JP2014235743A - Method and equipment for determining position of hand on the basis of depth image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and equipment for determining a position of a hand on the basis of a depth image.SOLUTION: A method of determining a position of a hand according to the present invention includes the steps of determining the main direction of a candidate area of a hand acquired on the basis of a depth image, searching for tip edge points of the hand along the main direction, and determining position points of the hand on the basis of the respective tip edge points. The method and equipment for determining a position of a hand enable the accurate determination of a position of a hand without being affected by light beam conditions and the shape of a hand.

Description

  The present invention relates generally to hand tracking and position determination, and more specifically to a hand position determination method and equipment based on a depth map.

  Hand tracking and positioning is a very important and key part of the hand motion identification system. For convenience of user operation, hand tracking and position determination should not require wearing special equipment such as special gloves or color markings. Other than that, the hand is a non-rigid object, fast moving, easily deformed, and has features such as shielding itself, so tracking and positioning of the hand is a very difficult task.

  Previously, many researchers used hand color features as hand features to track and determine positions, but this method is easily affected by light conditions. Considering that depth images are much less affected by light rays, more and more researchers are now tracking and locating hands based on depth images. However, since there is very little information on the depth image, it is a problem to be solved whether effective information can be extracted from any depth image and hand tracking and position determination can be performed. Some researchers use a depth threshold to divide a candidate hand area and check whether the area is a hand through shape features, edge features, or histogram statistical features. However, if the position determination of the hand candidate area is inaccurate or out of the range of the candidate area, the position of the hand cannot be determined by this method.

  U.S. Pat. No. 7372977b2 discloses a technique for real-time visual tracking using a depth camera. In this technology, first, the tracking target edge information is obtained from the depth image, then the tracking target is displayed using a mathematical contour expression model, and finally the tracking target edge information and the contour model are matched based on the depth data. Find the tracking position point. In the present invention, since hand tracking is mainly performed using contour information, considerable strictness is required for the hand contour.

  The paper "3D hand tracking using kalman filter in depth space" proposed a method for 3D hand tracking in real time using motion history image information. In this method, hand candidate areas are detected by motion clustering based on image information output from a 3D depth sensor, and then hand tracking and position determination are realized by a kalman filter. This technique can determine the position of the hand relatively well when the hand is moving, and can track the hand. However, when the hand is stationary, the tracking is likely to occur due to the influence of noise.

  An object of the present invention is to provide a hand position determination method and equipment based on a depth map.

  In an embodiment of the present invention, the main direction of the candidate area of the hand acquired based on the depth image is determined, the leading edge point of the hand is searched along the main direction, and based on each leading edge point A method for determining the position of a hand based on a depth image is provided, including determining a position point of the hand.

  In another embodiment of the present invention, a main direction determination device used to determine a main direction of a candidate area of a hand predicted based on a depth image, for searching for a leading edge point of the hand along the main direction. Provided is a hand position determination facility based on a depth image, including an edge point search device to be used, and a position point determination device to be used for determining a hand position point based on each of the leading edge points.

  The hand position determination technique based on the depth image according to the embodiment of the present invention can accurately determine the position of the hand and is not affected by the light beam condition or the shape of the hand.

6 is a scene conceptually showing hand tracking and position determination technology to which an embodiment of the present invention is applied. It is a flowchart of the position determination method of the hand based on the depth image of this invention Example. The candidate area block and the candidate area are determined based on the depth image. It is a conceptual diagram of the main direction of the candidate area of a hand. It is a flowchart which searches the front-end | tip edge point of the hand along the main direction of the Example of this invention. FIG. 6 is an example process for searching for a tip edge point of a hand along a main direction of an embodiment of the present invention. FIG. It is a flowchart which determines the position point of a hand based on each front-end | tip edge point of the Example of this invention. The functional arrangement block diagram of the position determination equipment based on the depth image of the Example of this invention is shown. 1 shows an overall hardware block diagram of a hand position determination system based on a depth image of an embodiment of the present invention. FIG.

  In order that those skilled in the art can better understand the present invention, the present invention will be described in detail below in combination with the drawings and embodiments for carrying out the invention.

