CN116863076A - Optimization method for obtaining three-dimensional information of sport ball - Google Patents
Optimization method for obtaining three-dimensional information of sport ball Download PDFInfo
- Publication number
- CN116863076A CN116863076A CN202310853826.1A CN202310853826A CN116863076A CN 116863076 A CN116863076 A CN 116863076A CN 202310853826 A CN202310853826 A CN 202310853826A CN 116863076 A CN116863076 A CN 116863076A
- Authority
- CN
- China
- Prior art keywords
- player
- dimensional
- calculating
- sphere
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000005457 optimization Methods 0.000 title claims abstract description 11
- 210000004556 brain Anatomy 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 206010034701 Peroneal nerve palsy Diseases 0.000 claims description 3
- 238000010586 diagram Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 3
- 230000000007 visual effect Effects 0.000 abstract description 3
- 210000001364 upper extremity Anatomy 0.000 description 21
- 238000005516 engineering process Methods 0.000 description 8
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000005452 bending Methods 0.000 description 2
- 230000000903 blocking effect Effects 0.000 description 2
- 210000000988 bone and bone Anatomy 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 210000002310 elbow joint Anatomy 0.000 description 2
- 210000001503 joint Anatomy 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000452 restraining effect Effects 0.000 description 2
- 210000000323 shoulder joint Anatomy 0.000 description 2
- 210000003857 wrist joint Anatomy 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 210000000245 forearm Anatomy 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0487—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Software Systems (AREA)
- Geometry (AREA)
- Architecture (AREA)
- Computer Hardware Design (AREA)
- Human Computer Interaction (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention discloses an optimization method for acquiring three-dimensional information of a sport ball. Comprising the following steps: calibrating a camera, acquiring three-dimensional coordinates of the position of the player, acquiring two-dimensional image coordinates of the sphere B, drawing an auxiliary cross line, estimating the exact position of an intersection point by changing the auxiliary cross line according to the context information, and displaying the distance between the position of the player and the sphere, namely d_layer; the height of the striking point and the distance from the striking point to the net are displayed, and the three-dimensional coordinates of the sphere are calculated. The d_layer guiding method provided by the invention can effectively reduce the error of the three-dimensional reconstruction result, and provides visual guidance for operators in monocular three-dimensional reconstruction according to the known information obtained by the computer vision method; the method is equivalent to a solution obtained by computer vision calculation, and the next solving action is constrained, so that larger errors exceeding a normal range can be avoided from being generated by operators, and the accuracy is further improved when the operators perceive the batting gesture of the badminton player from the two-dimensional image.
Description
Technical Field
The invention belongs to the technical field of computer vision and image processing, in particular to the technical field of analysis and image processing of sports competition technology, and particularly relates to a method for acquiring information such as three-dimensional tracks of balls which move in a parabolic manner only under the action of gravity and air resistance of the earth, such as table tennis, volleyball, basketball, football and the like.
Background
The three-dimensional information of a moving object is acquired from two-dimensional videos shot by a single camera, which is an important basic scientific problem of computer vision: monocular vision three-dimensional reconstruction (Monocular 3drecon construction).
A hybrid intelligent (Human-machine Hybrid Intelligence) method combining the monocular three-dimensional reconstruction capability of Human brain with an artificial intelligent technology breaks through the technical bottleneck of acquiring three-dimensional information of a moving object from game videos such as shuttlecocks shot by a single camera, and is not limited by various objections of data acquisition on a game site.
The Chinese patent 201410785238x relates to a method for acquiring three-dimensional information of a sport ball, which comprises video image acquisition, three-dimensional model construction and three-dimensional information acquisition. The method can acquire the speed, arc and drop point information of the ball from a single video stream, can acquire the three-dimensional track of the ball, and can further calculate technical and tactical parameters such as the ball striking speed of the ball, the bending radian of the three-dimensional track, the net passing height, the flight distance, the ball drop point and the like. By analyzing these technical and tactical information, coaches and athletes can be helped to improve the technical and tactical level, especially the winning ability of high-level athletes in ball games. The method not only can use videos shot by a training field, but also can use mass televisions on the Internet to rebroadcast the videos. The three-dimensional information acquired by the method cannot be obtained through naked human eyes or through the existing computer vision algorithm, and has unique value.
