WO2023130696A1 - 一种自动跟踪拍摄体育比赛的智能系统及其控制方法 - Google Patents

一种自动跟踪拍摄体育比赛的智能系统及其控制方法 Download PDF

Info

Publication number
WO2023130696A1
WO2023130696A1 PCT/CN2022/103544 CN2022103544W WO2023130696A1 WO 2023130696 A1 WO2023130696 A1 WO 2023130696A1 CN 2022103544 W CN2022103544 W CN 2022103544W WO 2023130696 A1 WO2023130696 A1 WO 2023130696A1
Authority
WO
WIPO (PCT)
Prior art keywords
player
players
tracking
control method
pan
Prior art date
Application number
PCT/CN2022/103544
Other languages
English (en)
French (fr)
Inventor
谈科峰
严新宇
Original Assignee
深眸科技(北京)有限公司
深眸远智(苏州)信息科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深眸科技(北京)有限公司, 深眸远智(苏州)信息科技有限公司 filed Critical 深眸科技(北京)有限公司
Publication of WO2023130696A1 publication Critical patent/WO2023130696A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/695—Control of camera direction for changing a field of view, e.g. pan, tilt or based on tracking of objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G—PHYSICS
    • G08—SIGNALLING
    • G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
    • G08C17/00—Arrangements for transmitting signals characterised by the use of a wireless electrical link
    • G08C17/02—Arrangements for transmitting signals characterised by the use of a wireless electrical link using a radio link
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/02—Constructional features of telephone sets
    • H04M1/04—Supports for telephone transmitters or receivers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
    • H04M1/724—User interfaces specially adapted for cordless or mobile telephones
    • H04M1/72403—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality
    • H04M1/72406—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality by software upgrading or downloading
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
    • H04M1/724—User interfaces specially adapted for cordless or mobile telephones
    • H04M1/72403—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality
    • H04M1/72409—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality by interfacing with external accessories
    • H04M1/72412—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality by interfacing with external accessories using two-way short-range wireless interfaces
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
    • H04M1/724—User interfaces specially adapted for cordless or mobile telephones
    • H04M1/72403—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality
    • H04M1/72409—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality by interfacing with external accessories
    • H04M1/72415—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality by interfacing with external accessories for remote control of appliances
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/66—Remote control of cameras or camera parts, e.g. by remote control devices
    • H04N23/661—Transmitting camera control signals through networks, e.g. control via the Internet
    • G—PHYSICS
    • G08—SIGNALLING
    • G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
    • G08C2201/00—Transmission systems of control signals via wireless link
    • G08C2201/90—Additional features
    • G08C2201/93—Remote control using other portable devices, e.g. mobile phone, PDA, laptop