  FIG. 1 is a scene conceptually showing a hand position determination technique to which an embodiment of the present invention is applied. As shown in FIG. 1, the user stands within the shooting range of a stereo camera such as a twin-lens camera, and the user is shot with this stereo camera. When the user moves his / her hand within the range of the camera, for example, the computer processing device determines the position information of the user's hand based on the depth image acquired based on the image captured by the stereo camera, and performs tracking of the hand Can do. For example, the black dots in FIG. 1 indicate the movement trajectory of the hand.

  As is well known to those skilled in the art, a depth image is an image in which the value of each pixel in the image indicates the distance between a certain point in the scene and the camera. Unlike a gray scale image, a depth image includes depth (distance) information of an object, and is therefore suitable for various applications that require stereoscopic information.

  FIG. 2 is a flowchart of the hand position determination method based on the depth image of the embodiment of the present invention.

  As shown in FIG. 2, in step S210, the main direction of the hand candidate area acquired based on the depth image is determined.

  As described above, obtaining a candidate area of a hand based on a depth image has already been a lot of research in the field, which is not a key point of the present invention, and those skilled in the art will know what existing suitable schemes. The hand candidate area can be acquired from the depth image. Here, in order to complete the description, a brief description of the acquisition method used in the present embodiment will be given.

  First, the candidate position point of the hand is predicted based on the history information.

  Since the hand movement is continuous, in this step, the position where the hand may appear now is predicted based on the previous movement state and the movement coefficient of the hand. The hand position point is a point used to represent the position of the hand, and may be, for example, the center point of the hand or the center of gravity point.

This step may be accomplished using any suitable point tracking method, such as a kalman filter or particle filter. In local range, hand movement approximates linear motion, so the simplest prediction method is to predict where a hand will appear in the current frame based on the hand position point and hand motion speed in the previous frame By doing so, the following formula is obtained.

In this, (S _{x, t-1} , S _{y, t-1} ) is the coordinates of the hand position point in the previous frame (t-1 time), and (S _{x, t} , S _{y, t} ) Is the coordinates of the candidate position point of the hand in the current frame (time t), Δt is the time interval between two consecutive frames, and the motion speed (V _{x, t-1} , V _{y, t-1} ) Can be calculated by the following equation.

In this, (S _{x, t-2} , S _{y, t-2} ) is the coordinates of the hand position point in the previous frame (time t-2).

  After acquiring the hand candidate position point, the candidate area block is determined based on the candidate position point and the depth value information history.

The candidate area block can have any shape, and the candidate area block is shown by taking the rectangle in FIG. 3 (a) as an example. The size of the candidate area block is not fixed and has adaptability, and its purpose is to include the detected hand as much as possible, but this candidate area block should not be too large. For a candidate area block whose size is fixed, such a flexible change in size is effective in reducing the computational complexity. Specifically, in the process of collecting images, near objects are large and distant objects are small, so when the hand approaches the camera, the area of the hand in the collected image becomes relatively large, and the candidate area block However, when the hand moves away from the camera, the area of the hand in the collected image becomes relatively small, and the candidate area block becomes relatively small accordingly. As an example, in the present invention, a candidate area block is represented using a rectangle, but if the radius of the candidate area block is determined by the following equation, it becomes half the side length of the rectangle.

In this, HandRadius is the radius of the candidate area block, α is a constant, and its size is the actual length of the corresponding object to be shot based on the fixed parameters of the camera used and the length of the pixels in the image To confirm. HisDepth Value represents the reference depth value of the candidate position point of the hand in the current frame, and can be used as the HisDepth Value by using the average depth value of the hand position point in the previous consecutive frames. It is shown as an expression.

In this, DepthValue _i is the actual depth value of the hand position point in the i-th frame, and n is the number of frames.

  After the candidate area block is determined, the candidate area is divided from among the candidate area blocks.

  Since the candidate area block includes not only the hand but also other image contents in the depth image, the candidate area of the hand must be divided from that. This division processing can be performed through various existing area division methods. For example, there are region growth, threshold division, and the like, but detailed description will not be given here. FIG. 3B shows a candidate area for divided hands.

  As mentioned above, the example of the method of acquiring the candidate area of a hand easily based on the depth image was introduced. As previously mentioned, one of ordinary skill in the art can obtain a hand candidate area from a depth image by any existing suitable method. Hereinafter, a process for determining the main direction of the hand candidate area will be described based on an embodiment of the present invention.

