WO2021169704A1

WO2021169704A1 - Method, device and apparatus for determining depth of gesture, and storage medium

Info

Publication number: WO2021169704A1
Application number: PCT/CN2021/073715
Authority: WO
Inventors: 黄少光; 许秋子
Original assignee: 深圳市瑞立视多媒体科技有限公司
Priority date: 2020-02-26
Filing date: 2021-01-26
Publication date: 2021-09-02
Also published as: CN111354029B; CN111354029A

Abstract

Provided are a method, device and apparatus for determining the depth of a gesture, and a storage medium capable of improving accuracy of depth information of a hand feature point measured by a depth camera. The method comprises: capturing a depth image of a hand, and acquiring a preset wrist feature point and a depth value thereof contained in the depth image of the hand (101); determining, according to the depth value, whether the preset wrist feature point is a noise point (102); if the preset wrist feature point is a noise point, acquiring a preset finger feature point closest to the preset wrist feature point, and using same as a target finger feature point (103); determining whether the target finger feature point is a noise point (104); if the target finger feature point is not a noise point, determining whether all of pixel points on a connection line between the preset wrist feature point and the target finger feature point are noise points (105); and if not all of the pixel points are noise points, calculating a mean depth of all normal points on the connection line between the preset wrist feature point and the target finger feature point, and using same as the depth of the preset wrist feature point (106).

Description

Method, device, equipment and storage medium for determining gesture depth

Technical field

The present invention relates to the field of image processing technology, and in particular to a method, device, device and storage medium for determining the depth of a gesture.

Background technique

In recent years, with the development of technologies such as human-computer interaction and machine vision, depth cameras have gradually been applied to application scenarios such as object recognition and scene modeling. Unlike ordinary color cameras that can only take 2D images of objects, depth cameras are a kind of 3D cameras. The depth images they take include not only color information, but also depth information, that is, the distance between the actual object feature points in the image and the camera. , Through the depth image, the three-dimensional coordinates of the object can be obtained, so as to restore the real scene and realize applications such as scene modeling.

At present, when using a depth camera to collect 3D images of a user’s hand, due to ambient light, hand shaking, etc., it is easy to cause noise in the collected hand depth information. For example, for the point of the fingertip, because the fingertip is a hand The end of the shook has the largest amplitude, so it is very likely that the depth information at it may be mistakenly detected as the depth information of the fingertip. Therefore, the accuracy of the existing depth information based on the depth camera to measure the depth information of the hand feature points needs to be improved.

Summary of the invention

The main purpose of the present invention is to provide a gesture depth determination method, device, equipment and storage medium, aiming to improve the accuracy of measuring the depth information of the hand feature points based on the depth camera.

The first aspect of the present invention provides a method for determining the depth of a gesture, and the method for determining the depth of a gesture includes:

Acquiring a hand depth image by a depth camera, and acquiring preset wrist feature points included in the hand depth image and depth values of the preset wrist feature points;

Judging whether the preset wrist feature point is a noise point according to the depth value;

When the preset wrist feature point is a noise point, acquiring the preset finger feature point closest to the preset wrist feature point as the target finger feature point;

Judging whether the feature point of the target finger is a noise point;

When the target finger feature point is not a noise point, determining whether pixels other than the target finger feature point are all noise points on the line connecting the preset wrist feature point and the target finger feature point;

When the pixel points other than the target finger feature points on the line connecting the preset wrist feature points and the target finger feature points are not all noise points, the preset wrist feature points and the target finger are calculated The median depth of all normal points on the line of the characteristic points is taken as the depth of the preset wrist characteristic points.

Optionally, in the first implementation of the first aspect of the present invention, the step of judging whether the preset wrist feature point is a noise according to the depth value includes:

Identifying the non-noise points of the hand region in the hand depth image according to the working depth range of the depth camera, and obtaining the depth value of the non-noise points;

Calculating the median value of the acquired depth values of all non-noise points, and calculating the absolute value of the difference between the depth value of the preset wrist feature point and the median value;

Judging whether the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold;

When the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold value, it is determined that the preset wrist feature point is a noise point.

