A method and apparatus for obstacle measurement based on depth map and u-depth histogram

Abstract

The application discloses a kind of obstacle measurement method and device based on depth map and U depth histogram, first, the depth map of obstacle scene collected by depth camera is acquired, generates U depth histogram, the pixel in U depth histogram greater than the first threshold value is regarded as the interest point belonging to obstacle, and the target frame of obstacle in U depth histogram is obtained by interest point clustering;Then according to the target frame in U depth histogram and the focal length of depth camera, the horizontal position coordinate of the front center point of obstacle in the body coordinate system of unmanned aerial vehicle with depth camera, and the width and length of obstacle are calculated;According to the target frame of obstacle in U depth histogram, the corresponding obstacle target frame is positioned in depth map, and the vertical position of the front center point of obstacle cuboid in the body coordinate system of unmanned aerial vehicle and the height of obstacle are calculated.The application has small amount of calculation, higher timeliness, can be deployed on the light weight platform with limited computing resources, and complete dynamic obstacle avoidance task.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of obstacle detection, and particularly relates to an obstacle measurement method and device based on a depth map and a U-depth histogram. BACKGROUND

[0002] The unmanned aerial vehicle (UAV) is widely used in civil and military fields due to its lightness, flexibility, strong maneuverability, and good concealment. With the development of UAV technology, autonomous flight of the UAV has become a hot research topic. As the environment in which the UAV autonomously performs a task is becoming more and more complex, the obstacle avoidance capability is crucial for autonomous flight of the UAV, and fast, accurate, and low-consumption perception of obstacles in a complex environment is a prerequisite for completing autonomous obstacle avoidance of the UAV.

[0003] Currently, methods for obstacle detection mainly include a vision-based method and a radar-based method. Compared with the radar-based method, the vision-based method has the advantages of lower device cost, the ability to identify different objects, higher measurement accuracy of object height and width, smaller data volume, and higher acquisition frequency, and is thus more suitable for being used on the UAV with limited computing power in a high real-time requirement scene.

[0004] For environment perception for obstacle avoidance, in addition to the position information and depth information of obstacles, the three-dimensional size information of obstacles is also important information for determining whether the obstacle avoidance path planning can be efficiently, quickly, and accurately completed. However, the prior art only concerns the depth and position information of obstacles and does not estimate the size of the obstacles. The prior art identifies the obstacle category and image coordinate information through a neural network, needs to collect images of obstacles to train the neural network, and is only effective for known obstacles with sufficient images. Meanwhile, the prior art usually has low timeliness and cannot be applied to a high-speed dynamic obstacle avoidance scene. SUMMARY

[0005] The application aims to provide an obstacle measurement method and device based on a depth map and a U-depth histogram, which can quickly obtain the depth, position, and size information of unknown obstacles and improve timeliness.

[0006] To achieve the above-mentioned purpose, the technical scheme of the application is as follows:

[0007] An obstacle measurement method based on a depth map and a U-depth histogram is applied to a UAV carrying a depth camera, and includes the following steps.

[0008] An obstacle scene depth map collected by the depth camera is obtained, and a U-depth histogram is generated. The horizontal position on the depth map is represented by the horizontal coordinate of the U-depth histogram, and the depth value on the depth map is represented by the vertical coordinate of the U-depth histogram. The pixel value on the U-depth histogram represents the number of pixels with the same horizontal position pixel depth value equal to the vertical coordinate of the U-depth histogram.

[0009] The pixels with pixel values greater than the first threshold value in the U-depth histogram are regarded as interest points belonging to the obstacle, and a target frame of the obstacle in the U-depth histogram is obtained by clustering the interest points;

[0010] According to the target frame in the U-depth histogram and the focal length of the depth camera, the horizontal position coordinate of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV and the width and length of the obstacle are calculated.