  The main direction is a linear direction in which the longest line segment that can be acquired in the image area exists. FIGS. 4A and 4B are conceptual diagrams of the main direction of the hand candidate area (the direction of the straight line indicated by the symbol A) in the embodiment of the present invention. The principal direction can be calculated by using various appropriate methods in this field. In this embodiment, the principal direction is calculated by using the principal component analysis method (PCA), and the direction vector of the straight line in the principal direction is calculated. The main direction is indicated by A (Ax, Ay), and Ay∈ [−1,1] and Ax∈ [−1,1]. Both the principal component analysis method and the direction vector of the straight line are known in the art, and will not be described in detail here.

  What can be selected is that the main direction calculated above can be adjusted to reduce the influence of noise. More specifically, because the hand movement is continuous, the main direction of the hand should not change too quickly in the two front and back frames, and if the change is too fast, the effects of noise often increase. Therefore, adjustment can be performed with respect to the main direction.

  Specifically, the difference value between the main direction to be calculated and the main direction of the hand of the image of the previous frame is determined, and if the difference value is larger than a predetermined threshold, the hand of the previous predetermined image is determined. The main direction to be calculated is adjusted based on the main direction.

For example, an example in which a predetermined threshold can be taken is 90 degrees. Of course, those skilled in the art may set other suitable angles based on experience and actual hand movement. If CurA (CurA.x, CurA.y) and LastA (LastA.x, LastA.y) represent the main direction of the current hand and the main direction of the previous frame, respectively, CurA (CurA.x, When the difference between CurA.y) and LastA (LastA.x, LastA.y) is greater than 90 degrees, CurA (CurA.x, CurA. Adjustments can be made to y). Taking the arithmetic mean as an example, the following formula can be used to calculate the arithmetic mean of CurA (CurA.x, CurA.y) and LastA (LastA.x, LastA.y) and use that as the current main direction: it can.

  As can be seen, an arithmetic average value of the main direction of the previous predetermined frame (for example, m frames before) and the current main direction can be calculated to obtain the current main direction.

  It should be understood that the main direction adjustment process is optional and not essential.

  Returning to FIG. 2, in step S220, the hand tip edge point is searched for along the main direction.

  The hand edge point can be found relatively easily along the main direction, even if the hand candidate area is inaccurate (for example, the hand is not in the hand candidate area). The tip edge point can be found. Hereinafter, the process of step S220 will be described in detail with reference to FIGS. 5 and 6A to 6D.

  As shown in FIG. 5, in step S2201, a first straight line perpendicular to the main direction is determined and shown as symbol B in FIG. 6 (a). The first straight line preferably passes through the candidate position point of the hand presented above, thereby improving the calculation efficiency. As can be seen, the first straight line is perpendicular to the main direction is merely an example, and the first straight line may form another angle with the main direction.

  Thereafter, in step S2202, the leading edge point is found on each straight line of several straight lines extending along the main direction, starting from several predetermined points within a predetermined range of the first straight line.

  There are many leading edge points of the hand. For example, all leading edge points can be found on several straight lines extending along the main direction shown in FIG. The predetermined range can be set by the user based on experience, and the purpose thereof is to include as much as possible a straight line (for example, a straight line shown in FIG. 6B) on which the tip edge point can be found. Yes, the range is not too large. As one possible example, the predetermined range may be a portion located in the candidate area block on a first straight line. The some predetermined points may be set as necessary. For example, when relatively high accuracy is required for the position point of the hand to be finally determined, a relatively large number of predetermined points must be selected. For example, each pixel within a predetermined range of the first straight line is used as a starting point, and Find the tip edge point on each straight line starting from the start point and extending along the main direction, and if the accuracy requirement is relatively low, select a relatively small number of predetermined points, for example, within a certain range with a certain interval. A starting point can be selected from each pixel.

  On each of the straight lines extending in the main direction, a possible leading edge point can be selected by the inclination of the change in the pixel depth value on the straight line. Specifically, the change in depth of two adjacent pixels on the hand should not be large, and if the change in depth is quite large, one of the two pixels is on the hand but the other Can be recognized as not on the hand (for example, on the background image). For this reason, in this embodiment, for each of the straight lines along the main direction, the depth value difference between two adjacent pixels on the straight line is calculated sequentially from the start point, and the depth values of two adjacent pixels are calculated. If the difference is larger than a predetermined threshold value, a pixel closer to the start point of the two adjacent pixels is set as a leading edge point. Among these, the predetermined threshold can be set by a person skilled in the art based on experience.

