents the detected distance to the nearest obstacle. The following formula can be used for correction: real_dist=checked_dist-(r- checked _esdf ). Among them, checked_dist can be the distance from point to point 4 in Figure 3, (r-checked _esdf ) represents the distance from point 3 to point 4. In this way, real_dist can be made more reasonable through correction to prevent the drone from exceeding the target position.

Step S42: Calculate the first safe speed of the UAV based on the current distance and the designated stopping distance of the UAV; the first safe speed is positively related to the current distance difference, and the current distance difference The value is the difference between the current distance and the specified stopping distance.

Specifically, the difference between the current distance and the designated stopping distance can be used as the input of the preset function, and the output of the preset function can be used as the first safe speed. For example, map_safe _spd represents the first safe speed, safe_dist represents the specified stopping distance, map_safe _spd =f(real_dist-safe_dist).

In the embodiment of the present application, the current distance between the drone and the detected object is determined based on local map information. Based on the current distance and the specified stopping distance of the drone, calculate the first safe speed of the drone; the first safe speed is positively related to the current distance difference, and the current distance difference is the difference between the current distance and the specified stopping distance . In this way, the closer the distance is, the smaller the planned first safe speed is, thereby achieving smooth speed planning. Moreover, local map information has stronger anti-noise ability, so the rationality of first safety can be ensured to a certain extent.

Optionally, the above-mentioned operation of calculating the second safe speed of the drone based on the single-frame depth image may specifically include:

Step S51: Based on the number of pixels falling into each depth interval in the single-frame depth image and the total number of pixels in the single-frame depth image, determine the proportion of the number of pixels corresponding to each depth interval.

Among them, the single-frame depth image can be continuously collected and updated by the drone in the flight direction as it approaches the target. Specifically, based on the depth map speed planning module, the depth histogram can be calculated based on the single-frame depth image, and the depth interval to which the depth value represented by each pixel in the single-frame depth image belongs, and the depth interval into which each pixel falls can be obtained. Of course, in practical applications, single-frame depth map speed planning can also be performed by directly calculating the closest point cloud, and the embodiments of the present application do not limit this. For any depth interval, the ratio of the number of pixels falling within the depth interval to the total number of pixels in a single frame depth map can be calculated to obtain the proportion of the number of pixels corresponding to the depth interval.

Step S52: Determine the weight value corresponding to each depth interval according to the proportion of the number of pixels corresponding to each depth interval; the weight value corresponding to the depth interval is positively correlated with the proportion of the number of pixels.

Step S53: Determine the second safe speed based on the speed value and weight value corresponding to each depth interval; the speed value corresponding to the depth interval is positively correlated with the depth value represented by the depth interval.

When the drone is closer to the target, the smaller the depth value, the greater the proportion of pixels corresponding to the depth interval. Therefore, by setting the speed value corresponding to the depth interval to be positively related to the depth value represented by the depth interval, and the weight value corresponding to the depth interval to be positively related to the proportion of the number of pixels, it can be made that when the drone is closer to the target, the lower speed value The corresponding weight value is larger, which to a certain extent can make the second safe speed gradually decrease as the drone approaches the target.

The speed value corresponding to each depth interval may be preset. For example, the velocity values corresponding to the depth range represented by each depth interval can be as shown in the following table:

Among them, p0, p1, p2, p3, p4, p5, p6, p7, p8, and p9 represent the proportion of the number of pixels corresponding to each depth interval. Furthermore, the pixel number ratio can be used to represent the original weight of the depth interval, and the original weight can be adjusted to obtain the weight value corresponding to each depth interval.

Depth_safe _spd represents the second safe speed, depth_safe _spd =Pnew*V. Among them, V represents the vector corresponding to the second row in the above table, and Pnew represents the adjusted vector corresponding to the third row in the above table. That is, the sum of the products of the adjusted weight value and the speed value of each depth interval can be calculated to obtain the second safe speed. It should be noted that the adjusted sum of the weight values corresponding to each depth interval may be equal to 1, and the weight value corresponding to each depth interval may determine the proportion of the speed value corresponding to each depth interval in the final second safe speed. Therefore, the method of calculating the sum of the products of the adjusted weight value and the speed value of each depth interval can be equivalent to using the adjusted weight value to perform a weighted average.