Optionally, in the second implementation manner of the first aspect of the present invention, the step of judging whether the feature point of the target finger is a noise point includes:

Acquiring the depth value of the feature point of the target finger, and calculating the absolute value of the difference between the depth value of the feature point of the target finger and the median value;

Judging whether the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold value, determining that the target finger feature point is a noise point;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is less than the preset threshold value, it is determined that the target finger feature point is not a noise point.

Optionally, in the third implementation manner of the first aspect of the present invention, after the step of determining whether the target finger feature point is a noise point, the method further includes:

When the target finger feature point is a noise point, obtain the depth of the normal point closest to the preset wrist feature point, and use the depth of the normal point closest to the preset wrist feature point as the preset wrist feature point depth.

Optionally, in the fourth implementation manner of the first aspect of the present invention, pixels other than the target finger feature point are determined on the line connecting the preset wrist feature point and the target finger feature point After the steps of whether the points are noise points, it also includes:

When on the line connecting the preset wrist feature point and the target finger feature point, pixel points other than the target finger feature point are noise points, obtain the normal point closest to the preset wrist feature point Depth, the depth of the normal point closest to the preset wrist feature point is taken as the depth of the preset wrist feature point.

A second aspect of the present invention provides a gesture depth determining device, the gesture depth determining device includes:

An acquisition module, configured to acquire a hand depth image through a depth camera, and acquire preset wrist feature points contained in the hand depth image and the depth value of the preset wrist feature points;

The first judgment module is configured to judge whether the preset wrist feature point is a noise point according to the depth value;

An acquiring module, configured to acquire a preset finger feature point closest to the preset wrist feature point as the target finger feature point when the preset wrist feature point is a noise point;

The second judgment module is used to judge whether the feature point of the target finger is a noise point;

The third judgment module is used for judging the pixel points other than the target finger feature points on the line between the preset wrist feature points and the target finger feature points when the target finger feature points are not noise points Are they all noisy;

A calculation module for calculating the preset wrist feature point when the pixel points other than the target finger feature point on the line connecting the preset wrist feature point and the target finger feature point are not all noise points The median depth of all normal points on the line with the feature point of the target finger is taken as the depth of the preset wrist feature point.

A third aspect of the present invention provides a gesture depth determining device. The gesture depth determining device includes a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor communicate with each other through a wire. Connect; the at least one processor calls the instructions in the memory, so that the gesture depth determination device executes the above gesture depth determination method.

A fourth aspect of the present invention provides a storage medium in which instructions are stored, which when run on a computer, cause the computer to execute the aforementioned gesture depth determination method.

The present invention collects the hand depth image through a depth camera, obtains the preset wrist feature points and the depth value of the preset wrist feature points contained in the hand depth image; judges whether the preset wrist feature points are noise points according to the depth value; when preset When the wrist feature point is a noise point, obtain the preset finger feature point closest to the preset wrist feature point as the target finger feature point; determine whether the target finger feature point is a noise point; when the target finger feature point is not a noise point, determine the preset On the connection line between the wrist feature point and the target finger feature point, whether the pixel points other than the target finger feature point are all noise; when the preset wrist feature point and the target finger feature point are on the line, except for the target finger feature point When the pixel points are not all noise points, the median depth of all normal points on the line connecting the preset wrist feature points and the target finger feature points is calculated as the depth of the preset wrist feature points. In this way, when the preset wrist feature points are noisy points, by finding the median depth of the normal points to replace the preset wrist feature points depth, more stable and accurate depth information of the hand feature points can be obtained, thereby improving the The depth camera measures the accuracy of the depth information of the feature points of the hand.