However, the above hybrid intelligent method has a large three-dimensional reconstruction error. The hybrid intelligent method is a combination of human brain intelligence and artificial intelligence, and requires operators to participate in a three-dimensional reconstruction process, particularly to be familiar with corresponding movement modes, and the operators can perceive the batting gesture of the athlete in the movement modes according to two-dimensional images in the brain aiming at the familiar movement modes, so as to guide the artificial intelligence (computer vision) method to finish the final three-dimensional reconstruction. Thus, the monocular three-dimensional reconstruction capability of the operator determines the accuracy of the three-dimensional reconstruction. Although the ability of three-dimensional reconstruction of the human brain can approach the optimal solution indefinitely in theory, recent practice has shown that the responsibility and ability of the operator is divergent. The three-dimensional reconstruction acquisition is carried out on the same batting of the same match, and the maximum error of operators can reach 1 meter. Although we adopt corresponding management means, such as emphasizing the accuracy and importance of data acquisition, the problem of large three-dimensional reconstruction error still cannot be fundamentally solved.
Disclosure of Invention
The invention discloses an optimization method for acquiring three-dimensional information of a sport ball body according to the defects of the prior art. The invention aims to solve the problem of providing a method capable of avoiding larger errors exceeding a normal range generated by an operator, helping the operator to further improve three-dimensional reconstruction precision when perceiving a motion gesture from a two-dimensional image and obtaining ball three-dimensional track information by using a camera single-machine video signal more accurately.
The invention is realized by the following technical scheme:
an optimization method for acquiring three-dimensional information of a moving sphere comprises the following steps:
step 1, calibrating a camera: constructing a corresponding relation between an image point and a model point, and calculating a projection matrix P;
step 2, acquiring three-dimensional coordinates of a position A of a player: clicking the position A of the player in the two-dimensional image by using a mouse to obtain the screen coordinate q of the player in the image; according to the projection matrix P and the screen coordinate q of the mouse cursor in the image, calculating the intersection point of the light corresponding to the mouse cursor and the ground; calculating a cross line which passes through the intersection point and is parallel to the ground, and drawing the cross line in a two-dimensional image; on the ground, calculating a two-dimensional projection straight line of a three-dimensional straight line corresponding to the screen coordinate q on the ground, calculating an intersection point of the two-dimensional projection straight line and an auxiliary cross line, and displaying a three-dimensional block diagram of a player; the operator clicks a mouse to determine the position A of the player;
step 3, acquiring two-dimensional image coordinates of the sphere B: clicking a sphere center point in the two-dimensional image by using a mouse to obtain a screen coordinate p of a sphere B in the image; calculating a straight line L corresponding to the screen coordinate p; calculating a straight line L corresponding to the two-dimensional image screen coordinates P of the sphere B according to the projection matrix P, and then calculating a projection straight line LG of the straight line L on the ground;
step 4, drawing auxiliary cross lines: according to the coordinates of the mouse cursor in the image, calculating the intersection point of the light corresponding to the mouse cursor and the ground; calculating a cross line which passes through the intersection point and is parallel to the ground, and drawing the cross line in a two-dimensional image; calculating an intersection point G of the auxiliary cross line and the projection straight line LG, wherein the intersection point G is a projection point of the sphere B on the ground because LG is a projection straight line of L on the ground;
step 5, the human brain estimates the exact position of the intersection point G by changing the auxiliary cross line according to the context information: when the auxiliary cross line is displayed, the distance between the position A of the player and the ball B, namely d_player, is displayed; the height z of the striking point B and the distance d_net of the striking point B to the net are displayed.