Hereinafter, this will be described with reference to FIG. For example, assuming that M ₀ is the starting point for straight line C, the depth value difference between adjacent pixels M ₁ and M ₀ is calculated first. If the difference is smaller than a predetermined threshold, this pixel M _{1 is} still Because it is considered to be above, it is necessary to continue the search. Thereafter, the depth value difference between the pixels M ₂ and M ₁ is calculated, and if the difference is less than a predetermined threshold, the pixel M _{2 is} still considered to be on the hand and the search needs to be continued. The search is continued according to this method until a pair of adjacent pixels M _i−1 and M _i whose depth value difference is larger than the threshold is found. At this time, the pixel M _i-1 is located on the hand and the pixel M _i is not on the hand, that is, the pixel M _i-1 near the start point M ₀ is the tip edge point of the hand. It becomes. For each straight line, by performing a search on a straight line starting from a predetermined point on the first straight line and extending along the main direction, all the points that can be the tip edge point can be found, and (d in FIG. 6 ).

The selection can be made by sieving each point that can be the leading edge point to reduce the influence of noise. More specifically, because of the influence of noise, some of the leading edge points found by the above processing are not genuine leading edge points, and therefore can be removed by the following equation.

  In this, edgepoint represents the tip edge point that has been found but has not been sifted, centerpoint and handRadius represent the position point of the hand tracked in the previous frame and the radius of the candidate area block, and dis ( edgepoint, centerpoint) represents the spatial distance between edgepoint and centerpoint, and a and b are both constants, and those skilled in the art can set by experience based on the ratio of human limbs. When the spatial distance between the found leading edge point and the position point of the hand of the previous frame does not satisfy the formula (6), the leading edge point can be removed as a noise edge point.

  It should be understood that the tip edge point sieving process is optional and not essential.

  Returning to FIG. 2 again, in step S230, the position point of the hand is determined based on each tip edge point. Hereinafter, the process of step S230 will be described with reference to FIG.

As shown in FIG. 7, in step S2301, the average value position point of each tip edge point is calculated. Specifically, the average value position point of each tip edge point is calculated by various methods such as an arithmetic average value, a square root average value, a geometric average value, and a weighted average value. For example, taking the arithmetic average value as an example, the average value position point of the leading edge point can be calculated by the following equation.

Where n is the number of tip edge points, edgepoint _i .x is the X-axis coordinate of the _i-th tip edge point, edgepoint _i .y is the y-axis coordinate of the i-th tip edge point, (Cx , Cy) is the coordinates of the average position point.

  Thereafter, in step S2302, a hand position point is determined based on the average value position point and the main direction described above.

After determining the average position point, the hand position point can be calculated by various suitable methods. In consideration of the fact that the connecting line from the hand position point to the average value position point should be approximately along the main direction of the hand, in this embodiment, the average value position point is the starting point and the opposite of the main direction. A hand position point is determined along the direction, as shown in the following equation.

  Where (handPx, handPy) is the coordinates of the current hand position point, handRadius is the radius of the candidate area block, (Ax, Ay) is the main direction of the hand, c is a constant, It can be set by experience based on the ratio.

  Thereby, the current position of the hand is determined. It is also possible to detect the accuracy of the position using information such as the history of the movement trajectory of the hand, the area size, and the shape. When determining that the current hand position is accurate, at least one of the current movement speed of the hand, the current movement acceleration of the hand, the area of the hand, the depth value of the position point, etc. One can be calculated (the method of calculation is well known in the art and does not tamper with the luxury here) and can be used to determine the position of the hand in the frame that follows to achieve hand tracking can do.

  The hand position determination method based on the depth image according to the embodiment of the present invention has been described above. In this method, the leading edge point of the hand is found using the main direction, and the actual position of the hand is determined based on the leading edge point. According to this method, even if the hand is not in the predicted hand candidate area (for example, the hand candidate area is mistakenly set on the arm), the hand edge point can be found, and thus good adaptability can be obtained. Have. In addition, since the method does not use hand shape information, the method is suitable for various hand shapes and satisfies the feature of the hand that there are many morphological changes. In addition, since the position is determined by the depth map, the method is less affected by light rays.