Description of the drawings

FIG. 1 is a schematic flowchart of an embodiment of a method for determining a gesture depth according to the present invention;

2 is a schematic diagram of preset hand feature points included in a hand depth image according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of modules of an embodiment of a gesture depth determining device of the present invention;

Fig. 4 is a schematic structural diagram of a gesture depth determination device provided by an embodiment of the present invention.

Detailed ways

The embodiment of the present invention provides a method, device, device and storage medium for determining the depth of a gesture. When the preset wrist feature point is a noisy point, the depth of the preset wrist feature point can be obtained by finding the median depth of the normal point instead of the preset wrist feature point. To obtain more stable and accurate depth information of the hand feature points, thereby improving the accuracy of measuring the depth information of the hand feature points based on the depth camera.

The terms "first", "second", "third", "fourth", etc. (if any) in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, without having to use To describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in a sequence other than the content illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those clearly listed. Steps or units, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products, or equipment.

For ease of understanding, the specific process of the embodiment of the method for determining the depth of the gesture of the present invention will be described below.

Referring to Fig. 1, Fig. 1 is a schematic flowchart of an embodiment of a method for determining a gesture depth according to the present invention. The method includes:

Step 101: Collect a hand depth image through a depth camera, and obtain preset wrist feature points included in the hand depth image and depth values of the preset wrist feature points;

It can be understood that the execution subject of the present invention may be a gesture depth determining device, and may also be a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example.

In this embodiment, the server is in communication connection with a depth camera. The depth camera has a depth image shooting function, and its specific model can be flexibly selected. For example, it can be kinectV1 (first generation kinect) or kinectV2 (second generation kinect).

First, the server collects a user's hand depth image through a depth camera, and obtains preset finger feature points contained in the hand depth image. Referring to FIG. 2, FIG. 2 is a schematic diagram of preset hand feature points included in a hand depth image according to an embodiment of the present invention. The hand depth image includes 21 preset hand feature points, specifically including 20 on the fingers The preset finger feature points: thumb T0, T1, T2, T3, index finger I0, I1, I2, I3, middle finger M0, M1, M2, M3, ring finger R0, R1, R2, R3, little finger L0, L1, L2 L3, and a preset wrist feature point W located on the wrist.

The server obtains the depth value of the preset wrist feature point included in the collected hand depth image, and the depth value represents the distance of the hand feature point from the depth camera. Specifically, the server may first collect the hand RGB image and the hand depth image through the depth camera, and align the two images through the alignment algorithm to locate the position of the hand and the position of the feature points of the hand in the RGB image, and then The RGB image and the hand depth image are coordinate mapped to obtain the depth information of the entire hand contained in the hand depth image, and obtain the two-dimensional information of the wrist feature points and the finger feature points through the deep learning model, and at the same time obtain The depth information of the entire hand contained in the hand depth image, and then the depth value of the preset wrist feature point is obtained from the depth information of the entire hand.

Step 102: Determine whether the preset wrist feature point is a noise point according to the depth value;

This step 102 may specifically include: identifying non-noise points in the hand region in the hand depth image according to the working depth range of the depth camera, and obtaining the depth value of the non-noise points; calculating all the obtained non-noise points And calculate the absolute value of the difference between the depth value of the preset wrist feature point and the median value; determine the absolute value of the difference between the depth value of the preset wrist feature point and the median value Whether the value is greater than or equal to a preset threshold; when the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to the preset threshold, it is determined that the preset wrist feature point is a noise point.

Take the kinectV2 camera as an example. The working depth range of the kinectV2 camera is 500～4500mm. If the object is too close or too far away from the camera, the depth information measurement will be inaccurate. For this reason, the server can obtain the depth information of the hand area in the hand depth image. For the depth values of all pixels, the pixels with a depth value of 500 to 4500 mm are regarded as non-noise points, so as to realize the recognition of non-noise points in the hand area in the hand depth image.

After identifying the non-noise points of the hand region in the hand depth image, calculate the median value of the depth values of all non-noise points, that is, the median. Specifically, the server obtains the depth value of each non-noise point and arranges the depth values from small to large to obtain a series of numbers. When the number of items in the series is odd, the depth value in the middle position is the median. When the number of items in the sequence is even, the median is the average of the two depth values in the middle. After that, the server calculates the absolute value of the difference between the depth value of the preset wrist feature points contained in the hand depth image and the median value, and determines whether the absolute value is greater than or equal to the preset threshold, and if the absolute value is greater than or equal to the preset threshold If the threshold is set, the server determines that the preset wrist feature point is noisy. On the contrary, if the absolute value is less than the preset threshold, it means that the preset wrist feature point is non-noise. At this time, the depth value of the preset wrist feature point is not used. Do any processing.

It should be noted that the foregoing preset threshold may be determined according to the length of a human palm. For example, the length of the feature point W to M3 in FIG. 2 may be used as the preset threshold. Because the palm can be moved when taking a photo, if the arm is not moving, the maximum forward and backward distance of the palm is the depth threshold of the palm space, and points beyond the depth threshold are regarded as abnormal.

Step 103: When the preset wrist feature point is a noise point, and when the preset wrist feature point is a noise point, acquire the preset finger feature point closest to the preset wrist feature point as the target finger feature point;

In this step, when the server determines that the preset wrist feature point is a noise point, the preset finger feature point closest to the preset wrist feature point is acquired as the target finger feature point.

Step 104: Judge whether the feature point of the target finger is a noise point;

This step 104 may specifically include: obtaining the depth value of the feature point of the target finger, calculating the absolute value of the difference between the depth value of the feature point of the target finger and the median value; judging the depth value of the feature point of the target finger and Whether the absolute value of the difference between the median value is greater than or equal to the preset threshold; when the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold, it is determined The target finger feature point is a noise point; when the absolute value of the difference between the depth value of the target finger feature point and the median value is less than the preset threshold, it is determined that the target finger feature point is not a noise point.

When judging whether the target finger feature point is a noise point, the server first obtains the depth value of the target finger feature point, calculates the absolute value of the difference between the depth value of the target finger feature point and the above median value, and then determines whether the absolute value is greater than or Equal to the aforementioned preset threshold. If the absolute value is greater than or equal to the preset threshold, the server determines that the target finger feature point is a noise; otherwise, if the absolute value is less than the preset threshold, then it determines that the target finger feature point is not a noise.

Step 105: When the target finger feature point is not a noise point, determine whether the pixel points other than the target finger feature point on the line connecting the preset wrist feature point and the target finger feature point are all noise points ；

In this step, when the target hand feature point is not a noise point, the server further determines whether the pixel points other than the target hand feature point are noise points on the connection line between the preset wrist feature point and the target hand feature point , The specific judgment method can refer to the above method of judging whether the target hand feature point is a noise point, which will not be repeated here.

Step 106: When the pixel points other than the target finger feature points are not all noise points on the line connecting the preset wrist feature points and the target finger feature points, calculate the preset wrist feature points and all the points. The median depth of all normal points on the line connecting the feature points of the target finger is used as the depth of the preset wrist feature points.

In this step, when the preset wrist feature point and the target hand feature point are on the line, and the pixel points other than the target hand feature point are not all noise points, the server obtains all normal points on the line (ie The median value of the depth value of the non-noise point, that is, the median depth, the median depth is taken as the depth of the preset wrist feature point.

Take the wrist feature point W in Figure 2 as an example. For the wrist feature point W, when it is determined to be a noise point, obtain the finger feature point T0 closest to W, use T0 as the target finger feature point and determine whether T0 is a noise point. If T0 is not a noise, then determine whether the pixels on the line between W and T0 are all noise, if not, then calculate the median depth of all normal points on the line between W and T0, and use the median depth As the depth of the characteristic point W of the wrist.

In this embodiment, through the above method, when the preset wrist feature points are noisy points, by finding the median depth of the normal points to replace the preset wrist feature points depth, more stable and accurate hand feature point depth information can be obtained. Thereby, the accuracy of measuring the depth information of the feature points of the hand based on the depth camera is improved.

Further, based on the first embodiment of the gesture depth determination method of the present invention, a second embodiment of the gesture depth determination method of the present invention is proposed.

In this embodiment, after the above step 104, it may further include: when the target finger feature point is a noisy point, acquiring the depth of the normal point closest to the preset wrist feature point, and setting it to the preset wrist feature point The depth of the nearest normal point is used as the depth of the preset wrist feature point.

Specifically, when the target hand feature point is a noisy point, the server may search for the normal point closest to the preset wrist feature point on the finger where the preset wrist feature point is located, centering on the preset wrist feature point, that is, For pixel points whose depth value is within a preset range, then the depth of the normal point is used as the depth of the preset wrist feature point.

Taking the wrist feature point W in FIG. 2 as an example, for the wrist feature point W, when it is determined to be a noisy point, the normal point closest to W is obtained, and the depth of the normal point can be used as the depth of the wrist feature point W.

Further, after the above step 105, it may further include: when the pixel points other than the target finger feature point are noise points on the line connecting the preset wrist feature point and the target finger feature point, obtaining the distance The depth of the normal point closest to the preset wrist feature point is the depth of the normal point closest to the preset wrist feature point as the depth of the preset wrist feature point.

Specifically, when the preset wrist feature point and the target hand feature point are on the line, and the pixel points other than the target hand feature point are all noise points, the server may place the finger on the finger where the preset wrist feature point is located. The preset wrist feature point is the center, and the normal point closest to the preset wrist feature point is retrieved, that is, the pixel point whose depth value is within the preset range, and then the depth of the normal point is taken as the depth of the preset wrist feature point .

Take the wrist feature point W in Figure 2 as an example. For the wrist feature point W, when it is determined to be a noise point, obtain the finger feature point T0 closest to W, use T0 as the target finger feature point and determine whether T0 is a noise point. If T0 is not a noise point, the normal point closest to W can be retrieved with W as the center, and the depth of the normal point can be used as the depth of the wrist feature point W.

In this embodiment, when the target hand feature point is a noise point, or when the preset wrist feature point and the target hand feature point are connected to the target hand feature point, the pixel points other than the target hand feature point are all noise points. The depth of the normal point closest to the preset wrist feature point is used as the depth of the preset wrist feature point, which further improves the accuracy of measuring the depth information of the hand feature point based on the depth camera.

The embodiment of the present invention also provides a gesture depth determination device.

Referring to FIG. 3, FIG. 3 is a schematic diagram of modules of an embodiment of a gesture depth determining device of the present invention. In this embodiment, the device for determining the gesture depth includes:

The acquisition module 301 is configured to acquire a hand depth image through a depth camera, and acquire preset wrist feature points contained in the hand depth image and the depth value of the preset wrist feature points;

The first determining module 302 is configured to determine whether the preset wrist feature point is a noise point according to the depth value;

The acquiring module 303 is configured to acquire the preset finger feature point closest to the preset wrist feature point as the target finger feature point when the preset wrist feature point is a noise point;

The second judgment module 304 is configured to judge whether the feature point of the target finger is a noise point;

The third judging module 305 is used for judging pixels other than the target finger feature points on the line connecting the preset wrist feature points and the target finger feature points when the target finger feature points are not noise points Whether the points are all noise points;

The calculation module 306 is configured to calculate the preset wrist feature when the pixel points other than the target finger feature point are not all noise points on the connection line between the preset wrist feature point and the target finger feature point The median depth of all normal points on the line connecting the point and the feature point of the target finger is taken as the depth of the preset wrist feature point.

Optionally, the first judgment module 302 is further configured to:

Optionally, the second judgment module 304 is further configured to:

Optionally, the device for determining the gesture depth further includes:

The first processing module is configured to obtain the depth of the normal point closest to the preset wrist feature point when the target finger feature point is a noise point, and use the depth of the normal point closest to the preset wrist feature point as The depth of the preset wrist feature point.

Optionally, the device for determining the gesture depth further includes:

The second processing module is configured to obtain a distance from the preset wrist when the pixel points other than the target finger feature points are noise points on the connection line between the preset wrist feature point and the target finger feature point. The depth of the normal point closest to the feature point is the depth of the normal point closest to the preset wrist feature point as the depth of the preset wrist feature point.

The functional implementation and beneficial effects of each module in the aforementioned gesture depth determination device correspond to the steps in the aforementioned gesture depth determination method embodiment, and will not be repeated here.

The device for determining the gesture depth in the embodiment of the present invention is described in detail above from the perspective of a modular functional entity, and the device for determining the gesture depth in the embodiment of the present invention is described in detail below from the perspective of hardware processing.

Referring to FIG. 4, FIG. 4 is a schematic structural diagram of a gesture depth determination device provided by an embodiment of the present invention. The gesture depth determination device 400 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 410 (for example, one or more processors) and a memory 420. Or more than one storage medium 430 (for example, one or one storage device with a large amount of storage) storing application programs 433 or data 432. Among them, the memory 420 and the storage medium 430 may be short-term storage or persistent storage. The program stored in the storage medium 430 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the device 400 for determining the gesture depth. Furthermore, the processor 410 may be configured to communicate with the storage medium 430, and execute a series of instruction operations in the storage medium 430 on the gesture depth determining device 400.

The gesture depth determination device 400 may also include one or more power supplies 440, one or more wired or wireless network interfaces 450, one or more input and output interfaces 460, and/or one or more operating systems 431, such as Windows Serve , Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that the structure of the gesture depth determination device shown in FIG. 4 does not constitute a limitation on the gesture depth determination device, and may include more or less components than shown in the figure, or combine certain components, or different The layout of the components.

The present invention also provides a storage medium. The storage medium may be a non-volatile storage medium or a volatile storage medium. The storage medium stores a gesture depth determination program, and the gesture depth determination program is processed. When the device is executed, the steps of the method for determining the depth of the gesture as described above are implemented.

The method and beneficial effects achieved when the gesture depth determination program running on the processor is executed can refer to the various embodiments of the gesture depth determination method of the present invention, which will not be repeated here.

Those skilled in the art can understand that if the above integrated modules or units are implemented in the form of software functional units and sold or used as independent products, they can be stored in a storage medium. Based on this understanding, the technical solution of the present invention essentially or the part that contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium. , Including several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disks or optical disks and other media that can store program codes. .

The above embodiments are only used to illustrate the technical solutions of the present invention, not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, a person of ordinary skill in the art should understand that: The recorded technical solutions are modified, or some of the technical features are equivalently replaced; these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

A method for determining the depth of a gesture, characterized in that the method for determining the depth of a gesture includes the following steps:

Acquiring a hand depth image by a depth camera, and acquiring preset wrist feature points included in the hand depth image and depth values of the preset wrist feature points;

Judging whether the preset wrist feature point is a noise point according to the depth value;

When the preset wrist feature point is a noise point, acquiring the preset finger feature point closest to the preset wrist feature point as the target finger feature point;

Judging whether the feature point of the target finger is a noise point;

When the target finger feature point is not a noise point, determining whether pixels other than the target finger feature point are all noise points on the line connecting the preset wrist feature point and the target finger feature point;

When the pixel points other than the target finger feature points on the line connecting the preset wrist feature points and the target finger feature points are not all noise points, the preset wrist feature points and the target finger are calculated The median depth of all normal points on the line of the characteristic points is taken as the depth of the preset wrist characteristic points.
The method for determining the depth of a gesture according to claim 1, wherein the step of judging whether the preset wrist feature point is a noise according to the depth value comprises:

Identifying the non-noise points of the hand region in the hand depth image according to the working depth range of the depth camera, and obtaining the depth value of the non-noise points;

Calculating the median value of the acquired depth values of all non-noise points, and calculating the absolute value of the difference between the depth value of the preset wrist feature point and the median value;

Judging whether the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold;

When the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold value, it is determined that the preset wrist feature point is a noise point.
The method for determining the depth of a gesture according to claim 2, wherein the step of judging whether the feature point of the target finger is a noise point comprises:

Acquiring the depth value of the feature point of the target finger, and calculating the absolute value of the difference between the depth value of the feature point of the target finger and the median value;

Judging whether the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold value, determining that the target finger feature point is a noise point;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is less than the preset threshold value, it is determined that the target finger feature point is not a noise point.
The method for determining the depth of a gesture according to any one of claims 1 to 3, wherein after the step of determining whether the feature point of the target finger is a noise point, the method further comprises:

When the target finger feature point is a noise point, obtain the depth of the normal point closest to the preset wrist feature point, and use the depth of the normal point closest to the preset wrist feature point as the preset wrist feature point depth.
The method for determining the depth of a gesture according to any one of claims 1 to 3, wherein the line connecting the predetermined wrist feature point and the target finger feature point is determined, except for the target finger feature point. After the steps of whether the pixels other than the dots are all noises, it also includes:

When on the line connecting the preset wrist feature point and the target finger feature point, pixel points other than the target finger feature point are noise points, obtain the normal point closest to the preset wrist feature point Depth, the depth of the normal point closest to the preset wrist feature point is taken as the depth of the preset wrist feature point.
A gesture depth determining device, characterized in that the gesture depth determining device includes:

An acquisition module, configured to acquire a hand depth image through a depth camera, and acquire preset wrist feature points contained in the hand depth image and the depth value of the preset wrist feature points;

The first judgment module is configured to judge whether the preset wrist feature point is a noise point according to the depth value;

An acquiring module, configured to acquire a preset finger feature point closest to the preset wrist feature point as the target finger feature point when the preset wrist feature point is a noise point;

The second judgment module is used to judge whether the feature point of the target finger is a noise point;

The third judgment module is used for judging the pixel points other than the target finger feature points on the line between the preset wrist feature points and the target finger feature points when the target finger feature points are not noise points Are they all noisy;

A calculation module for calculating the preset wrist feature point when the pixel points other than the target finger feature point on the line connecting the preset wrist feature point and the target finger feature point are not all noise points The median depth of all normal points on the line with the feature point of the target finger is taken as the depth of the preset wrist feature point.
7. The gesture depth determining device according to claim 6, wherein the first judgment module is further configured to:

Identifying the non-noise points of the hand region in the hand depth image according to the working depth range of the depth camera, and obtaining the depth value of the non-noise points;

Calculating the median value of the acquired depth values of all non-noise points, and calculating the absolute value of the difference between the depth value of the preset wrist feature point and the median value;

Judging whether the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold;

When the absolute value of the difference between the depth value of the preset wrist feature point and the median value is greater than or equal to a preset threshold value, it is determined that the preset wrist feature point is a noise point.
8. The gesture depth determining device according to claim 7, wherein the second judgment module is further configured to:

Acquiring the depth value of the feature point of the target finger, and calculating the absolute value of the difference between the depth value of the feature point of the target finger and the median value;

Judging whether the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is greater than or equal to the preset threshold value, determining that the target finger feature point is a noise point;

When the absolute value of the difference between the depth value of the target finger feature point and the median value is less than the preset threshold value, it is determined that the target finger feature point is not a noise point.
A gesture depth determination device, wherein the gesture depth determination device includes a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor are interconnected by wires;

The at least one processor invokes the instructions in the memory, so that the gesture depth determination device executes the gesture depth determination method according to any one of claims 1-7.
A storage medium with a computer program stored on the storage medium, wherein the computer program implements the gesture depth determination method according to any one of claims 1-7 when the computer program is executed by a processor.