[0011] According to the target frame of the obstacle in the U-depth histogram, the corresponding obstacle target frame is located in the depth map, and the vertical position of the front center point of the obstacle cuboid in the body coordinate system of the UAV and the height of the obstacle are calculated.

[0012] Further, the first threshold value is calculated by the following formula:

[0013]

[0014] where f is the focal length of the depth camera, μ is the size of a pixel, d bin is the pixel value of the current pixel point in the U-depth histogram, T h is a preset obstacle height threshold in the obstacle avoidance scene.

[0015] Further, the calculation of the horizontal position coordinate of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV and the width and length of the obstacle is calculated by the following formula:

[0016]

[0017]

[0018]

[0019] l o =d b -d t

[0020] where f is the focal length of the depth camera, (i l ,d t ) represents the top-left vertex coordinate of the target frame in the U-depth histogram, (i r ,d b ) represents the bottom-right vertex coordinate of the target frame in the U-depth histogram, w o represents the width of the obstacle, l o represents the length of the obstacle, and (x, y) represents the horizontal position coordinate of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV.

[0021] Further, the vertical position of the obstacle cuboid front center point in the body coordinate system of the unmanned aerial vehicle and the height of the obstacle are calculated, comprising:

[0022]

[0023]

[0024] wherein f is the focal length of the depth camera, z represents the vertical position of the obstacle cuboid front center point in the body coordinate system of the unmanned aerial vehicle, the vertical position of the obstacle cuboid front center point in the body coordinate system of the unmanned aerial vehicle, h t represents the height coordinate of the left upper vertex of the obstacle target frame in the depth map, h b represents the height coordinate of the right lower vertex of the obstacle target frame in the depth map, and h represents the height of the obstacle.

[0025] Further, the corresponding obstacle target frame in the depth map is located according to the target frame of the obstacle in the U-depth histogram, comprising:

[0026] The pixel points with the horizontal position index in [i l , i r ] range and the depth value in [d t , d b ] range in the depth map are clustered, and the clustering result is framed with a rectangular frame to obtain the corresponding obstacle target frame in the depth map. l , d t ) represents the left upper vertex coordinate of the target frame in the U-depth histogram, and (i r , d b ) represents the right lower vertex coordinate of the target frame in the U-depth histogram.

[0027] The application also provides an obstacle measuring device based on a depth map and a U-depth histogram, which is applied to an unmanned aerial vehicle carrying a depth camera, and comprises:

[0028] A histogram generation module is configured to acquire an obstacle scene depth map collected by the depth camera and generate a U-depth histogram, wherein the horizontal coordinate of the U-depth histogram represents the horizontal position on the depth map, the vertical coordinate represents the depth value on the depth map, and the pixel value on the U-depth histogram represents the number of pixels with the same horizontal position pixel depth value equal to the vertical coordinate of the U-depth histogram on the depth map;

[0029] A histogram clustering module is configured to regard the pixels with a pixel value greater than a first threshold value in the U-depth histogram as interest points belonging to the obstacle, and acquire a target frame of the obstacle in the U-depth histogram through interest point clustering.

[0030] a length-width estimation module, configured to calculate a horizontal position coordinate of a front center point of the obstacle in a body coordinate system of the depth camera-carrying UAV and a width and a length of the obstacle according to the target frame in the U-depth histogram and a focal length of the depth camera;

[0031] a height estimation module, configured to locate a corresponding obstacle target frame in a depth map according to the target frame of the obstacle in the U-depth histogram, and calculate a vertical position of the front center point of the obstacle cuboid in the body coordinate system of the UAV and a height of the obstacle.

[0032] Further, the first threshold is calculated by the following formula:

[0033]

[0034] wherein f is the focal length of the depth camera, μ is the size of a pixel, d bin is a pixel value of the current pixel point in the U-depth histogram, T h is a preset obstacle height threshold in an obstacle avoidance scene.