  It is worth pointing out that the method for determining the position of the hand based on the depth image according to the embodiment of the present invention can also be used for detecting the position of the finger, especially when the fingers are separated from each other. An effect is obtained. Specifically, when the fingers are separated from each other, the application of steps S210 and S220 of the method for determining the position of the hand according to the embodiment of the present invention may cause a continuous contact point (for example, a finger) Branch point) can be detected. Subsequently, an appropriate threshold value can be set, and the continuous contacts of adjacent fingers can be screened. Thereafter, an average value position point is calculated for several tip edge points of each finger, and this finger position can be obtained.

  Hereinafter, the hand position determining equipment based on the depth image according to the embodiment of the present invention will be described with reference to FIG.

  FIG. 8 shows a functional arrangement block diagram of the hand position fixing equipment 800 based on the depth image according to the embodiment of the present invention.

  As shown in FIG. 8, the hand position determination device 800 is a main direction determination device 810 used to determine the main direction of the candidate area of the hand predicted based on the depth image, the tip of the hand along the main direction. An edge point searching device 820 used for searching for an edge point, and a position point determining device 830 used for determining a hand position point based on each of the leading edge points are included.

  The hand position determination facility 800 further includes a pre-processing device (not shown) and a post-processing device (not shown), and the pre-processing device predicts a candidate position point of the hand based on the history information, The candidate area block is determined based on the hand candidate position point and the depth value information history, and is used to divide the candidate area from the candidate area block. The post-processing device is configured to determine at least one of the current movement speed of the hand, the current movement acceleration of the hand, the area of the hand, and the depth value of the position point based on the determined hand position point. Used.

  For specific functions and operations of the main direction determining device 810, the edge point searching device 820, the position point determining device 830, the pre-processing device, and the post-processing device, refer to the related description from FIG. 1 to FIG. I won't repeat the explanation here.

  Hereinafter, an overall hardware block diagram of a hand position determination system 900 based on a depth image according to an embodiment of the present invention will be described with reference to FIG. As shown in FIG. 9, the hand position determination system 900 inputs externally related images and information such as left and right images taken by the camera, camera parameters, depth images, and previous motion parameters of the hand. For example, an input device 910 corresponding to a keyboard, a mouse, a camera, etc .; used for the implementation of the hand position determination method based on the depth image according to the above-described embodiment of the present invention or the above-described hand position determination apparatus; Central processing unit or other IC chip with processing capability, etc., applicable processing equipment 920; obtained in the process of performing the above-mentioned hand position determination, such as the determined position point coordinates, hand movement trajectory, etc. Used to output the results, eg output equipment 930 applicable to monitors, printers, etc .; and related to the above-mentioned hand position determination process in a volatile and non-volatile manner, eg depth images, Used to store data such as movement history information, current position point of hand, current movement speed, depth value of current position point, eg random access memory (RAM), read only memory (ROM), hard disk or A storage facility 940 corresponding to various volatile and non-volatile memories such as a semiconductor memory is included.

  While the basic principles of the present invention have been described in connection with specific embodiments, it should be pointed out that those skilled in the art will understand all or any of the steps and components of the method and apparatus of the present invention. Can be implemented in any computer (including processors, memory media, etc.) or computer network by hardware, firmware, software, or combinations thereof, which will be described by those skilled in the art to explain the present invention. In the situation you read, you can do it using their basic computer programming skills.

  For this reason, the object of the present invention can be realized by using one or a set of computer programs on any computer. The above-mentioned computer may be a known publicly known device. For this reason, the object of the present invention is realized only by providing a program product including a program code for realizing the above-described method or apparatus. That is, such a program product also constitutes the present invention, and a storage medium having such a program product also constitutes the present invention. Of course, the storage medium may be any known storage medium or any storage medium developed in the future.

  It should be further pointed out that, in the apparatus and method of the present invention, it is understood that each member or each step can be disassembled and / or newly combined. These decompositions and / or new combinations should be regarded as a method having an effect equivalent to that of the present invention. In addition, the above-described series of processing steps can be performed in chronological order according to the description order, but is not necessarily performed in chronological order. Certain steps can be performed in parallel or independently of each other.

The form for carrying out the above-mentioned invention does not become the restriction | limiting of the protection scope of this invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives may occur depending on design requirements and other factors. Any amendments, equivalent replacements and improvements in the spirit and principle of the present invention shall all be included in the protection scope of the present invention.