Use pnew to represent the weight value after the depth interval adjustment. pnew can include ph, new and pl, new. Among them, the value of h can be 0 or 1, and ph can represent the weight corresponding to the high-speed interval. The value of l can be from 2 to 9, and pl can represent the weight corresponding to the low-speed interval. The adjusted weight value can be calculated based on the following formula:

Ph, new=ph, new*Ph/ph, Pl, new=pl, new*Pl/pl; where P represents a vector and p represents a numerical value. That is, for all high-speed sections, the ratio of the adjusted weight value to the original weight value can be calculated to obtain the change multiple, and the product of the change multiple and the original weight value of each high-speed section can be calculated to obtain the weight value corresponding to each high-speed section. For all low-speed sections, the ratio of the adjusted weight value to the original weight value can be calculated to obtain the change factor. The product of the change factor and the original weight value of each low-speed section can be calculated to obtain the weight value corresponding to each low-speed section.

Further, ph,new and pl,new can be calculated based on the following formulas:

ph,new=f(ph), pl,new=1-ph,new;

During specific adjustments, the weight of intervals with smaller original weights can be compressed, and then the weights of other intervals can be increased. Figure 4 is a schematic diagram of a function curve provided by an embodiment of the present application. As shown in Figure 4, when the proportion of the number of pixels is smaller, the compression rate can be larger, that is, the adjusted weight can be smaller. Assuming ph<pl, ph can be compressed according to the function shown in Figure 4, so that the adjusted ph is smaller and the adjusted pl is larger. That is, the lower the credibility of the interval with the smaller original weight, the greater the credibility of other intervals, thereby achieving the bias of the speed mapping.

Among them, the f(x) mapping function corresponding to f(ph) can be obtained through fourth-order curve fitting, and f(x) can be equal to AX. Among them, A can be [a1, a2, a3, a4], and X=[x ⁴ , x ³ , x ² , x ¹ ]. That is, f(x)=a1*x ⁴ +a2*x ³ +a3*x ⁴ +a4*x ¹ . The values of a1, a2, a3 and a4 can be set according to actual needs. For example, a1, a2, a3 and a4 can be: -47.63, 44.2, -9.96, 0.9055, -0.01439 respectively.

In the embodiment of this application, the pixel number proportion of each depth interval is counted through a single frame depth image, and the weight value corresponding to each depth interval is determined based on the pixel number proportion. The weight value and speed value corresponding to each depth interval can be determined based on 2. Safe speed. In this way, since there is no need for coordinate conversion operations on the point cloud, speed planning can be implemented only by knowing the depth information. Therefore, computing resource consumption can be saved to a certain extent.

It should be noted that after obtaining the final target safe speed based on the local map speed planning results and the single-frame depth map speed planning results: safe _spd = min (map_safe _spd , depth_safe _spd ), the target safe speed can be decomposed. Specifically, when performing collision query, the collision query can be performed by separating vertical and horizontal channels. Exemplarily, FIG. 5 is a channel schematic diagram of a collision query provided by an embodiment of the present application. As shown in FIG. 5 , collision query can be performed on the horizontal channel 01 and the vertical channel 02 respectively. Correspondingly, the second safe speed determined based on the single-frame depth map can also be decomposed into the second safe speed in the vertical direction and the second safe speed in the horizontal direction, and then the target safety speed in the vertical direction and the target safety in the horizontal direction can be obtained respectively. Speed, the target safe speed in the horizontal direction can be decomposed according to the direction of movement, and finally the target safe speed in the xyz three dimensions is obtained.

Optionally, the detected objects may include objects located in a space within a preset range indicated by the target flight direction, and the space within the preset range moves with the movement of the UAV in space. The space within the preset range indicated by the target flight direction is the space within the preset range in the target flight direction. For example, the space in the preset range can be adjacent to the nose of the drone, and as the drone continues to fly forward in the target flight direction, the space in the preset range also changes continuously. Correspondingly, if an object in the current surrounding environment falls into the space of the preset range indicated by the target flight direction, the object can be determined to be the detected object. In this way, objects that interfere with the drone's flight can be accurately identified as detected objects, so that in subsequent operations, intelligent behavioral decisions can be made based on whether the detected objects are target objects, that is, whether to approach the detected object or Detour around the detected object.