[0035] Further, the calculation of the horizontal position coordinate of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV and the width and the length of the obstacle is calculated by the following formula:

[0036]

[0037]

[0038]

[0039] l o =d b -d t

[0040] wherein f is the focal length of the depth camera, (i l ,d t ) represents the top-left vertex coordinate of the target frame in the U-depth histogram, (i r ,d b ) represents the bottom-right vertex coordinate of the target frame in the U-depth histogram, w o represents the width of the obstacle, l o represents the length of the obstacle, and (x, y) represents the horizontal position coordinate of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV.

[0041] Further, the calculation of the vertical position of the front center point of the obstacle cuboid in the body coordinate system of the UAV and the height of the obstacle includes:

[0042]

[0043]

[0044] wherein f is the focal length of the depth camera, z represents the vertical position of the front face center point of the obstacle cuboid in the body coordinate system of the UAV, the vertical position of the front face center point of the obstacle cuboid in the body coordinate system of the UAV, h t represents the height coordinate of the top-left vertex of the obstacle target frame in the depth map, h b represents the height coordinate of the bottom-right vertex of the obstacle target frame in the depth map, and h represents the height of the obstacle.

[0045] Further, the height estimation module locates the corresponding obstacle target frame in the depth map according to the target frame of the obstacle in the U-depth histogram, and performs the following operations:

[0046] The pixel points with the horizontal position index in the range of [i l ,i r ] and the depth value in the range of [d t ,d b ] in the depth map are clustered, and the clustering result is framed with a rectangular frame to obtain the corresponding obstacle target frame in the depth map, wherein (i l ,d t ) represents the top-left vertex coordinate of the target frame in the U-depth histogram, and (i r ,d b ) represents the bottom-right vertex coordinate of the target frame in the U-depth histogram.

[0047] The application provides an obstacle measurement method and device based on a depth map and a U-depth histogram. The depth map obtained by a depth camera and the U-depth histogram obtained from the depth map are used. According to the statistical characteristics of the depth information, the depth, position and size information of an unknown obstacle can be quickly obtained without prior knowledge of the obstacle. The application has small calculation amount and high timeliness, can be deployed on a lightweight platform with limited computing resources, and can complete a dynamic obstacle avoidance task. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The flowchart of the obstacle measurement method based on the depth map and the U-depth histogram is provided.

[0049] Figure 2 The U-depth histogram generation diagram for the depth map is provided. DETAILED DESCRIPTION

[0050] The technical solutions of the application will be further described in detail below in combination with the drawings and examples. The following examples do not constitute a limitation on the application.

[0051] The application provides a method for obstacle detection and size measurement based on a depth map and a U-depth histogram.

[0052] In one embodiment, as shown in Figure 1 A method for obstacle detection and size measurement based on a depth map and a U-depth histogram is applied to a UAV with a depth camera, and includes the following steps.

[0053] In step S1, a depth map of an obstacle scene collected by a depth camera is acquired, and a U-depth histogram is generated, wherein the horizontal coordinate of the U-depth histogram represents the horizontal position on the depth map, the vertical coordinate represents the depth value on the depth map, and the pixel value on the U-depth histogram represents the number of pixels with the same horizontal position and the same depth value as the vertical coordinate of the U-depth histogram.

[0054] In this embodiment, a depth image is acquired by a depth camera in a scene with an obstacle, and a depth map of the obstacle scene is obtained. The method for acquiring the depth image includes a laser radar depth imaging method, a computer stereo vision imaging method, a coordinate measuring machine method, a Moire fringe method, a structured light method and the like. In this embodiment, the depth map is completed based on the computer stereo vision imaging method. The depth map is represented by a matrix with a channel number of 1 and a size of n x m. n and m represent the width and height of the image respectively. The pixel value of each point in the depth map represents the depth value of the point.