Definitions

  • the invention relates to the technical field of intelligent shooting, in particular to an intelligent system for automatically tracking and shooting sports games and a control method thereof.
  • the prior art provides a camera technology for football detection and tracking, for example, Chinese Patent Document 1: CN106488127B, which discloses a camera switching control method and device based on football detection and tracking.
  • This Chinese patent is a camera switching control method and device based on football detection and tracking.
  • the panoramic camera can be used to continue shooting when the football is blocked by the player; it can also intercept the lens from the panoramic camera based on the number of moving targets in the current video content, and zoom the lens to the output standard picture for shooting; and it can be triggered
  • the camera is preset at the position, and when the football position coincides with the trigger position, the camera corresponding to the trigger position is switched to shoot.
  • This patent can switch different camera lenses according to the position of the football, complete the game video shooting on the football field under the unattended situation, and can better meet the needs of the fans when watching the game.
  • this prior art first needs panoramic camera to shoot, obtains the video information of athletes and players on the field and after the detection processing of a plurality of control modules, is switched to corresponding camera by switching module and shoots. Therefore, this prior art still needs a dedicated imaging device, and the system configuration of the device is complicated, and the system cost is still high.
  • Chinese patent document 2 CN105245845A and Chinese patent document 3: CN105208349A; Both disclose a kind of method that the scene of a game controls the camera to follow up automatically according to the number of people on the field; Number of people control The camera automatically follows and shoots, recognizes the number of people on the field and automatically controls the camera.
  • the above-mentioned patents require the cooperation of far-infrared low-light high-speed high-definition cameras, the main server in the central management computer room, tracking units, and high-definition zoomable 360-degree rotating cameras. This technology still cannot rely on the simplification of the system architecture, and the system layout cost still cannot be reduced.
  • the existing smart tracking and shooting applications based on the mobile phone platform are relatively simple, and no other patent documents have been found to record the application of using the mobile phone platform to track and shoot team sports events. Specifically speaking, for the sports event scene, the intelligent tracking technology of the existing mobile phone platform still has the following deficiencies:
  • the present invention proposes an intelligent system and its control method for automatically tracking and shooting sports games.
  • the intelligent system proposed by the present invention can be realized by using only one pan-tilt and an intelligent user terminal, without deploying a panoramic camera as in Patent Document 1, and without switching to a specific shooting camera.
  • the invention only needs a specially designed pan-tilt to cooperate with a user's smart hand-held terminal to realize automatic tracking and shooting of sports games.
  • the pan-tilt of the present invention is a device that can turn the platform horizontally and vertically.
  • the cloud platform is provided with Bluetooth module, WIFI or other communication modules that can be connected to the intelligent terminal, and the user can select the connection mode between the handheld intelligent terminal and the cloud platform according to specific needs.
  • the gimbal is matched and connected with the Bluetooth of the smart terminal through the Bluetooth module.
  • the WIFI module of the smart terminal can also be connected to the WIFI module of other devices, or through the cellular network (3G/4G/5G, etc.), the automatic tracking and shooting of sports games can be transmitted to the other devices in real time for live broadcast or storage.
  • the pan/tilt is automatically controlled by the application program APP running on the smart terminal.
  • the core of this system also includes an application program APP developed on the smart terminal to realize the real-time control of the pan/tilt by the smart terminal.
  • the application program APP of the smart terminal transmits the tracking decision to the gimbal through a communication link such as Bluetooth or WiFi, and the gimbal drives the smart terminal to follow the target to rotate horizontally and vertically in a controlled state, thereby realizing intelligent tracking and shooting of sports games .
  • the definition of smart terminal should generally include high-definition wide-angle lens, AI processing module, CPU/GPU processor, communication module and other related modules.
  • the tracking decision strategy proposed by the present invention is based on a very simple observation of a ball game. That is, in team ball games, the movement direction of most players is centered on the position of the ball. Especially the players who are closer to the ball run and fight more aggressively. Therefore, the tracking decision-making strategy of the present invention firstly judges the movement trend on the field by analyzing the movement direction and movement track of the overall players. After that, the speed and movement of individual players are analyzed, and one or several key players are found based on this analysis, and they are locked and tracked.
  • the goal of the present invention is to align the center of the smart terminal lens with the key player, so the present invention determines the coordinates of the center point of the key player based on the coordinates of the key player on the screen of the smart terminal. Tracking of key players. In addition, the above-mentioned tracking process needs to be repeated periodically, such as repeating judgment and finding key players every 2-3 seconds, and locking and tracking them in the next cycle.
  • an intelligent system for automatically tracking and shooting sports games proposed by the present invention includes a cloud platform and a smart phone.
  • the smart phone includes a first communication module
  • the cloud platform includes a second communication module.
  • the first communication module of the smart phone and the second communication module of the cloud platform are Bluetooth modules, WIFI modules or other communication modules connected to the smart phone.
  • the communication connection mode between the cloud platform and the smart phone is carried out by Bluetooth module, WIFI or other means capable of realizing communication connection.
  • the user can choose the connection method between the smart phone and the cloud platform according to the specific needs.
  • the gimbal is automatically controlled by the application program APP running on the smart phone.
  • the communication between the smart phone and the pan/tilt is bidirectional, while the control of the smart phone to the pan/tilt is unidirectional, that is, the smart phone can only issue control commands to the pan/tilt in one direction to control the action of the pan/tilt.
  • the pan-tilt for automatic tracking and shooting of the present invention includes a mobile phone clip, which is used for physical connection with the smart phone body.
  • the cloud platform also includes a vertical rotation motor that can rotate in the vertical direction, and a horizontal rotation motor that can rotate in the horizontal direction; the mutual cooperation of the vertical rotation motor and the horizontal rotation motor can realize the full-angle and full-angle rotation of the smartphone body.
  • the cloud platform also includes a cloud platform handle and a base, and a button control area of the cloud platform is also set above the cloud platform handle and the base.
  • the present invention also proposes a control method for an intelligent system that automatically tracks and shoots sports games, comprising the following steps:
  • the smartphone shoots the players in the shot, and after recognizing the number of players in the shot, performs action recognition on the joint nodes of each player.
  • the system classifies according to the action types determined in S103, and reserves the required action types, and filters the unnecessary action types, and the players whose action types are reserved are used as candidate key players.
  • the smart phone sends a control command to control the pan/tilt to automatically track the locked key players through Bluetooth or WIFI.
  • the error between the target center point and the smartphone lens or screen center point is continuously calculated to obtain the error_value value; the error_value value presets a range, as long as the calculated error_value value is in this range The lens does not rotate within the specified interval; once the error_value value deviates from the preset range, the smartphone sends a command to control the rotation of the gimbal, so that the error_value falls within the preset range again.
  • the present invention realizes the tracking, photographing and live broadcast of the hot spots of the game by constantly selecting and locking key players on the field.
  • Fig. 1 is a control flow chart of an intelligent system for automatically tracking and shooting sports games and its control method according to the present invention.
  • Fig. 2 is a schematic front view of the connection structure between the smart phone and the pan-tilt in the intelligent system structure for automatic tracking and shooting of the present invention.
  • Fig. 3 is a schematic diagram of the back of the connection structure between the smart phone and the pan-tilt in the intelligent system structure for automatic tracking and shooting of the present invention.