Optionally, the size of the space in the preset range is set with horizontal and vertical boundary sizes based on the shape of the drone. In this way, the shape of the drone corresponds to the horizontal and vertical boundary dimensions of the space in which the preset range is set, so that the space in the preset range can be more adapted to the shape characteristics of the drone, which can ensure to a certain extent Objects that interfere with the flight of the drone can be detected, thereby ensuring the safety of the drone and avoiding missing targets.

Specifically, the boundary size of the space in the preset range in the corresponding direction can be set according to the maximum size of the UAV in each direction, so that the space in the preset range can cover the UAV. In one implementation, the boundary size of the space in the preset range in each direction can be set to be larger than the maximum size of the drone in each direction, so as to avoid missing objects that may interfere with the flight of the drone as much as possible. Wherein, when the maximum size of the drone in this direction is larger, the boundary size in the corresponding direction of the preset range of space can be larger.

Optionally, the boundary size of the space in the preset range in the vertical direction is smaller than the boundary size in the horizontal direction. In one embodiment, the drone may be a multi-rotor drone. In order to improve the power performance, the size of the blade plane of this type of UAV is larger, that is, the size of the UAV in the horizontal direction is larger. At the same time, in order to improve the maneuverability of the UAV in the horizontal plane, the vertical size of such UAVs is often smaller than the plane size of the blades. In the embodiment of the present application, by setting the boundary size of the preset range of space in the vertical direction to be smaller than the boundary size in the horizontal direction, the space in the above-mentioned preset range can be used more efficiently to search for detected objects, ensuring no interference. Objects flown by humans and machines can be detected in a preset range of space. At the same time, it can also exclude, to a certain extent, objects in spaces that have little impact on flight safety from falling into the preset range, further reducing the risk of UAVs in space objects. Perceived computational overhead.

Furthermore, compared to setting the boundary size in the vertical direction to be equal to the boundary size in the horizontal direction, in the embodiment of the present application, the boundary size in the vertical direction of the space of the preset range is set to be smaller than the boundary size in the horizontal direction. The boundary size can avoid the problem of identifying the ground as a detected object, thus affecting the flight efficiency of the drone, when the drone is flying at a lower height and is closer to the ground.

Assuming that the boundary size in the vertical direction and the boundary size in the horizontal direction are both a, and the target object in this operation task is not the ground, then when the distance between the drone and the ground is less than a, it will The ground is identified as the detected object, and since the target object this time is not the ground, the detected object is identified as an obstacle. Correspondingly, if the drone is controlled to fly in a direction closer to the ground at this time, for example, if the user performs a stick operation on the control stick of the first control device, the drone will not respond normally to the user's stick operation. However, in this case, the ground does not actually pose a threat to the drone. In the embodiment of the present application, since the vertical boundary size of the preset range of space is smaller than the horizontal boundary size, as long as the distance between the drone and the ground is smaller than the vertical boundary size, it will not be The ground is recognized as an obstacle. Even when the distance between the drone and the ground is less than a, it can still respond normally to the user's swinging operation, thus adapting to the scene of hitting the swing near the ground.

It should be noted that the boundary size in the vertical direction can also be set to be equal to the boundary size in the horizontal direction. Correspondingly, in this case, if the object in the surrounding environment is a specified object, then when the object falls into the space of the preset range and the proportion of the object in the space of the preset range is greater than the preset ratio, The object is then determined to be the detected object. The specified object may include the ground, and the preset ratio may be set according to actual needs. In this way, you can avoid directly identifying obstacles on the ground that will interfere with the flight of the drone and set a higher tolerance for the ground. In the embodiment of this application, by setting a more reasonable ground filtering strategy, the processing of the ground can be made more reasonable, thereby ensuring safety while improving the response sensitivity of the UAV in near-ground scenarios.

In one embodiment, the object configuration of the drone is closer to a pie shape. Therefore, the space of the preset range can be spherical, and the query range in the horizontal direction of the space of the preset range can be closer to a round cake to avoid being too conservative. Figure 6 is a schematic diagram of a preset range provided by an embodiment of the present application. As shown in Figure 6, a plane rectangular coordinate system can be constructed in the normal plane tangential to the collision point trajectory, wherein the vertical axis in the coordinate system can be The z-axis of the navigation coordinate system is parallel, and the horizontal axis can be parallel to the horizontal plane of the navigation coordinate system. The area between line X and line Y can be obtained by translating the horizontal axis up and down according to the upper and lower tolerance. Since the border size in the vertical direction is smaller, the distance between line Y and the horizontal axis is smaller. Among them, the object query process can be queried sequentially from left to right as shown by the interlaced small circles in Figure 6. If an object is found in the corridor area formed between line X and line Y, it is considered that there is indeed a detected object. If no object is detected in the corridor area formed between line

Optionally, the above target flight direction can be determined through the following steps:

Step S61: Receive a flight control instruction sent by the first control device of the drone; the flight control instruction is generated based on the user's control operation of the control stick of the first control device.

Step S62: Determine the flight direction determined in response to the flight control instruction as the target flight direction.

The first control device may be a remote controller of the drone, and the control operation may be a lever operation on a control rod of the first control device. Assume that the user moves the stick in direction A, then the first control device can generate a flight control instruction indicating direction A, and the drone can receive the flight control instruction, and in response to the flight control instruction, determine the direction A indicated by the flight control instruction as Target flight direction. Correspondingly, upon receiving a flight control instruction generated based on the user's control operation of the control stick of the first control device, if the detected object in the target flight direction is the target object, the drone can fly along the stick. Continue flying in the indicated target flight direction to approach the target. If the detected object in the target flight direction is not the target object, the drone does not follow the target flight direction indicated by the stick, but changes the flight direction to bypass the detected object. In this way, it can be ensured that the drone responds to flight control instructions while ensuring flight safety.

Optionally, the above operation of obtaining the description information of the target object in the space may specifically include: receiving the description information of the target object sent by the second control device of the drone; wherein, the description of the target object The information is determined by the second control device based on the object selected by the user. In this way, the user can select the target object in this mission as needed based on the second control device, so that the drone's behavior decision-making for the detected object is more in line with the actual needs of the user. Specifically, the second control device may be the same device as the first control device. The second control device may also be a different device from the first control device. For example, the second control device may be an electronic device used by the user. In actual applications, the second control device can display a scene view to the user, and the scene view can provide multiple candidate objects. When the user's touch operation on the candidate object is detected, the candidate object selected by the touch operation is determined as the target object, and the description information of the selected candidate object is determined as the description information of the target object. Among them, the scene view can be a 3D map of the established working area, an orthophoto, etc. When obtaining the description information of the selected candidate object, the description information may be extracted based on images of the selected candidate object collected in historical operations.

Optionally, in another UAV flight control method, the description information of the target object in the space can be obtained; the planned trajectory of the UAV can be obtained; and the plan can be detected based on the sensor mounted on the UAV. The description information of the detected object in the space touched by the trajectory; if the description information of the detected object matches the description information of the target object, the drone is controlled to decelerate along the planned trajectory until it is close to the target object. The detected object; if the description information of the detected object does not match the description information of the target object, adjust the planned trajectory so that the UAV detours along the adjusted planned trajectory. Detected object. The implementation of each step in the flight control method may refer to the implementation of the same or similar steps mentioned above. The planned trajectory in the flight control method can be the trajectory that the UAV is currently flying to follow, and the planned trajectory can be the trajectory in the current flight direction. The above planned trajectory can be a plan automatically planned by the drone based on its own perception of the surrounding environment. It can also be a trajectory planned in response to external control instructions. The space touched by the planned trajectory may be a preset range of space touched by the planned trajectory.

For example, in one implementation, the operation of obtaining the planned trajectory of the UAV may include: obtaining a flight direction instruction, and obtaining the planned trajectory of the UAV according to the flight direction instruction. Wherein, obtaining the flight direction instruction may be receiving the flight direction instruction sent by the control device. Alternatively, it can also be an automatically generated flight direction instruction based on its own perception of the surrounding environment. Correspondingly, the flight trajectory indicated by the flight direction instruction can be determined as the planned trajectory of the UAV.

To sum up, the flight control method of the UAV provided by the embodiment of the present application automatically detects whether the description information of the detected object in the space touched by the planned trajectory matches the description information of the target object, and automatically controls the drone based on the detection result. Man-machine avoids or approaches the detected object, which can improve control efficiency and reduce labor costs to a certain extent. At the same time, it can avoid collisions between the drone and the detected object to ensure the safety of the drone's flight, while ensuring that the drone can get close to the target object in space, thus taking into account flight safety and scene requirements.

Figure 7 is a system flow diagram involved in a flight control method provided by an embodiment of the present application. As shown in Figure 7, target detection and tracking can be performed on a two-dimensional target (ie, a detected object) to determine the detected object. Whether the object is a target object. Then, the obstacle avoidance decision-making module confirms the position information of the two-dimensional target in the three-dimensional navigation coordinate system based on the target tracking results, the position and attitude of the aircraft and the gimbal, and when the UAV flies towards the two-dimensional target and the two-dimensional target is In the case of a target object, the speed planning module performs the next operation. The speed planning module can determine the target safe speed and control the underlying flight control system of the drone based on the target safe speed to control the drone to move along the target flight direction and decelerate at the target safe speed until it approaches the detected object. Among them, when performing speed planning, the environmental speed limit can be combined with the environment speed limit to avoid the determined target safe speed exceeding the environmental speed limit. On the contrary, the obstacle bypass module can perform the next operation to control the drone to bypass the detected object based on the underlying flight control system.

During flight, the drone will avoid obstacles based on detected objects. However, in scenarios where drones are used to perform inspection tasks, the drone needs to be able to approach the inspection target for inspection instead of bypassing the inspection target. In the embodiment of this application, the user only needs to select the inspection targets of this inspection task. During the flight of the drone, based on the detection results of the detected objects, the inspection targets and non-inspection targets in the space will be analyzed. It can distinguish the detected objects, automatically switch the response mode for the detected objects, and realize intelligent obstacle avoidance behavior decision-making. When the detected object is the inspection target, the local map and single-frame depth image are used to implement smooth speed planning to park near the target more safely. When the detected object is not the inspection target, a detour is performed to avoid the detected object to ensure flight safety.

Figure 8 is a block diagram of a UAV flight control device provided by an embodiment of the present application. The device may include: a memory 301 and a processor 302.

The memory 301 is used to store program codes;

The processor 302 calls the program code, and when the program code is executed, is used to perform the following operations:

Obtain description information of target objects in space;

Obtain the target flight direction of the drone;

To sum up, the flight control device of the unmanned aerial vehicle provided by the embodiment of the present application obtains the description information of the target object in the space and obtains the target flight direction of the unmanned aerial vehicle. Based on the sensor mounted on the drone, the description information of the detected object in the space indicated by the target flight direction is detected. If the description information of the detected object matches the description information of the target object, the drone is controlled to decelerate along the target flight direction until it approaches the detected object. If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object. In this way, by automatically detecting whether the description information of the detected object matches the description information of the target object, and automatically controlling the drone to avoid the detected object or approach the detected object based on the detection results, control efficiency can be improved to a certain extent and labor costs can be reduced. . At the same time, it can avoid collisions between the drone and the detected object to ensure the safety of the drone's flight, while ensuring that the drone can get close to the target object in space, thus taking into account flight safety and scene requirements.

Optionally, when controlling the drone to decelerate along the target flight direction to approach the detected object, the processor 302 is also configured to execute:

When the distance between the drone and the detected object is not greater than the specified stopping distance, the drone is controlled to stop flying.

Optionally, the processor 302 is also used to execute:

After the UAV stops flying, the UAV is controlled to perform a target operation on the detected object.

Optionally, the specified stopping distance is set based on one or more of the following factors:

The density of objects in said space;

Description information of the target object;

The type of operation of the drone.

Optionally, the processor 302 is specifically configured to execute:

Determine the target safe speed based on local map information and/or single-frame depth images; the target safe speed is positively related to the current distance between the drone and the detected object;

Control the drone to move along the target flight direction and decelerate at the target safe speed until it approaches the detected object.

Optionally, the local map information is generated based on multi-frame depth images of the current surrounding environment collected by the drone.

Optionally, the multi-frame depth images are obtained by sensors mounted on the drone in different directions.

Optionally, the processor 302 is also specifically configured to execute:

Based on the local map information, calculate the first safe speed of the drone;

Based on the single-frame depth image, calculate the second safe speed of the drone;

The target safe speed is determined based on the first safe speed and the second safe speed.

Optionally, the processor 302 is also specifically configured to execute:

Based on the local map information, determine the current distance between the drone and the detected object;

Based on the current distance and the designated stopping distance of the UAV, the first safe speed of the UAV is calculated; the first safe speed is positively related to the current distance difference, and the current distance difference is the The difference between the current distance and the specified stopping distance.