[0055] Then, a corresponding U-depth histogram is generated according to the acquired original depth map. As shown in Figure 2 The number of different depths in the entire depth map is counted and recorded as s. The size of the U-depth histogram corresponding to the depth map with a size of n x m is s x m. The horizontal coordinate of the U-depth histogram represents the horizontal position on the depth map, the vertical coordinate represents the depth value on the depth map, and the pixel value on the U-depth histogram represents the number of pixels with the same horizontal position and the same depth value as the vertical coordinate of the U-depth histogram.

[0056] For example, Figure 2 The horizontal coordinates of the U-depth histogram are consistent with those of the original depth map. The depth values of the pixels in the original depth map are different from 0 to 5, that is, six different depth values are included. Therefore, the U-depth histogram has six corresponding coordinates, which are 0, 1, 2, 3, 4 and 5 from top to bottom.

[0057] For the first horizontal position of the U depth histogram, among all the pixels in the same horizontal position of the original depth map corresponding to the first horizontal position, there is one pixel with a depth value of 0, three pixels with a depth value of 1, no pixel with a depth value of 2, one pixel with a depth value of 3, and no pixel with a depth value of 4 or 5. Therefore, the pixel values corresponding to the first horizontal position of the U depth histogram from top to bottom are 1, 3, 0, 1, 0, and 0. The pixel values corresponding to other horizontal positions of the U depth histogram are obtained in the same manner, which will not be described herein.

[0058] In step S2, the pixels with pixel values greater than the first threshold value in the U depth histogram are regarded as interest points belonging to the obstacle, and the target box of the obstacle in the U depth histogram is obtained through interest point clustering.

[0059] When the obstacle is located directly in front of the depth camera, there are more pixels with the same depth value in the obtained original depth map. Therefore, the pixel value corresponding to the depth of the current obstacle in the U depth histogram is relatively greater than other pixel values. According to the above characteristics, the obstacle interest points can be found according to the U depth histogram.

[0060] In the embodiment, the pixels with pixel values greater than a certain threshold T PI on the U depth histogram are regarded as interest points belonging to the obstacle, where the threshold T PI is defined as:

[0061]

[0062] where f is the focal length of the depth camera, μ is the size of one pixel, d bin is the pixel value of the current pixel point in the U depth histogram, and T h is a preset obstacle height threshold in the obstacle avoidance scene.

[0063] After all the interest points are screened through the threshold, distance threshold clustering can be used to cluster the interest points with similar horizontal positions and pixel values into a class, and the clustering result is boxed with a rectangular frame, so as to obtain the target box of the obstacle in the U depth histogram. Specifically, the pixels with an image plane Euclidean distance less than a distance threshold dif dis from the interest points and a pixel value difference less than a pixel threshold dif pix are regarded as belonging to the same obstacle, and the clustering result is boxed with a rectangular frame, that is, the target box of the obstacle in the U depth histogram is obtained. In the embodiment, the top-left and bottom-right vertex coordinates of the target box are denoted as (i l , d t ) and (i r , d b ), respectively.

[0064] Step S3, according to the target frame in the U-depth histogram and the focal length of the depth camera, the horizontal position coordinate of the front center point of the obstacle cuboid in the body coordinate system of the depth camera-carrying UAV and the width and length of the obstacle are calculated.

[0065] Considering the passability of the UAV in the UAV obstacle avoidance scene, it is reasonable and convenient to model a single obstacle as a three-dimensional cuboid, while retaining sufficient safety margin. Therefore, the measurement of the obstacle is to estimate the length, width and height of the cuboid modeled by the obstacle. According to the rectangular frame of the obstacle in the U-depth histogram, the horizontal position coordinates x and y of the front center point of the obstacle cuboid in the body coordinate system of the depth camera-carrying UAV are calculated:

[0066]

[0067]

[0068] And the width estimation w of the obstacle cuboid o And the approximate length l o Estimation:

[0069]

[0070] l o =d b -d t .