Step 6, calculating three-dimensional coordinates of the sphere B: an intersection point of the straight line GB which takes the foot drop G as a starting point and is vertically upward and the ray L is calculated, and the intersection point is the three-dimensional coordinate of the sphere B.
In step 2 of the method of the present invention, the screen coordinates q of the player position a in the image select the neck position using the upper end of the torso.
In step 5 of the method of the present invention, the distance between the player position a and the ball B, i.e. d_player, should satisfy the following mean and standard deviation according to the height of the player:
in step 5 of the method of the present invention, the distance between the player position a and the ball B, i.e., d_player, should satisfy the following average and standard deviation according to the striking mode:
the d_player is the Euclidean distance (distance from player to impact point) between the three-dimensional coordinates of sphere B and the three-dimensional coordinates of player position A.
The upper limb of the human body consists of a sports apparatus (such as a badminton racket), two rigid bones (such as a big arm and a small arm) and three joints (such as a shoulder joint, an elbow joint and a wrist joint). Therefore, the maximum extension length that the upper limb can reach is "big arm length+small arm length+badminton racket length". If the operator sees that the upper limb is fully extended at the moment of the player's impact in the two-dimensional image, it can be inferred that d_player is approaching the maximum of the effective range. Conversely, if the operator sees that the upper limb is fully curled at the moment of the player's impact in the two-dimensional image, it can be inferred that d_player is approaching the minimum of the effective range.
That is, when the operator "estimates the exact position of the intersection point G by changing the auxiliary cross line based on the context information", if the value of d_layer is found to exceed the maximum value or to be smaller than the minimum value, it is highly likely that an error occurs to the operator. Therefore, the introduction of the d_layer can help operators to correct errors, improve the estimation accuracy and finally improve the accuracy of the three-dimensional reconstruction method.
Furthermore, the effective range of d_player is quite different for different striking techniques. When killing balls, the upper limbs are fully stretched to form whipping ball striking actions. At this time, d_layer is close to the maximum value of the effective range. When a ball is served, the upper limb bends, and the action of shooting force is hardly caused. At this time, d_layer is close to the minimum value of the effective range. Therefore, the operator is familiar with the typical d_layer value of different batting technologies (different attitudes), which is equivalent to the prior knowledge of d_layer, so that the estimation accuracy can be further improved, and the error of three-dimensional reconstruction can be reduced.
The specific practice can be to use the value of d_player when the operator actually measures the upper arm and forearm lengths and hits the ball to establish the perception of the effective range.
The method of the invention takes the distance d_layer from the three-dimensional coordinates of the sphere B to the three-dimensional coordinates of the player position A as a guide. It treats the upper limb of the human body as a whole and considers that there is an effective range of maximum extension length and minimum extension length of the upper limb. And determining: the maximum extension length is determined by the physical length of the player's upper limb and the racket, and the minimum extension length is determined by the batting posture of the badminton player.
Counting the effective range of the d_layer can guide the operator to make more reasonable depth estimation. The method of the invention avoids the batting data beyond the effective range and can avoid the occurrence of larger error in three-dimensional reconstruction. Meanwhile, aiming at athletes with different heights and different batting technologies, operators can obtain more accurate three-dimensional reconstruction results after being familiar with the effective range of the d_player.
The d_layer guiding method provided by the invention is an optimization of a hybrid intelligent three-dimensional reconstruction method, and can effectively reduce errors of three-dimensional reconstruction results.
The d_layer method provides visual guidance for operators in monocular three-dimensional reconstruction according to known information obtained by a computer vision method. The method is equivalent to a solution obtained by computer vision calculation, and is used for restraining the next solving action of an operator. These new constraints not only can avoid the operator from generating larger errors beyond the normal range, but can also help the operator to further improve accuracy in perceiving the batting posture of the badminton player from the two-dimensional image.