  • Figure 4 and Figure 5 are schematic diagrams of joint node sequences.
  • Fig. 6 is a schematic diagram of time-varying joint nodes of an action in soccer.
  • Fig. 7 is a block diagram of an intelligent system for automatically tracking and shooting sports games proposed by the present invention.
  • the present invention proposes an intelligent system for automatically tracking and photographing sports games, including a pan-tilt and an intelligent terminal.
  • the smart terminal is, for example, a smart phone with a camera, a smart tablet computer or a smart camera.
  • the preferred smart terminal of the present invention is a smart phone with a camera.
  • FIG. 7 A structural block diagram of an intelligent system for automatically tracking and shooting sports games is shown in FIG. 7 , including a cloud platform 12 and a smart phone 11 .
  • the smart phone 11 includes a first communication module 111
  • the platform 12 includes a second communication module 121 .
  • the first communication module 111 of the smart phone 11 and the second communication module 121 of the cloud platform 12 are bluetooth modules, WIFI modules or other communication modules connected to the smart phone 11 .
  • the communication connection mode between the cloud platform 12 and the smart phone 11 is carried out by using a bluetooth module, WIFI or other means capable of realizing communication connection.
  • the user can choose the connection mode between the smart phone 11 and the cloud platform 12 according to specific needs.
  • the cloud platform 12 is matched and connected with the Bluetooth of the smartphone 11 through the Bluetooth module.
  • the WIFI module of the cloud platform 12 and the WIFI module of the smart phone 11 form a small local area network to connect.
  • the WIFI module can also be connected with the WIFI modules of other devices, so that the automatic tracking and shooting of sports games can be transmitted to the other devices in real time. Live or store.
  • the cloud platform 12 is automatically controlled by the application program APP running on the smart phone 11 .
  • the communication between the smart phone 11 and the cloud platform 12 is two-way, and the control of the smart phone 11 to the cloud platform 12 is unidirectional, that is, the smart phone 11 can only issue control commands to the cloud platform 12 in one direction. Control the motion of the cloud platform 12.
  • the physical connection method between the pan-tilt and the smart terminal is not particularly limited in this patent, as long as the smart terminal can be firmly installed on the pan-tilt, all fixing methods can be applied.
  • the smart terminal is clamped by the clamping device on the platform.
  • Fig. 2 and Fig. 3 have shown the connection structure of smart phone and pan-tilt in the smart system structure of automatic tracking and shooting of the present invention, including: pan-tilt and smart phone.
  • the cloud platform includes a mobile phone clip 1, which is used for physical connection with the smart phone body 6, that is, the smart phone body 6 is fixed on the cloud platform.
  • the cloud platform also includes a vertical rotation motor 2 that can rotate in a vertical direction, and a horizontal rotation motor 3 that can rotate in a horizontal direction; angle to rotate.
  • Cloud Terrace also comprises Cloud Terrace handle and base 4, and the button control area 5 of Cloud Terrace is also set on Cloud Terrace handle and base 4 top, when automatic tracking shooting sports game, user's hand holds Cloud Terrace handle and base 4, and passes The button control area 5 controls and operates the pan-tilt.
  • Fig. 1 is a specific control flow of an intelligent system for automatically tracking and shooting sports games proposed by the present invention.
  • the tracking decision-making strategy of the present invention firstly judges the movement trend on the field by analyzing the movement direction and movement track of the overall players. After that, the speed and movement of individual players are analyzed, and one or several key players are found based on this analysis, and they are locked. Therefore, the tracking decision-making strategy of the present invention first recognizes the player's action type, which includes: running, jumping, standing, etc.
  • the number of marked joint nodes can be set according to the specific competition item to be automatically tracked and photographed. For example, 14, 17 or 19 joint nodes are used. The more the number of joint nodes is set, the more delicate the tracking of the player's limbs and muscles will be, and the more accurate the judgment and tracking of the player's movements will be. However, the more the joint nodes are marked, the better, because the more the joint nodes are marked, the greater the calculation of the system for the analysis of joint node changes, which will put forward higher requirements for the data processing capability of the system.
  • the device used for system data calculation is the smart phone body 6.
  • the automatic tracking and shooting system proposed by the present invention can allow the user to select the brand and model of the corresponding smart phone body 6 in the APP installed on the smart phone body 6, so that the system can automatically match a reasonable number of joint nodes in specific applications.
  • the smartphone shoots the players in the shot, and after recognizing the number of players in the shot, performs action recognition on the joint nodes of each player.
  • S103 judge after recognizing the action of each player, and determine the action type of each player.
  • the judgment of the action type depends on the complexity of the action type to be judged, such as distinguishing some simple actions, such as standing still or running, which can be judged by using a static joint node graph.
  • Figure 4 and Figure 5 schematically give examples of labeling of static joint nodes.
  • Figure 4 and Figure 5 are marked with 14 joint nodes, including the player's head, neck, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left and right ankles.
  • the time series of joint nodes is used as input, ⁇ joint node @t1, joint node @t2, joint node @t3...joint node N@t N ⁇ (where N is a natural number, t N is the unit time), and the pre-customized
  • the trained neural network is used as a model, and the customized neural network judges the input joint node sequence to determine which action is running and which action is shooting.
  • FIG. 6 is a schematic diagram of the sequence of joint nodes changing with time when a player shoots a football. From left to right in Figure 6 are the sequence diagrams of the joint nodes at time t1-t5. According to the pre-customized model, the system can judge that the action is a football shooting action. The judgment of other actions is also similar to this.
  • the system classifies according to the action types judged in S103, and retains the required action types, and filters the unnecessary action types, and the players whose action types are reserved are used as candidate key players. For example, in a sports game, more attention is paid to players who are in a running state. Therefore, for players whose action type is identified as standing, the system will filter them, and keep players in a running state as candidate key players.
  • S105 Calculate the displacement of the candidate key player of the reserved action type within a unit time.
  • the displacement of the candidate key player is calculated within a unit time, such as 0.5 second, 1 second, 1.5 second, and 2 seconds.
  • a unit time such as 0.5 second, 1 second, 1.5 second, and 2 seconds.
  • the gimbal system makes judgments on the player's movement and displacement during the static period, and locks key players; then, during the motion period, the gimbal receives control commands to realize rotation and tracking. Reciprocate in this way to realize continuous tracking.
  • the smart phone sends a control command to control the pan/tilt to automatically track the locked key players through Bluetooth or WIFI.
  • the PID algorithm proportion integration differentiation
  • the specific method is to continuously calculate the error between the center point of the target (that is, the locked player) and the center point of the lens (screen), that is, the error_value value. Adjust and optimize the error_value value to achieve the best effect, that is, the error_value value presets a range, as long as the calculated error_value value is within this range, the lens will not rotate; once the error_value value deviates from the preset range, the smartphone The control program APP immediately issues a command to control the rotation of the gimbal, so that the error_value falls into the preset range again. With such constant repetition, it is possible to automatically track and control the smartphone to take pictures of the game.
  • the present invention also needs to further control the micro-jitter problem during shooting. Specifically, first, when the overall movement speed of the athlete is slow, the gimbal does not need to rotate too fast, that is, the gimbal will automatically follow the player's movement speed to adjust the rotation speed. Second, players usually pass faster when passing the center line of the field, but when the two end lines of the field are crowded, the relative displacement speed on the screen will also slow down. The rotation speed of the gimbal can also be slowed down. In order to achieve this control goal, we control two parameters to achieve the goal. First of all, we control the number of key players to be tracked.
  • the gimbal when only one player is tracked, the gimbal is definitely the most sensitive, but the picture is also the most likely to be unstable and cause micro-jitter.
  • the rotation speed of the gimbal is fast when passing the midfield, and the rotation speed of the gimbal is slow when it is on both sides of the court.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Computational Linguistics (AREA)
  • Psychiatry (AREA)
  • Social Psychology (AREA)
  • Studio Devices (AREA)

Abstract

本发明提出了一种自动跟踪拍摄体育比赛的智能系统及其控制方法，所述智能系统包括云台和智能手机，智能手机包括第一通信模块，云台包括第二通信模块。智能手机通过第一通信模块和云台的第二通信模块进行通信连接。本发明还提出了一种控制方法，对每个球员的人体标记出多个关节节点并进行动作识别，判断出每个球员的动作类型并进行分类，过滤掉不需要的动作类型而保留需要的动作类型的球员作为候选关键球员。在单位时间内计算出候选关键球员的位移。对候选关键球员在单位时间内的位移大小进行排序。选择一名或多名单位时间内位移大小排序靠前的候选关键球员标记为关键球员并锁定。最后自动跟踪被锁定的关键球员。

Description

一种自动跟踪拍摄体育比赛的智能系统及其控制方法 技术领域
本发明涉及智能拍摄技术领域，具体涉及一种自动跟踪拍摄体育比赛的智能系统及其控制方法。
背景技术
团队体育赛事跟踪拍摄
在团队体育赛事的跟踪拍摄上，目前市场上典型的自动化解决方案是利用一个或多个广角镜头架在高处，对整个体育场进行全景拍摄。对拍摄的全景视频进行基于深度学习的分析和处理之后，将相关的场景裁切出来，如球员对球争抢的画面等。这种系统设计通常需要专用的摄像头，并将其架设在高处，安装在场边。这一系统的缺点为：1.整个系统价格通常比较昂贵，仅使用于专业俱乐部、体育场馆等；2.设备需要提前在固定位置安装，或者比较笨重不便于携带。
现有技术提供了一种足球检测跟踪的摄像技术，例如中国专利文献1：CN106488127B，其公开了一种基于足球检测跟踪的摄像机切换控制方法及装置。该中国专利是基于足球检测跟踪的摄像机切换控制方法及装置。利用全景摄像机在足球被运动员遮挡时能够继续拍摄；还能够基于当前视频内容里处于运动状态的目标数量，从全景摄像机中截取镜头，并将镜头缩放到输出的标准画面进行拍摄；并且能够在触发位置预先设置摄像机，当足球位置与触发位置重合时，切换与该触发位置相应的摄像机进行拍摄。该专利可以根据足球的位置切换不同的摄像机镜头，完成无人值守情况下足球场上的比赛视频拍摄，能够更 好的满足球迷们在观看比赛时的需要。但该现有技术首先需要全景摄像机进行拍摄，获得运动员和球员在球场上的视频信息并经过多个控制模块的检测处理后，由切换模块切换到相应的摄像机进行拍摄。因此该现有技术仍然需要专用的摄像设备，且设备的系统构成复杂，系统成本依然较高。
中国专利文献2：CN105245845A和中国专利文献3：CN105208349A；均公开了一种比赛现场根据场上人数控制摄像头自动跟拍的方法；上述两件专利公开文献记载的摄像头自动跟拍方法能够根据场上人数控制摄像头自动跟拍，对场上人数进行识别并对摄像头进行自动控制。但上述专利需要设置远红外微光高速高清照相机、中心管理机房的主服务器、跟踪单元和高清可变焦360度转动摄像机的共同协作才能完成。该技术依然无法拜托系统架构的简化，系统布置成本依然无法降低。
基于手机平台智能跟踪拍摄
现存基于手机平台的智能跟踪拍摄应用还比较简单，目前尚未发现其他专利文献记载过利用手机平台对团队体育赛事跟踪拍摄的应用。针对体育赛事场景具体说来，现有手机平台的智能跟踪技术还存在如下不足：
(1)只能针对简单场景中的单一目标进行跟踪，画面背景不可太复杂，比如一张人脸，一个在移动中的物体等。在参与人数众多的大球类团队体育赛事中，这一假设显然是不成立的。
(2)在跟踪主体的选择上要么需要预先定义好某种固定特征，如跟踪画面中出现的人脸；要么需要用户人工干预，如用户用手指在屏幕上框选出待跟踪目标等，而这也是在动态的体育赛事中无法实现的。比如有多个人脸出现在画面里，该跟踪谁的脸，如果预先框选出的目标丢失了，该如何去更新跟踪目标。经过实测，目前具备手机智能跟踪功能的技术在实际多人团体赛事场景中，仅 能保持对目标跟踪几秒钟。
(3)在传统跟踪算法设计当中，针对人的跟踪主要考虑静态特征，比如人脸特征，体态特征、衣服颜色特征等，作为主要的特征来区分画面中的不同目标。在体育比赛中，由于球员相距镜头较远，难以利用脸部和身体体态特征加以识别，而球员球衣颜色统一，也难于对个别球员进行区分。
发明内容
针对现有技术存在的问题，本发明提出了一种自动跟踪拍摄体育比赛的智能系统及其控制方法。本发明提出的智能系统仅利用一个云台和智能用户终端即可实现，无需如专利文献1那样部署全景摄像机，也无需切换到具体的拍摄摄像机。同时也无需如专利文献2和3那样需要远红外微光高速高清照相机和高清可变焦360度转动摄像机。
本发明仅需要专门设计的云台配合用户的智能手持终端即可实现自动跟踪拍摄体育比赛。
同时，在拍摄识别算法上，在大球类体育赛事的视频跟踪拍摄存在很多的挑战，比如需要跟踪的对象要从“人”转变为“球“。从计算机视觉的角度来说，实时跟踪运动中的“球”这类快速移动的小目标，基于智能终端的摄像头现有的分辨率和AI芯片处理的速度都是非常困难的。这就需要我们根据实际场景设计出符合要求的跟踪算法来解决这些问题。而本发明的算法是基于对捕捉视频的理解，特别是根据对视频中人的移动和动作的分析理解来做出跟踪判断的。
首先，本发明的云台为一个在水平方向和垂直方向上均可转台的设备。所述云台内设置有蓝牙模块、WIFI或其他可连接智能终端的通信模块，用户可根 据具体的需要自行选择手持智能终端与云台的连接方式。例如云台通过蓝牙模块与智能终端的蓝牙进行匹配连接。或者由云台的WIFI模块与智能终端的WIFI模块组网进行连接。此外智能终端的WIFI模块还能与其他设备的WIFI模块连接，或者通过蜂窝网络(3G/4G/5G等)将自动跟踪拍摄体育比赛的画面实时传输到该其他设备进行直播或存储。云台和智能终端成功建立连接后，云台由智能终端上运行的应用程序APP进行自动控制。
本系统的核心除了可水平和垂直转动的云台外，还包括智能终端上为了实现智能终端实时控制云台而研发的应用程序APP。智能终端的应用程序APP把跟踪决策通过蓝牙或WiFi等通讯链路传给云台，云台在受控状态下带动智能终端跟随目标在水平及垂直方向转动，从而实现对体育比赛的智能跟踪拍摄。智能终端的定义一般应包含高清广角镜头、AI处理模块、CPU/GPU处理器、通讯模块及其他相关的模块。由于普通智能终端上不论超广角镜头(一般视角为120度左右)或普通广角(一般视角为60度左右)均视角有限，难于实现对体育比赛的全场跟踪。所以如采用智能终端或类似独立摄像头对体育比赛进行跟踪，需要对镜头拍摄方向进行控制，从而实现对体育活动的自动跟踪拍摄。体育活动影像通过高清广角镜头被捕捉，传送到AI处理模块分析处理，得出镜头跟踪的方向及转速等关键决策。