Optionally, the processor 302 is also specifically configured to execute:

Based on the number of pixels falling into each depth interval in the single-frame depth image and the total number of pixels in the single-frame depth image, determine the proportion of the number of pixels corresponding to each depth interval;

Determine the weight value corresponding to each depth interval according to the proportion of the number of pixels corresponding to each depth interval; the weight value corresponding to the depth interval is positively related to the proportion of the number of pixels;

The second safe speed is determined based on the speed value and weight value corresponding to each depth interval; the speed value corresponding to the depth interval is positively related to the depth value represented by the depth interval.

Optionally, the processor 302 is also specifically configured to execute:

The minimum value of the first safety speed and the second safety speed is determined as the target safety speed.

Optionally, the local map information is generated based on the multi-frame depth image when the distance between the drone and the detected object reaches a preset mapping distance, and the preset Mapping distance is directly related to the specified stopping distance of the drone.

Optionally, the detected object includes an object located in a preset range of space indicated by the target flight direction, and the preset range of space moves with the movement of the UAV in space.

Optionally, the size of the space in the preset range is set with horizontal and vertical boundary sizes based on the shape of the drone.

Optionally, the boundary size of the space in the preset range in the vertical direction is smaller than the boundary size in the horizontal direction.

Optionally, the processor 302 is also used to execute:

Receive flight control instructions sent by the first control device of the UAV; the flight control instructions are generated based on the user's control operation of the control lever of the first control device;

The flight direction determined in response to the flight control instruction is determined as the target flight direction.

Optionally, the processor 302 is also specifically configured to execute:

Receive the description information of the target object sent by the second control device of the drone; wherein the description information of the target object is determined by the second control device based on the object selected by the user.

The operations performed by the above device are similar to the corresponding steps in the above method and can achieve the same technical effect. To avoid repetition, they will not be described again here.

Further, embodiments of the present application also provide a computer-readable storage medium, which stores a computer program that, when run on a computer, causes the computer to perform each step in the above method, and can achieve the same results. The technical effects will not be repeated here to avoid repetition.

Furthermore, embodiments of the present application also provide a computer program product containing instructions, which when the instructions are run on a computer, cause the computer to execute the above method.

Furthermore, embodiments of the present application also provide an unmanned aerial vehicle, which is used to implement each step in the above method and can achieve the same technical effect. To avoid duplication, it will not be described again here.

The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in One location, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.

Various component embodiments of the present application may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor may be used in practice to implement some or all functions of some or all components in the computing processing device according to embodiments of the present application. The present application may also be implemented as an apparatus or device program (eg, computer program and computer program product) for performing part or all of the methods described herein. Such a program implementing the present application may be stored on a computer-readable medium, or may be in the form of one or more signals. Such signals may be downloaded from an Internet website, or provided on a carrier signal, or in any other form.

For example, FIG. 9 is a block diagram of a computing processing device provided by an embodiment of the present application. As shown in the figure, the diagram illustrates a computing processing device that can implement the method according to the present application. The computing processing device conventionally includes a processor 410 and a computer program product or computer-readable medium in the form of memory 420 . Memory 420 may be electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM. The memory 420 has a storage space 430 for program codes for executing any method steps in the above-described methods. For example, the storage space 430 for program codes may include individual program codes respectively used to implement various steps in the above method. These program codes can be read from or written into one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. Such computer program products are typically portable or fixed storage units as described with reference to FIG. 10 . The storage unit may have storage segments, storage spaces, etc. arranged similarly to the memory 420 in the computing processing device of FIG. 9 . The program code may, for example, be compressed in a suitable form. Typically, the storage unit includes computer readable code, ie code that can be read by, for example, a processor such as 410, which code, when executed by a computing processing device, causes the computing processing device to perform each of the methods described above. step.

Each embodiment in this specification is described in a progressive manner. Each embodiment focuses on its differences from other embodiments. The same and similar parts between the various embodiments can be referred to each other. Reference herein to "one embodiment," "an embodiment," or "one or more embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. In addition, please note that the examples of the word "in one embodiment" here do not necessarily all refer to the same embodiment.

In the instructions provided here, a number of specific details are described. However, it is understood that embodiments of the present application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this description.

In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The application may be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the element claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, third, etc. does not indicate any order. These words can be interpreted as names.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still implement the foregoing various The technical solutions described in the embodiments may be modified, or some of the technical features thereof may be equivalently replaced; however, these modifications or substitutions shall not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions in the embodiments of the present application.