Drawings
FIG. 1 is a camera calibration of step 1 of the method of the present invention;
FIG. 2 is a three-dimensional coordinate of the player position A obtained in step 2 of the method of the present invention;
FIG. 3 is a straight line corresponding to the player position A in step 3 of the method of the present invention;
FIG. 4 is a method step 4 of the present invention for drawing auxiliary cross-hairs and displaying auxiliary information;
FIG. 5 is a view of the method of the present invention with the selection of a reasonable striking point under the direction of d_player;
FIG. 6 is a graph showing the coordinates of a striking point calculated by the method of the present invention.
Detailed Description
The present invention is specifically described below by way of examples, which are only for further illustration of the present invention, but are not to be construed as limiting the scope of the present invention, and some insubstantial modifications and adaptations of the present invention by those skilled in the art based on the foregoing teachings are intended to be within the scope of the present invention.
The method is an optimization of a method for acquiring three-dimensional information of the sport ball in the Chinese patent 201410785238 x. The patent method in which the human brain estimates the exact position of G by changing the way of the auxiliary cross-hair based on the context information is a major source of reconstruction errors.
The method of the invention proposes a d_layer boot method. The d_layer guiding method provides visual guiding for operators in monocular three-dimensional reconstruction according to known information obtained by a computer vision method. The method is equivalent to a solution obtained by computer vision calculation, and is used for restraining the next solving action of an operator. These new constraints not only can avoid the operator from generating larger errors beyond the normal range, but can also help the operator to further improve accuracy in perceiving the batting posture of the badminton player from the two-dimensional image.
d_layer boot method: d_player is the Euclidean distance (distance from player to impact point) of the three-dimensional coordinates of sphere B and player position A.
The example takes a badminton as an example.
The upper limb of the human body consists of a sports apparatus (badminton racket), two rigid bones (big arm and small arm) and three joints (shoulder joint, elbow joint and wrist joint). Therefore, the maximum extension length that the upper limb can reach is "big arm length+small arm length+badminton racket length".
If the operator sees that the upper limb is fully extended at the moment of the player's impact in the two-dimensional image, it can be inferred that d_player is approaching the maximum of the effective range. Conversely, if the operator sees that the upper limb is fully curled at the moment of the player's impact in the two-dimensional image, it can be inferred that d_player is approaching the minimum of the effective range.
When the operator estimates the exact position of G by changing the auxiliary cross-hair based on the context information, it is highly likely that an error occurs in the operator if the value of d_layer is found to exceed the maximum value or to be less than the minimum value. Therefore, the introduction of the d_layer can help operators to correct errors, improve the estimation accuracy and finally improve the accuracy of the three-dimensional reconstruction method.
The effective range of d_player is quite different for different striking techniques. When killing balls, the upper limbs are fully stretched to form whipping ball striking actions. At this time, d_layer is close to the maximum value of the effective range. When a ball is served, the upper limb bends, and the action of shooting force is hardly caused. At this time, d_layer is close to the minimum value of the effective range. Therefore, the operator is familiar with the typical d_layer value of different batting technologies (different attitudes), which is equivalent to the prior knowledge of d_layer, so that the estimation accuracy can be further improved, and the error of three-dimensional reconstruction can be reduced.
With reference to the attached drawings, the method for acquiring the three-dimensional information of the moving sphere by using d_layer guidance comprises the following specific steps:
step 1: and (5) calibrating a camera. And constructing a corresponding relation between the image points and the model points, and calculating a projection matrix P. Camera calibration is shown in fig. 1.
Step 2: and acquiring the three-dimensional coordinates of the position A of the player. Clicking the position A of the player in the two-dimensional image with a mouse obtains the screen coordinates q of the player in the image. And then, calculating the intersection point of the light corresponding to the mouse cursor and the ground according to the projection matrix P and the screen coordinate q of the mouse cursor in the image. A cross line is calculated that passes through the intersection point and is parallel to the ground, and the cross line is drawn in a two-dimensional image. On the ground, calculating a two-dimensional projection straight line of the three-dimensional straight line corresponding to q on the ground, calculating an intersection point of the two-dimensional projection straight line and an auxiliary cross line, and displaying a three-dimensional block diagram of a player. The operator mouse click determines the player position a.