具体而言，本发明提出的跟踪决策策略是基于一个非常简单的对球类比赛的观察。即，团队球类运动中，多数球员的运动方向是以球所在位置为中心移动的。特别是离球越近的球员奔跑、拼抢越积极。所以本发明的跟踪决策策略首先通过分析整体球员的移动方向和移动轨迹来对球场上的移动趋势做出判断。之后，再分析个别球员的速度和动作，根据该分析来找到一名或几名关键球员，并对他们进行锁定跟踪。本发明的目标是使智能终端镜头拍摄的中心对准关键 球员，所以本发明依据关键球员在智能终端屏幕上的坐标来确定他们的中心点坐标，通过控制云台不断逼近这个中心点从而实现对关键球员的跟踪。此外，上述跟踪过程需要周期性不断重复，比如每2-3秒重复一次判断和寻找关键球员，并在接下来的周期里，对其进行锁定跟踪。
从目标搜寻和锁定的过程来说，比如，在球类项目当中，相对于慢速移动的球员，我们对处于高速奔跑状态中的球员更感兴趣。再进一步，在多位处于奔跑当中的球员里，我们又对当前控球的球员更感兴趣。而这些针对球员动作的识别与区分涉及到对球员动作和姿态的理解与识别。
从针对云台的控制具体来说，我们主要考虑两个因素。第一，运动员整体移动速度较慢的时候，云台也不需要转动太快。这主要是从人的视觉舒适度上来考虑的。第二，球员在通过场地中线的时候通常通过速度较快，但在场地的两个端线的时候人员比较集中，相对在屏幕上的位移速度也会慢下来。云台的转动速度也可以慢下来。为了实现这一控制目标，我们对两个参数控制来实现目标。首先，我们控制待跟踪关键球员的数量，比如只跟踪一名球员的时候，云台肯定最灵敏，但也是画面最容易不稳定，最容易产生微抖动的时候。其次，我们可以控制云台的转速，在通过中场的时候云台转速快，在球场两侧的时候云台转速慢。
具体来说，本发明提出的一种自动跟踪拍摄体育比赛的智能系统包括云台和智能手机。智能手机包括第一通信模块，云台包括第二通信模块。智能手机的第一通信模块和云台的第二通信模块为蓝牙模块、WIFI模块或其他科连接智能手机的通信模块。
云台与智能手机的通信连接方式采用蓝牙模块、WIFI或其他能够实现通信连接的方式进行。用户可根据具体的需要自行选择智能手机与云台的连接方式。 云台和智能手机成功建立连接后，云台由智能手机上运行的应用程序APP进行自动控制。
在本发明中，智能手机与云台的通信是双向的，而智能手机对云台的控制是单向的，即只能由智能手机向云台单向发出控制命令来控制云台的动作。
本发明自动跟踪拍摄的云台包括手机夹，其用于与智能手机本体物理连接。云台还包括能够在垂直方向上转动的垂直转动电机、能够在水平方向上转动的水平转动电机；垂直转动电机和水平转动电机的相互配合能够实现智能手机本体进行全方位全角度进行转动。云台还包括云台手柄及底座，在云台手柄及底座上方还设置云台的按键控制区。
本发明还提出了一种自动跟踪拍摄体育比赛的智能系统的控制方法，包括如下步骤：
S101，对每个球员的人体标记出多个关节节点；所标记的关节节点的数量可以根据具体所要自动跟踪拍摄的比赛项目来设置。
S102，智能手机拍摄镜头内的球员，识别出镜头内球员的人数后，对每个球员的关节节点进行动作识别。
S103，对每个球员的动作识别后进行判断，判断出每个球员的动作类型。进一步地，首先进行静态判断以判断出简单的动作；然后进行复杂动作判断，把关节节点的时间序列作为输入：{关节节点@t1，关节节点@t2,关节节点@t3……关节节点N@t N}(其中，N为自然数，t N为单位时间)，把预先定制化训练好的神经网络作为模型，由定制化神经网络对输入的关节节点序列进行判断。
S104，系统根据S103所判断出的动作类型进行分类，并对需要的动作类型进行保留，而不需要的动作类型进行过滤，动作类型被保留的球员作为候选关键球员。
S105，在单位时间内计算出所保留动作类型的候选关键球员的位移。
S106，根据计算结果对候选关键球员在单位时间内的位移大小进行排序。
S107，选择一名或多名单位时间内位移大小排序靠前的候选关键球员标记为关键球员并锁定。
S108，智能手机发出控制命令，通过蓝牙或WIFI控制云台自动跟踪被锁定的关键球员。
进一步地，对于被锁定的关键球员，不断的计算目标中心点和智能手机镜头或屏幕中心点的误差，得到error_value值；所述error_value值预设一个区间，只要计算得出的error_value值位于该区间内，镜头就不转动；一旦error_value值偏离预设区间，智能手机发出命令控制云台转动，使得error_value重新落入预设区间内。
本发明通过不断选择和锁定赛场上的关键球员，来实现对比赛热点场景的跟踪、拍摄和直播。
附图说明
图1为本发明所述的一种自动跟踪拍摄体育比赛的智能系统及其控制方法的控制流程图。
图2为本发明自动跟踪拍摄的智能系统结构中智能手机与云台连接结构正面示意图。
图3为本发明自动跟踪拍摄的智能系统结构中智能手机与云台连接结构背面示意图。
图4和图5为关节节点序列示意图。
图6为足球运动中一个动作的关节节点随时间变化的示意图。
图7为本发明提出的自动跟踪拍摄体育比赛的智能系统的框图。
具体实施方式
本发明提出的一种自动跟踪拍摄体育比赛的智能系统，包括云台和智能终端。其中智能终端例如是具有摄像头的智能手机、智能平板电脑或智能相机。
基于通用性考虑，本发明优选的智能终端为具有摄像头的智能手机。
下面以智能手机为例来进一步对本发明进行说明。
一种自动跟踪拍摄体育比赛的智能系统的结构框图如图7所示，包括云台12和智能手机11。智能手机11包括第一通信模块111，云台12包括第二通信模块121。智能手机11的第一通信模块111和云台12的第二通信模块121为蓝牙模块、WIFI模块或其他科连接智能手机11的通信模块。
云台12与智能手机11的通信连接方式采用蓝牙模块、WIFI或其他能够实现通信连接的方式进行。用户可根据具体的需要自行选择智能手机11与云台12的连接方式。例如云台12通过蓝牙模块与智能手机11的蓝牙进行匹配连接。或者由云台12的WIFI模块与智能手机11的WIFI模块组建成小局域网进行连接，此外WIFI模块还能与其他设备的WIFI模块连接，从而将自动跟踪拍摄体育比赛的画面实时传输到该其他设备进行直播或存储。云台12和智能手机11成功建立连接后，云台12由智能手机11上运行的应用程序APP进行自动控制。
在本发明中，智能手机11与云台12的通信是双向的，而智能手机11对云台12的控制是单向的，即只能由智能手机11向云台12单向发出控制命令来控制云台12的动作。
云台与智能终端的物理连接方式在本专利中不做特别限定，只要能够实现将智能终端牢固安装在云台上的所有固定方式均可应用。例如通过云台上的夹 持装置夹住智能终端。
图2和图3示出了本发明自动跟踪拍摄的智能系统结构中智能手机与云台的连接结构，包括：云台和智能手机。其中云台包括手机夹1，其用于与智能手机本体6物理连接，即将智能手机本体6固定在云台上。云台还包括能够在垂直方向上转动的垂直转动电机2、能够在水平方向上转动的水平转动电机3；垂直转动电机2和水平转动电机3的相互配合能够实现智能手机本体6进行全方位全角度进行转动。云台还包括云台手柄及底座4，在云台手柄及底座4上方还设置云台的按键控制区5，在自动跟踪拍摄体育比赛时，用户手拿着云台手柄及底座4，并通过按键控制区5对云台进行控制操作。
图1为本发明提出的一种自动跟踪拍摄体育比赛的智能系统的具体控制流程。
对于团体性的球类运动而言，多数球员的运动方向是以随球所在位置为中心移动的。特别是离球越近的球员奔跑、拼抢越积极。所以本发明的跟踪决策策略首先通过分析整体球员的移动方向和移动轨迹来对球场上的移动趋势做出判断。之后，再分析个别球员的速度和动作，根据该分析来找到一名或几名关键球员，并对他们进行锁定。因此本发明的跟踪决策策略首先识别出球员的动作类型，该动作类型包括：奔跑、跳跃、站立等。
如图1所示：
S101，为了识别球员的动作类型，需要对每个球员的人体标记出多个关节节点；所标记的关节节点的数量可以根据具体所要自动跟踪拍摄的比赛项目来设置。例如采用14个、17个或者19个关节节点。关节节点数量设置得越多，对球员的肢体和肌肉的追踪就越细腻，因而对球员动作的判断和跟踪就越精确。然而关节节点的标记也并不是越多越好，因为对关节节点的标记越多，系统对 关节节点变化的分析运算量就越大，这对系统的数据处理能力将提出更高的要求。在本发明中，充当系统数据计算的设备是智能手机本体6，对于具体需要标记多少数量关节节点才能合理匹配用户使用本系统时的智能手机本体6，这对于大多数用户来说过于专业，因而本发明提出的自动跟踪拍摄系统可由用户在其智能手机本体6上安装的APP中选择相应的智能手机本体6的品牌和型号，从而在具体应用时能够由系统自动匹配合理的关节节点数量。
S102，智能手机拍摄镜头内的球员，识别出镜头内球员的人数后，对每个球员的关节节点进行动作识别。
S103，对每个球员的动作识别后进行判断，判断出每个球员的动作类型。对于动作类型的判断，取决于需要判断的动作类型的动作复杂程度，比如区分一些简单的动作，如站立静止或奔跑，这采用静止的关节节点图就可以做出判断。图4和图5示意性地给出了静态关节节点的标记示例。图4和图5均采用14个关节节点进行标记，包括球员的头部、颈部、左肩、右肩、左肘、右肘、左腕、右腕、左髋、右髋、左膝、右膝、左踝和右踝。而对于一些更复杂的动作，比如足球中的射门或者篮球中的投篮，就需要根据关节节点序列随时间的变化才可以做出判断。比如把关节节点时间序列作为输入，{关节节点@t1，关节节点@t2,关节节点@t3……关节节点N@t N}(其中，N为自然数，t N为单位时间)，把预先定制化训练好的神经网络作为模型，由定制化神经网络对输入的关节节点序列进行判断，从而决定哪个动作是奔跑，哪个动作是射门等。图6为球员在足球射门时的关节节点随时间变化的序列示意图。图6中由左至右分别为时刻t1-t5时刻关节节点的序列图，根据预先定制好的模型，系统即可判断出该动作为足球射门动作。其他动作的判断也与此类似。
S104，系统根据S103所判断出的动作类型进行分类，并对需要的动作类型 进行保留，而不需要的动作类型进行过滤，动作类型被保留的球员作为候选关键球员。例如，在体育比赛中，更关注的是处于奔跑状态的球员，因此，对于动作类型被识别为站立状态的球员，系统将对其进行过滤，而保留奔跑状态的球员作为候选关键球员。
S105，在单位时间内计算出所保留动作类型的候选关键球员的位移。在单位时间内，计算候选关键球员的位移，该单位时间例如0.5秒、1秒、1.5秒，2秒。在实际处理当中，由于云台在不停的转动，这会让我们比较难于对球员的位移做出判断。比如云台往右转动时，有球员做出了轻微的往左的运动，这时我们将无法判断他的准确位移。为了解决这个问题，我们人为的把云台分成静止周期和运动周期。即云台在静止周期的时候系统对球员的运动和位移做出判断，锁定关键球员；之后在运动周期的时候，云台接受控制命令，实现转动和追踪。如此往复，实现连续追踪。
S106，根据计算结果对候选关键球员在单位时间内的位移大小进行排序。
S107，选择一名或多名单位时间内位移大小排序靠前的候选关键球员标记为关键球员并锁定。
S108，智能手机发出控制命令，通过蓝牙或WIFI控制云台自动跟踪被锁定的关键球员。
进一步的，对于被锁定的关键球员，采用PID算法(proportion integration differentiation)使镜头的中心点对准下一刻需要瞄准的目标。具体做法就是不断的计算目标中心点(即锁定球员)和镜头(屏幕)中心点的误差，即error_value值。对error_value值进行调整优化使其达到最佳的效果，也就是error_value值预设一个区间，只要计算得出的error_value值位于该区间内，镜头就不转动；一旦error_value值偏离预设区间，智能手机的控制程序APP 立刻发出命令控制云台转动，使得error_value重新落入预设区间内。如此不断的反复，就可以自动追踪控制智能手机拍摄比赛的画面。
此外，本发明还需要进一步控制拍摄时的微抖动问题。具体来说，第一，运动员整体移动速度较慢的时候，云台也不需要转动太快，即云台会自动跟随球员的运动速度来调整转动速度。第二，球员在通过场地中线的时候通常通过速度较快，但在场地的两个端线的时候人员比较集中，相对在屏幕上的位移速度也会慢下来。云台的转动速度也可以慢下来。为了实现这一控制目标，我们对两个参数控制来实现目标。首先，我们控制待跟踪关键球员的数量，比如只跟踪一名球员的时候，云台肯定最灵敏，但是画面也最容易不稳定，产生微抖动。其次，我们可以控制云台的转速，在通过中场的时候云台转速快，在球场两侧的时候云台转速慢。
以上阐述了本发明的技术方案,一切不脱离本发明的技术方案之实质的替代，都应在本发明的权利要求的范围内。

Claims (10)

  1. 一种自动跟踪拍摄体育比赛的智能系统，包括：云台和智能终端；智能终端包括第一通信模块，云台包括第二通信模块；智能终端和云台通过第一通信模块和第二通信模块进行通信连接；云台由智能手机上运行的应用程序APP进行自动控制。
  2. 如权利要求1所述的一种自动跟踪拍摄体育比赛的智能系统，其特征在于：智能终端的第一通信模块和云台的第二通信模块为蓝牙模块、WIFI模块；所述云台包括能够在垂直方向上转动的垂直转动电机、能够在水平方向上转动的水平转动电机；垂直转动电机和水平转动电机的相互配合能够实现智能手机本体进行全方位全角度进行转动；云台还包括云台手柄及底座，在云台手柄及底座上方还设置云台的按键控制区，在自动跟踪拍摄体育比赛时，用户手拿着云台手柄及底座，并通过按键控制区对云台进行控制操作。
  3. 如权利要求1所述的一种自动跟踪拍摄体育比赛的智能系统，其特征在于：智能终端与云台的通信是双向的，智能终端仅单向控制云台。
  4. 一种如权利要求1所述的自动跟踪拍摄体育比赛的智能系统的控制方法，包括以下步骤：
    智能手机拍摄对进行的体育比赛进行拍摄，然后：
    S101，对每个球员的人体标记出多个关节节点；所标记的关节节点的数量根据具体所要自动跟踪拍摄的比赛项目来设置；
    S102，智能手机拍摄镜头内的球员，识别出镜头内球员的人数后，对每个球员的关节节点进行动作识别；
    S103，对每个球员的动作识别后进行判断，判断出每个球员的动作类型；
    S104，根据S103所判断出的动作类型进行分类，并对需要的动作类型进行保留，而不需要的动作类型进行过滤，动作类型被保留的球员作为候选关键球 员；
    S105，在单位时间内计算出所保留动作类型的候选关键球员的位移；
    S106，根据计算结果对候选关键球员在单位时间内的位移大小进行排序；
    S107，选择一名或多名单位时间内位移大小排序靠前的候选关键球员标记为关键球员并锁定；
    S108，智能手机发出控制命令，通过蓝牙或WIFI控制云台自动跟踪被锁定的关键球员。
  5. 一种如权利要求4所述的控制方法，其中，所述步骤S103还进一步包括：
    首先根据静态关节节点判断简单的动作进行；然后进行复杂动作判断，把关节节点按照时间顺序的序列作为输入：{关节节点@t1，关节节点@t2,关节节点@t3……关节节点N@t N}，其中，N为自然数，t N为单位时间；把预先定制化训练好的神经网络作为模型，由定制化神经网络对输入的关节节点时间序列进行判断。
  6. 一种如权利要求4所述的控制方法，其中，所述步骤S108后，还进一步包括：对于被锁定的关键球员，不断的计算被锁定的关键球员和智能手机镜头或屏幕中心点的误差，得到error_value值；所述error_value值预设一个区间，只要计算得出的error_value值位于该区间内，镜头就不转动；一旦error_value值偏离预设区间，智能手机发出命令控制云台转动，使得error_value重新落入预设区间内。
  7. 一种如权利要求4所述的控制方法，进一步地，将云台分成静止周期和运动周期，即云台在静止周期的时候系统对球员的运动和位移做出判断，锁定关键球员；之后在运动周期的时候，云台接受控制命令，实现转动和追踪；由于球员通常移动速度很快，这两个周期都很短，在1/10秒的数量级上。
  8. 一种如权利要求4所述的控制方法，进一步地，当球员整体移动速度较慢的时，云台会自动跟随球员的运动速度来调整转动速度。
  9. 一种如权利要求4所述的控制方法，进一步地，球员在通过场地中线的时云台转速比在球场两侧的时云台转速快。
  10. 一种如权利要求7所述的控制方法，进一步地，所述静止周期小于0.5秒和运动周期均小于3秒。
PCT/CN2022/103544 2022-01-05 2022-07-04 一种自动跟踪拍摄体育比赛的智能系统及其控制方法 WO2023130696A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210005422 2022-01-05
CN202210005422.2 2022-01-05

Publications (1)

Publication Number Publication Date
WO2023130696A1 true WO2023130696A1 (zh) 2023-07-13

Family

ID=82521244

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/103544 WO2023130696A1 (zh) 2022-01-05 2022-07-04 一种自动跟踪拍摄体育比赛的智能系统及其控制方法

Country Status (2)

Country Link
CN (1) CN114827478B (zh)
WO (1) WO2023130696A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105443956A (zh) * 2015-12-26 2016-03-30 武汉智能鸟无人机有限公司 一种带跟踪功能的手持云台
CN106056035A (zh) * 2016-04-06 2016-10-26 南京华捷艾米软件科技有限公司 一种基于体感技术的幼儿园智能监控方法
US20160337573A1 (en) * 2015-05-15 2016-11-17 Lak Wang Chan Camera, method, and system for filming golf game
CN107871120A (zh) * 2017-11-02 2018-04-03 汕头市同行网络科技有限公司 基于机器学习的体育赛事理解系统及方法
CN111314527A (zh) * 2019-09-20 2020-06-19 深圳市启明创新科技开发有限公司 一种智能跟随云台及方法
WO2021162305A1 (ko) * 2020-02-15 2021-08-19 김지훈 스포츠 동영상 기반 플랫폼 서비스를 제공하는 서버의 동작 방법

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1657927A1 (en) * 2004-11-12 2006-05-17 Saab Ab Image-based movement tracking
CN109903312B (zh) * 2019-01-25 2021-04-30 北京工业大学 一种基于视频多目标跟踪的足球球员跑动距离统计方法
WO2021026789A1 (zh) * 2019-08-13 2021-02-18 深圳市大疆创新科技有限公司 基于手持云台的拍摄方法、手持云台及存储介质
CN111783692A (zh) * 2020-07-06 2020-10-16 广东工业大学 一种动作识别方法、装置及电子设备和存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160337573A1 (en) * 2015-05-15 2016-11-17 Lak Wang Chan Camera, method, and system for filming golf game
CN105443956A (zh) * 2015-12-26 2016-03-30 武汉智能鸟无人机有限公司 一种带跟踪功能的手持云台
CN106056035A (zh) * 2016-04-06 2016-10-26 南京华捷艾米软件科技有限公司 一种基于体感技术的幼儿园智能监控方法
CN107871120A (zh) * 2017-11-02 2018-04-03 汕头市同行网络科技有限公司 基于机器学习的体育赛事理解系统及方法
CN111314527A (zh) * 2019-09-20 2020-06-19 深圳市启明创新科技开发有限公司 一种智能跟随云台及方法
WO2021162305A1 (ko) * 2020-02-15 2021-08-19 김지훈 스포츠 동영상 기반 플랫폼 서비스를 제공하는 서버의 동작 방법

Also Published As

Publication number Publication date
CN114827478A (zh) 2022-07-29
CN114827478B (zh) 2022-09-16

Similar Documents

Publication Publication Date Title
US11380100B2 (en) Methods and systems for ball game analytics with a mobile device
US11157742B2 (en) Methods and systems for multiplayer tagging for ball game analytics generation with a mobile computing device
US11842572B2 (en) Device, system, and method of computer vision, object tracking, image analysis, and trajectory estimation
US11638854B2 (en) Methods and systems for generating sports analytics with a mobile device
Thomas et al. Computer vision for sports: Current applications and research topics
WO2019229748A1 (en) Golf game video analytic system
US9615015B2 (en) Systems methods for camera control using historical or predicted event data
CN109961039B (zh) 一种个人进球视频捕捉方法及系统
CN107798692A (zh) 一种球类图像识别方法及其系统
JP7078577B2 (ja) 動作類似度評価装置、方法およびプログラム
JP7198661B2 (ja) オブジェクト追跡装置及びそのプログラム
CN114037923A (zh) 一种目标活动热点图绘制方法、系统、设备及存储介质
WO2023130696A1 (zh) 一种自动跟踪拍摄体育比赛的智能系统及其控制方法
CN114025183A (zh) 直播方法、装置、设备、系统和存储介质
CN108970091A (zh) 一种羽毛球动作分析方法及系统
CN115475373B (zh) 运动数据的展示方法、装置、存储介质及电子装置
CN113302906A (zh) 图像处理设备、图像处理方法、计算机程序和存储介质
NO20161098A1 (en) Camera system for filming sports venues
CA2633197A1 (en) Method and system for creating event data and making same available to be served
US11900678B2 (en) System for tracking, locating and calculating the position of an object in a game involving moving objects
CN108833777A (zh) 一种摄像摄影装置及进行摄像摄影的方法
US11707663B1 (en) System for tracking, locating and predicting the position of a ball in a game of baseball or similar
CN115278082B (zh) 视频拍摄方法、视频拍摄装置及电子设备
JP7289630B2 (ja) 画像処理装置
JP7233886B2 (ja) 画像処理装置