Claims (18)

A hand position determination method based on a depth image,
Determine the main direction of the candidate area of the hand acquired based on the depth image,
A method for determining a hand position, comprising: searching for a tip edge point of the hand along the main direction; and determining the hand position point based on each tip edge point. Predicting the candidate position point of the hand based on history information;
The hand position determining method according to claim 1, further comprising: determining a candidate area block based on the candidate position point and depth value information history; and dividing the candidate area from the candidate area block. Determining the main direction of the candidate area of the hand acquired based on the depth image is
Calculate the main direction of the candidate area of the hand,
A difference value between the main direction and the main direction of the hand of the image of the previous frame is determined, and when the difference value is larger than a predetermined threshold, the main of the hand of the image of the previous frame is determined. 3. The method of determining a position of a hand according to claim 1, comprising adjusting the main direction of the candidate area based on a direction. Looking for the tip edge point of the hand along the main direction,
A first straight line perpendicular to the main direction is determined, and on each straight line of several straight lines extending along the main direction, starting from several predetermined points within a predetermined range of the first straight line, respectively. The method of claim 2, comprising finding the tip edge point.   5. The hand position determining method according to claim 4, wherein the first straight line passes through the candidate position point, and the predetermined range is a portion located in the candidate area block on the first straight line. Looking for the tip edge point along the main direction is
A depth value difference between two adjacent pixels on the straight line is calculated from the start point, and a pixel closer to the start point among two adjacent pixels having the depth value difference larger than a predetermined threshold is defined as a leading edge point. The method for determining the position of a hand according to claim 4 or 5, comprising: Looking for the tip edge point along the main direction is
A spatial distance between the front edge point and the position point of the image of the previous frame is calculated, and when the spatial distance exceeds a predetermined threshold range, the front edge point is set as a noise edge point. The method for determining a position of a hand according to claim 6, further comprising: Determining the hand position point based on each tip edge point,
3. The hand according to claim 1, further comprising: calculating an average value position point of each tip edge point; and determining the hand position point based on the average value position point and the main direction. Position determination method.   The method further comprises determining at least one of a current movement speed of the hand, a current movement acceleration of the hand, an area of the hand, and a depth value of the position point based on the hand position point. Item 1. The method for determining the position of the hand according to Item 1. A hand positioning device based on depth images,
A main direction determining device used to determine the main direction of the candidate area of the hand predicted based on the depth image;
An edge point search device used for searching for the leading edge point of the hand along the main direction, and a position point determination device used for determining the position point of the hand based on each leading edge point; Location fixing equipment. Predicting the candidate position point of the hand based on history information;
11. The hand position determining facility according to claim 10, further comprising a pre-processing device that determines a candidate area block based on the candidate position point and depth value information history, and divides the candidate area from the candidate area block. The main direction determining device is:
Calculate the main direction of the candidate area of the hand,
A difference value between the main direction and the main direction of the hand of the image of the previous frame is determined, and when the difference value is larger than a predetermined threshold, the main of the hand of the image of the previous frame is determined. 12. The hand positioning apparatus according to claim 10 or 11, wherein adjustment is made to a main direction of the candidate area based on a direction. The edge point search device comprises:
A first straight line perpendicular to the main direction is determined, and on each straight line of several straight lines extending along the main direction, starting from several predetermined points within a predetermined range of the first straight line, respectively. 12. The hand location facility of claim 11, comprising finding the tip edge point.   14. The hand position determining facility according to claim 13, wherein the first straight line passes through the candidate position point, and the predetermined range is a portion located in the candidate area block on the first straight line. The edge point search device comprises:
A depth value difference between two adjacent pixels on the straight line is calculated from the start point, and a pixel closer to the start point among two adjacent pixels having the depth value difference larger than a predetermined threshold is defined as a leading edge point. 15. The hand positioning facility according to claim 13 or 14, comprising: The edge point search device comprises:
A spatial distance between the front edge point and the position point of the image of the previous frame is calculated, and when the spatial distance exceeds a predetermined threshold range, the front edge point is set as a noise edge point. 16. The hand location facility of claim 15 comprising: The position point determination device includes:
12. The hand according to claim 10 or 11, comprising calculating an average value position point of each tip edge point, and determining the hand position point based on the average value position point and the main direction. Positioning equipment. And a post-processing device for determining at least one of the current movement speed of the hand, the current movement acceleration of the hand, the area of the hand, and the depth value of the position point based on the position point of the hand. The hand positioning apparatus according to claim 10.

Images