The player position A in the two-dimensional image is clicked by a mouse, and the neck position at the upper end of the trunk is selected in the example.
As shown in fig. 2, the player position a is acquired.
Step 3: two-dimensional image coordinates of the sphere B are acquired. Clicking the center point of the sphere in the two-dimensional image by using a mouse to obtain the screen coordinate p of the sphere B in the image. And calculating a straight line L corresponding to p. And calculating a straight line L corresponding to the two-dimensional image coordinate P of the sphere B according to the projection matrix P, and then calculating a projection straight line LG of the straight line L on the ground.
As shown in fig. 3, the player position a corresponds to a straight line.
Step 4: and drawing an auxiliary cross line. And calculating the intersection point of the light rays corresponding to the mouse cursor and the ground according to the coordinates of the mouse cursor in the image. A cross line is calculated that passes through the intersection point and is parallel to the ground, and the cross line is drawn in a two-dimensional image. An intersection point G of the auxiliary reticle and the projection straight line LG is calculated. Since LG is a projected straight line of L on the ground, G is a projected point of sphere B on the ground.
As shown in fig. 4, an auxiliary cross line is drawn and auxiliary information is displayed.
Step 5: the human brain estimates the exact position of the intersection point G by changing the way of the auxiliary cross-hair based on the context information.
When the auxiliary cross line is displayed, the distance between the player position a and the sphere B, i.e. d_player, is displayed. The height z of the striking point B and the distance d_net of the striking point B to the net are displayed. These auxiliary information help to improve accuracy.
As shown in fig. 5, a reasonable striking point is selected under the guidance of d_player.
Step 6: the three-dimensional coordinates of sphere B are calculated. An intersection point of the straight line GB and the ray L which are vertically upward and take the foot drop G as a starting point is calculated. This intersection is the three-dimensional coordinates of sphere B.
As shown in fig. 6, the coordinates of the striking point are calculated.
The method of the invention regards the upper limb of the human body as a whole and designs a new constraint condition d_layer. The method guides (constrains) the human brain to make more accurate estimation by displaying the d_layer value, thereby improving the measurement accuracy of the hybrid intelligent method.
The invention further carries out constraint definition on the effective range of the statistical d_layer by collecting the existing data.
The d_layer guiding method of the invention regards the upper limb of the human body as a whole, the d_layer has an effective range, and the accurate value of the effective range is obtained from the competition data in a statistics way, thereby avoiding errors exceeding the effective range to the maximum extent, further reducing errors and improving precision.
For example, athletes of different heights may have different upper limb lengths. This results in a difference in the effective range of d_layer. For example, if a danish player is 1.95 meters high and a japan player is 1.74 meters high, their d_player effective ranges must be different. The acquisition accuracy can be improved if the operator is familiar with the effective range of d_layer under different height conditions.
The 123335 beats of badminton batting data are collected, and the mean value and standard deviation of d_player of players with different heights are obtained. The higher the height of the athlete, the greater the d_layer mean value, as shown in the following table. Table 1 mean and standard deviation of d_player for athletes of different heights.
In addition, the player adopts different techniques to hit the ball, and the bending degree of the upper limbs is also different. Table 2 below shows that, when a badminton player kills a ball, the upper limb is extended to the maximum extent, d_layer=1.23±0.16, in order to pursue the height of the striking point. When a badminton player plays a badminton, the player does not play a badminton, so that the upper limb extends to the minimum extent, and d_player=0.69+/-0.22. The effective range of d_player under different batting technical conditions is familiar, and the acquisition accuracy can be improved. Table 2 means and standard deviations of d_player for different striking techniques.
For example, when a player of 1.8 m height kills a ball, the upper arm extension reaches the maximum, the maximum value of d_player obtained according to the mean value and standard deviation of d_player of players of different heights in table 1 is 1.01+2×0.26=1.53 m, and the maximum value of d_player obtained according to the mean value and standard deviation of d_player of different batting technologies in table 2 is 1.23+2×0.16=1.55 m. The minimum of the two constraints is 1.53 meters. If the acquisition personnel determines the three-dimensional coordinates of the sphere B in step 6, d_layer exceeds the theoretical upper limit of 1.53 meters, the invention gives a warning to the acquisition personnel. The acquisition personnel further corrects the step 5 according to the warning, corrects the error just, and can improve the precision of the three-dimensional reconstruction method.