Claims

A flight control method for a UAV, characterized in that the method includes:

Obtain description information of target objects in space;

Obtain the target flight direction of the drone;

Based on the sensor mounted on the drone, detect the description information of the detected object in the space indicated by the target flight direction;

If the description information of the detected object matches the description information of the target object, control the drone to decelerate along the target flight direction to approach the detected object;

If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object.
The method according to claim 1, characterized in that in the case of controlling the UAV to decelerate along the target flight direction to approach the detected object, the method further includes:

When the distance between the drone and the detected object is not greater than the specified stopping distance, the drone is controlled to stop flying.
The method of claim 2, further comprising:

After the drone stops flying, the drone is controlled to perform a target operation on the detected object.
The method according to claim 2 or 3, characterized in that the designated stopping distance is set based on one or more of the following factors:

The density of objects in said space;

Description information of the target object;

The type of operation of the drone.
The method according to any one of claims 1 to 3, characterized in that controlling the drone to decelerate along the target flight direction to approach the detected object includes:

Determine the target safe speed based on local map information and/or single-frame depth images; the target safe speed is positively related to the current distance between the drone and the detected object;

Control the drone to move along the target flight direction and decelerate at the target safe speed until it approaches the detected object.
The method of claim 5, wherein the local map information is generated based on multi-frame depth images of the current surrounding environment collected by the drone.
The method according to claim 6, characterized in that the multi-frame depth images are obtained by sensors mounted on the drone in different directions.
The method of claim 5, wherein determining the target safe speed based on local map information and/or a single frame depth image includes:

Based on the local map information, calculate the first safe speed of the drone;

Based on the single-frame depth image, calculate the second safe speed of the drone;

The target safe speed is determined based on the first safe speed and the second safe speed.
The method of claim 8, wherein calculating the first safe speed of the drone based on the local map information includes:

Based on the local map information, determine the current distance between the drone and the detected object;

Based on the current distance and the designated stopping distance of the UAV, the first safe speed of the UAV is calculated; the first safe speed is positively related to the current distance difference, and the current distance difference is the The difference between the current distance and the specified stopping distance.
The method of claim 8, wherein calculating the second safe speed of the drone based on the single-frame depth image includes:

Based on the number of pixels falling into each depth interval in the single-frame depth image and the total number of pixels in the single-frame depth image, determine the proportion of the number of pixels corresponding to each depth interval;

Determine the weight value corresponding to each depth interval according to the proportion of the number of pixels corresponding to each depth interval; the weight value corresponding to the depth interval is positively related to the proportion of the number of pixels;

The second safe speed is determined based on the speed value and weight value corresponding to each depth interval; the speed value corresponding to the depth interval is positively related to the depth value represented by the depth interval.
The method of claim 8, wherein determining the target safe speed based on the first safe speed and the second safe speed includes:

The minimum value of the first safety speed and the second safety speed is determined as the target safety speed.
The method according to claim 6, characterized in that the local map information is based on the multi-frame when the distance between the drone and the detected object reaches a preset mapping distance. Depth images are generated, and the preset mapping distance is positively related to the specified stopping distance of the drone.
The method according to any one of claims 1 to 3, characterized in that the detected object includes an object located in a preset range of space indicated by the target flight direction, and the space of the preset range changes with the Describe the movement of the drone in space.
The method according to any one of claims 13, characterized in that the size of the space in the preset range sets horizontal and vertical boundary sizes based on the shape of the drone.
The method according to claim 13, characterized in that the boundary size of the space in the preset range in the vertical direction is smaller than the boundary size in the horizontal direction.
The method according to any one of claims 1-3, characterized in that the method further includes:

Receive flight control instructions sent by the first control device of the UAV; the flight control instructions are generated based on the user's control operation of the control lever of the first control device;

The flight direction determined in response to the flight control instruction is determined as the target flight direction.
The method according to any one of claims 1 to 3, characterized in that said obtaining the description information of the target object in the space includes:

Receive the description information of the target object sent by the second control device of the drone; wherein the description information of the target object is determined by the second control device based on the object selected by the user.
A flight control device for an unmanned aerial vehicle, characterized in that the device includes a memory and a processor;

The memory is used to store program code;

The processor calls the program code, and when the program code is executed, is used to perform the following operations:

Obtain description information of target objects in space;

Obtain the target flight direction of the drone;

Based on the sensor mounted on the drone, detect the description information of the detected object in the space indicated by the target flight direction;

If the description information of the detected object matches the description information of the target object, control the drone to decelerate along the target flight direction to approach the detected object;

If the description information of the detected object does not match the description information of the target object, the drone is controlled to change its flight direction to bypass the detected object.
The device according to claim 1, wherein when controlling the drone to decelerate along the target flight direction to approach the detected object, the processor is further configured to execute:

When the distance between the drone and the detected object is not greater than the specified stopping distance, the drone is controlled to stop flying.
The device according to claim 19, wherein the processor is further configured to execute:

After the drone stops flying, the drone is controlled to perform a target operation on the detected object.
The device according to claim 19 or 20, characterized in that the designated stopping distance is set based on one or more of the following factors:

The density of objects in said space;

Description information of the target object;

The type of operation of the drone.
The device according to any one of claims 18-20, characterized in that the processor is specifically configured to execute:

Determine the target safe speed based on local map information and/or single-frame depth images; the target safe speed is positively related to the current distance between the drone and the detected object;

Control the drone to move along the target flight direction and decelerate at the target safe speed until it approaches the detected object.
The device according to claim 22, wherein the local map information is generated based on multi-frame depth images of the current surrounding environment collected by the drone.
The device according to claim 23, wherein the multi-frame depth images are obtained by sensors mounted on the drone in different directions.
The device according to claim 22, wherein the processor is further specifically configured to execute:

Based on the local map information, calculate the first safe speed of the drone;

Based on the single-frame depth image, calculate the second safe speed of the drone;

The target safe speed is determined based on the first safe speed and the second safe speed.
The device according to claim 25, wherein the processor is further configured to execute:

Based on the local map information, determine the current distance between the drone and the detected object;

Based on the current distance and the designated stopping distance of the UAV, the first safe speed of the UAV is calculated; the first safe speed is positively related to the current distance difference, and the current distance difference is the The difference between the current distance and the specified stopping distance.
The device according to claim 25, wherein the processor is further configured to execute:

Based on the number of pixels falling into each depth interval in the single-frame depth image and the total number of pixels in the single-frame depth image, determine the proportion of the number of pixels corresponding to each depth interval;

Determine the weight value corresponding to each depth interval according to the proportion of the number of pixels corresponding to each depth interval; the weight value corresponding to the depth interval is positively related to the proportion of the number of pixels;

The second safe speed is determined based on the speed value and weight value corresponding to each depth interval; the speed value corresponding to the depth interval is positively related to the depth value represented by the depth interval.
The device according to claim 25, wherein the processor is further configured to execute:

The minimum value of the first safety speed and the second safety speed is determined as the target safety speed.
The device according to claim 23, wherein the local map information is based on the multi-frame when the distance between the drone and the detected object reaches a preset mapping distance. Depth images are generated, and the preset mapping distance is positively related to the specified stopping distance of the drone.
The device according to any one of claims 18 to 20, wherein the detected object includes an object located in a space within a preset range indicated by the target flight direction, and the space within the preset range changes with the Describe the movement of the drone in space.
The device according to any one of claims 30, wherein the size of the space in the preset range is set with horizontal and vertical boundary sizes based on the shape of the drone.
The device according to claim 30, characterized in that the boundary size of the space in the preset range in the vertical direction is smaller than the boundary size in the horizontal direction.
The device according to any one of claims 18-20, characterized in that the processor is also used to execute:

Receive flight control instructions sent by the first control device of the UAV; the flight control instructions are generated based on the user's control operation of the control lever of the first control device;

The flight direction determined in response to the flight control instruction is determined as the target flight direction.
The device according to any one of claims 18-20, characterized in that the processor is further specifically configured to execute:

Receive the description information of the target object sent by the second control device of the drone; wherein the description information of the target object is determined by the second control device based on the object selected by the user.
A drone, characterized in that the drone is used to implement the method described in any one of claims 1-17.
A computer-readable storage medium, characterized by comprising instructions that, when run on a computer, cause the computer to execute the method described in any one of claims 1-17.
A computer program product containing instructions, characterized in that, when the instructions are run on a computer, the computer is caused to execute the method described in any one of claims 1-17.