[0071] Where f is the focal length of the depth camera, (i l ,d t ) represents the top-left vertex coordinates of the target frame in the U-depth histogram, (i r ,d b ) represents the bottom-right vertex coordinates of the target frame in the U-depth histogram, w o represents the width of the obstacle, l o represents the length of the obstacle, and (x, y) represents the horizontal position coordinates of the front center point of the obstacle in the body coordinate system of the depth camera-carrying UAV.

[0072] Step S4, according to the target frame of the obstacle in the U-depth histogram, the corresponding obstacle target frame is located in the depth map, and the vertical position of the front center point of the obstacle cuboid in the body coordinate system of the UAV and the height of the obstacle are calculated.

[0073] The depth map is consistent with the horizontal position of the U-depth histogram, so the horizontal position index is in the range of [i l ,i r ] in the depth map, and the original depth map depth value is in the range of [d t ,d bthe pixel clustering in the range, the clustering result is framed with a rectangular frame, the left upper and right lower vertex coordinates of the rectangular frame are recorded as (i l ,h t ) and (i r ,h b ) respectively, and then the corresponding obstacle position in the depth map is obtained.

[0074] Therefore, the vertical position z of the front center point of the obstacle in the body coordinate system of the unmanned aerial vehicle and the height h of the obstacle can be calculated according to the following formula:

[0075]

[0076]

[0077] wherein f is the focal length of the depth camera, z represents the vertical position of the front center point of the obstacle cuboid in the body coordinate system of the unmanned aerial vehicle, h t represents the height coordinate of the left upper vertex of the obstacle target frame in the depth map, h b represents the height coordinate of the right lower vertex of the obstacle target frame in the depth map, and h represents the height of the obstacle.

[0078] The method does not require prior information of the obstacle and image data set, and does not use a complex depth image target detection algorithm. According to the statistical characteristics of the obstacle in the depth map, the position detection and size measurement of the obstacle are completed only by using the obstacle scene depth map. The algorithm is simple and has small calculation amount, and is suitable for deployment on the unmanned aerial vehicle platform with high real-time requirement and limited computing power, and can quickly complete obstacle perception.

[0079] In another embodiment, the application also provides an obstacle measurement device based on a depth map and a U-depth histogram, applied to an unmanned aerial vehicle carrying a depth camera. The obstacle measurement device based on the depth map and the U-depth histogram comprises:

[0080] A histogram generation module is configured to acquire an obstacle scene depth map collected by the depth camera, and generate a U-depth histogram. The horizontal position on the depth map is represented by the horizontal coordinate of the U-depth histogram, the depth value on the depth map is represented by the vertical coordinate of the U-depth histogram, and the pixel value on the U-depth histogram represents the number of pixels with the same horizontal position pixel depth value equal to the vertical coordinate of the U-depth histogram.

[0081] A histogram clustering module is configured to regard the pixels with a pixel value greater than a first threshold value in the U-depth histogram as interest points belonging to the obstacle, and acquire a target frame of the obstacle in the U-depth histogram through interest point clustering.

[0082] The length and width estimation module is used to calculate the horizontal position coordinates of the center point of the obstacle's front face in the body coordinate system of the drone equipped with the depth camera, as well as the width and length of the obstacle based on the target box in the U depth histogram and the focal length of the depth camera;

[0083] The height estimation module is used to locate the corresponding obstacle target frame in the depth map based on the target frame of the obstacle in the U depth histogram, and calculate the vertical position of the center point of the front of the obstacle cuboid in the drone's body coordinate system and the height of the obstacle.

[0084] For the specific limitations of the obstacle measurement device based on the depth map and U depth histogram, please refer to the limitations of the obstacle measurement method based on the depth map and U depth histogram above, which will not be repeated here. The various modules in the above-mentioned obstacle measurement device based on the depth map and U depth histogram can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0085] The memory and processor are electrically connected, directly or indirectly, to enable data transmission or interaction. For example, these components may be electrically connected via one or more communication buses or signal lines. The memory stores a computer program executable on the processor, and the processor executes the computer program stored in the memory to implement the network topology layout method according to the embodiments of the present invention.

[0086] The memory may be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. The memory is used to store a program, and the processor executes the program after receiving an execution instruction.

[0087] The processor can be an integrated circuit chip with data processing capability. The processor can be a general purpose processor, including a central processing unit (CPU), a network processor (NP), or the like. The processor can implement or execute the methods, steps, and logical block diagrams disclosed in the embodiments of the present application. The general purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0088] The above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Those skilled in the art can make various corresponding changes and modifications to the present application without departing from the spirit and essence of the present application, and these corresponding changes and modifications shall all belong to the protection scope of the claims of the present application.

Claims

1. An obstacle measurement method based on depth map and U-depth histogram, applied to a drone equipped with a depth camera, characterized in that: The obstacle measurement method based on the depth map and the U depth histogram includes: Obtain a depth map of the obstacle scene captured by the depth camera and generate a U depth histogram, where the abscissa of the U depth histogram represents the horizontal position on the depth map, the ordinate represents the depth value on the depth map, and the pixel value on the U depth histogram represents the number of pixels whose depth values ​​at the same horizontal position on the depth map are equal to the ordinate of the U depth histogram; Pixels in the U depth histogram whose pixel values ​​are greater than the first threshold are regarded as interest points belonging to the obstacle, and the target box of the obstacle in the U depth histogram is obtained by clustering the interest points; Based on the target frame in the U depth histogram and the focal length of the depth camera, calculate the horizontal position coordinates of the center point of the obstacle in the body coordinate system of the drone equipped with the depth camera, as well as the width and length of the obstacle; According to the target frame of the obstacle in the U depth histogram, the corresponding obstacle target frame is located in the depth map, and the vertical position of the center point of the front of the obstacle cuboid in the drone's body coordinate system and the height of the obstacle are calculated.

2. The obstacle measurement method based on depth map and U depth histogram according to claim 1, characterized in that: The first threshold is calculated by the following formula: Where f is the focal length of the depth camera, μ is the size of a pixel, and d bin is the pixel value of the current pixel in the U depth histogram, T h It is the obstacle height threshold in the preset obstacle avoidance scenario.

3. The obstacle measurement method based on depth map and U depth histogram according to claim 1, characterized in that: The calculation formula for calculating the horizontal position coordinates of the center point of the obstacle in the body coordinate system of the drone equipped with the depth camera and the width and length of the obstacle is as follows: l o =d b -d t Where, f is the focal length of the depth camera, (i l ,d t ) represents the upper left vertex coordinate of the target box in the U depth histogram, (i r ,d b ) represents the coordinate of the lower right vertex of the target box in the U depth histogram, w o Indicates the width of the obstacle, l o represents the length of the obstacle, and (x, y) represents the horizontal position coordinates of the center point of the obstacle in the body coordinate system of the drone equipped with the depth camera.

4. The obstacle measurement method based on depth map and U depth histogram according to claim 1, characterized in that: Calculating the vertical position of the center point of the front face of the obstacle cuboid in the drone's body coordinate system and the height of the obstacle includes: Among them, f is the focal length of the depth camera, z represents the vertical position of the center point of the front of the obstacle cuboid in the body coordinate system of the drone, h represents the vertical position of the center point of the front of the obstacle cuboid in the body coordinate system of the drone, t Represents the height coordinate of the upper left vertex of the obstacle target box in the depth map, h b The height coordinate of the lower right vertex of the obstacle target box in the depth map, h represents the height of the obstacle.

5. The obstacle measurement method based on depth map and U depth histogram according to claim 1, characterized in that: The step of locating the corresponding obstacle target frame in the depth map according to the target frame of the obstacle in the U depth histogram includes: Index the horizontal position in the depth map into [i l ,i r ] range while the depth map depth value is within [d t ,d b ] range, and the clustering result is framed by a rectangular frame to obtain the target frame corresponding to the obstacle in the depth map. (i l ,d t ) represents the upper left vertex coordinate of the target box in the U depth histogram, (i r ,d b ) represents the coordinate of the lower right vertex of the target box in the U depth histogram.

6. An obstacle measurement device based on depth map and U depth histogram, applied to a drone equipped with a depth camera, characterized in that: The obstacle measurement device based on the depth map and the U depth histogram includes: A histogram generation module is used to obtain a depth map of an obstacle scene collected by a depth camera and generate a U depth histogram, wherein the abscissa of the U depth histogram represents the horizontal position on the depth map, the ordinate represents the depth value on the depth map, and the pixel value on the U depth histogram represents the number of pixels whose depth values ​​at the same horizontal position on the depth map are equal to the ordinate of the U depth histogram; A histogram clustering module is used to take pixels in the U depth histogram whose pixel values ​​are greater than a first threshold as points of interest belonging to obstacles, and obtain a target box of the obstacle in the U depth histogram through interest point clustering; The length and width estimation module is used to calculate the horizontal position coordinates of the center point of the obstacle's front face in the body coordinate system of the drone equipped with the depth camera, as well as the width and length of the obstacle based on the target box in the U depth histogram and the focal length of the depth camera; The height estimation module is used to locate the corresponding obstacle target frame in the depth map based on the target frame of the obstacle in the U depth histogram, and calculate the vertical position of the center point of the front of the obstacle cuboid in the drone's body coordinate system and the height of the obstacle.

7. The obstacle measurement device based on depth map and U depth histogram according to claim 6, characterized in that: The first threshold is calculated by the following formula: Where f is the focal length of the depth camera, μ is the size of a pixel, and d bin is the pixel value of the current pixel in the U depth histogram, T h It is the obstacle height threshold in the preset obstacle avoidance scenario.

8. The obstacle measurement device based on depth map and U depth histogram according to claim 6, characterized in that: The calculation formula for calculating the horizontal position coordinates of the center point of the obstacle in the body coordinate system of the drone equipped with the depth camera and the width and length of the obstacle is as follows: l o =d b -d t Where, f is the focal length of the depth camera, (i l , d t ) represents the upper left vertex coordinate of the target box in the U depth histogram, (i r , d b ) represents the coordinate of the lower right vertex of the target box in the U depth histogram, w o Indicates the width of the obstacle, l o represents the length of the obstacle, and (x, y) represents the horizontal position coordinates of the center point of the obstacle in the body coordinate system of the drone equipped with the depth camera.

9. The obstacle measurement device based on depth map and U depth histogram according to claim 6, characterized in that: Calculating the vertical position of the center point of the front face of the obstacle cuboid in the drone's body coordinate system and the height of the obstacle includes: Among them, f is the focal length of the depth camera, z represents the vertical position of the center point of the front of the obstacle cuboid in the body coordinate system of the drone, h represents the vertical position of the center point of the front of the obstacle cuboid in the body coordinate system of the drone, t Represents the height coordinate of the upper left vertex of the obstacle target box in the depth map, h b The height coordinate of the lower right vertex of the obstacle target box in the depth map, h represents the height of the obstacle.

10. The obstacle measurement device based on depth map and U depth histogram according to claim 6, characterized in that: The height estimation module locates the corresponding obstacle target frame in the depth map based on the target frame of the obstacle in the U depth histogram and performs the following operations: Index the horizontal position in the depth map into [i l ,i r ] range while the depth map depth value is within [d t , d b ] range, and the clustering result is framed by a rectangular frame to obtain the target frame corresponding to the obstacle in the depth map. (i l , d t ) represents the upper left vertex coordinate of the target box in the U depth histogram, (i r , d b ) represents the coordinate of the lower right vertex of the target box in the U depth histogram.

Images