Second, when a player of 1.8 m height kills a ball, the player's batting gesture is passive due to control by an opponent, and the upper arm extension may be below a theoretical minimum. The minimum value of d_layer, i.e. 1.01-2 x 0.26=0.91 meters, and the minimum value of 1.23-2 x 0.16=0.49 meters, is calculated according to tables 1 and 2, 0.49 meters. The minimum of the two constraints is 0.49 meters. When d_layer is below the theoretical lower limit of 0.49 meters, the invention alerts (rather than warns) the acquisition personnel. And (6) facing the reminding, the acquisition personnel can choose to ignore the reminding, and the step (6) is directly completed. Therefore, the reminding does not interfere with the collection efficiency of the collection personnel.
Meanwhile, when the player with the height of 1.8 meters stops the ball, the collector selects the wrong technology (such as killing the ball). The theoretical lower limit of the blocking ball is 0.37 m, the theoretical upper limit of the blocking ball is 1.37 m, the theoretical lower limit of the killing ball is 0.49 m, and the theoretical upper limit of the killing ball is 1.53 m. If d_player=0.4 meters currently, because the collector selects the wrong technique (e.g., killing), 0.4 meters is less than the theoretical lower limit of killing by 0.49 meters, triggering the lower limit alert of "killing technique", the collector can quickly realize that he or she has selected the wrong striking technique, and then correct the striking technique to bring d_player=0.4 meters to a reasonable range.
In summary, the above method performs rigid constraint on the behavior of the collection personnel according to the statistical data of tables 1 and 2, that is, the effective ranges of the two constraints. The constraint can not only enable the monocular three-dimensional reconstruction result to be in an effective error range, but also avoid the error of the batting technology.
Claims (4)
1. An optimization method for acquiring three-dimensional information of a moving sphere is characterized by comprising the following steps:
step 1, calibrating a camera: constructing a corresponding relation between an image point and a model point, and calculating a projection matrix P;
step 2, acquiring three-dimensional coordinates of a position A of a player: clicking the position A of the player in the two-dimensional image by using a mouse to obtain the screen coordinate q of the player in the image; according to the projection matrix P and the screen coordinate q of the mouse cursor in the image, calculating the intersection point of the light corresponding to the mouse cursor and the ground; calculating a cross line which passes through the intersection point and is parallel to the ground, and drawing the cross line in a two-dimensional image; on the ground, calculating a two-dimensional projection straight line of a three-dimensional straight line corresponding to a coordinate screen coordinate q on the ground, calculating an intersection point of the two-dimensional projection straight line and an auxiliary cross line, and displaying a three-dimensional block diagram of a player; the operator clicks a mouse to determine the position A of the player;
step 3, acquiring two-dimensional image coordinates of the sphere B: clicking a sphere center point in the two-dimensional image by using a mouse to obtain a screen coordinate p of a sphere B in the image; calculating a straight line L corresponding to the coordinate p; calculating a straight line L corresponding to the two-dimensional image screen coordinates P of the sphere B according to the projection matrix P, and then calculating a projection straight line LG of the straight line L on the ground;
step 4, drawing auxiliary cross lines: according to the coordinates of the mouse cursor in the image, calculating the intersection point of the light corresponding to the mouse cursor and the ground; calculating a cross line which passes through the intersection point and is parallel to the ground, and drawing the cross line in a two-dimensional image; calculating an intersection point G of the auxiliary cross line and the projection straight line LG, wherein the intersection point G is a projection point of the sphere B on the ground because LG is a projection straight line of L on the ground;
step 5, the human brain estimates the exact position of the intersection point G by changing the auxiliary cross line according to the context information: when the auxiliary cross line is displayed, the distance between the position A of the player and the ball B, namely d_player, is displayed; displaying the height z of the striking point B and the distance d_net from the striking point B to the net;
step 6, calculating three-dimensional coordinates of the sphere B: an intersection point of the straight line GB which takes the foot drop G as a starting point and is vertically upward and the ray L is calculated, and the intersection point is the three-dimensional coordinate of the sphere B.
2. The optimization method for acquiring three-dimensional information of a sport ball according to claim 1, wherein: in the step 2, the screen coordinates q of the player position a in the image select the neck position using the upper end of the trunk.
3. The optimization method for acquiring three-dimensional information of a sport ball according to claim 1, wherein: in the step 5, the distance between the player position a and the ball B, i.e., d_player, should satisfy the following average and standard deviation according to the height of the player:
4. the optimization method for acquiring three-dimensional information of a sport ball according to claim 1, wherein: in the step 5, the distance between the player position a and the ball B, i.e., d_player, should satisfy the following average and standard deviation according to the batting mode:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310853826.1A CN116863076A (en) | 2023-07-12 | 2023-07-12 | Optimization method for obtaining three-dimensional information of sport ball |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310853826.1A CN116863076A (en) | 2023-07-12 | 2023-07-12 | Optimization method for obtaining three-dimensional information of sport ball |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116863076A true CN116863076A (en) | 2023-10-10 |
Family
ID=88218572
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310853826.1A Pending CN116863076A (en) | 2023-07-12 | 2023-07-12 | Optimization method for obtaining three-dimensional information of sport ball |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116863076A (en) |
-
2023
- 2023-07-12 CN CN202310853826.1A patent/CN116863076A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11836929B2 (en) | Systems and methods for determining trajectories of basketball shots for display | |
| US11263461B2 (en) | Systems and methods for monitoring objects at sporting events | |
| US10019806B2 (en) | Determining x,y,z,t biomechanics of moving actor with multiple cameras | |
| US6098458A (en) | Testing and training system for assessing movement and agility skills without a confining field | |
| US11798318B2 (en) | Detection of kinetic events and mechanical variables from uncalibrated video | |
| US20050223799A1 (en) | System and method for motion capture and analysis | |
| Kelly et al. | A virtual coaching environment for improving golf swing technique | |
| WO2017062428A1 (en) | Systems and methods for monitoring objects in athletic playing spaces | |
| AU2016372567A1 (en) | Method and device for analysis of movement of a piece of sports equipment | |
| CN109069903B (en) | System and method for monitoring objects in a sporting event | |
| CN104504694A (en) | Method for acquiring three-dimensional information of moving ball | |
| WO2011009302A1 (en) | Method for identifying actions of human body based on multiple trace points | |
| US20080136907A1 (en) | Method and measuring device for motional performance | |
| CN110270077A (en) | A kind of batting assessment system and its method | |
| CN105879349A (en) | Method and system for displaying golf ball falling position on putting green on display screen | |
| CN103404122A (en) | Motion-vector-based comparison of moving objects | |
| CN116863076A (en) | Optimization method for obtaining three-dimensional information of sport ball | |
| CN115414648B (en) | Football evaluation method and football evaluation system based on motion capture technology | |
| CN115414646B (en) | Basketball auxiliary training method and device | |
| TWI782649B (en) | Badminton smash measurement system and method | |
| TWI805124B (en) | Virtual reality system for baseball pitcher fatigue analysis and sport injury diagnosis | |
| Curtis et al. | Cricket batting technique analyser/trainer using fuzzy logic | |
| Sujae et al. | Technology enhanced teaching and coaching of complex sport skills–An example of the acro-volley (sepaktakraw) power smash (kuda) and normal relay (sila) serve techniques | |
| Kumar | Automated decisions in aware cricket ground: a proposal | |
| Yao et al. | Research on Position Optimization of Badminton Doubles Based on Wireless Network Sensor and Human-computer Interactive Training |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |