A multi-position photography cloud platform control system

Abstract

The application relates to the technical field of multi-camera photography cloud platform control, and discloses a multi-camera photography cloud platform control system, which comprises a central control host, at least one cloud platform sub-node and a double-channel communication link. The central control host comprises a state machine director logic module, a predictive verification engine, a dynamic path re-planning module and a state transition readiness verification module. The predictive verification engine continuously performs kinematics simulation and safety verification on all potential state transitions in the background. If it is detected that there is a risk of a slip between the cloud platforms, the dynamic path re-planning module will generate a corrected safe motion path. The state transition readiness verification module queries the internal readiness state of the target cloud platform sub-node before actually executing the state transition, so as to ensure that parameters such as focusing and exposure are ready, and the transition can be executed. The application realizes the automatic, high-safety and high-picture-quality collaborative control of the multi-camera photography cloud platform, and improves the professional director production efficiency and picture quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-camera photography gimbal control, in particular to a multi-camera photography gimbal control system. BACKGROUND

[0002] In professional photography production, photography gimbals are widely used for picture stabilization, motion control and subject tracking. Currently, the control panel of most photography gimbals is usually integrated on the handle. The operator directly controls the gimbal by holding the handle to realize photography angle conversion, subject tracking and other operations. The control circuit and the gimbal motor part are usually directly connected through hardware lines.

[0003] The inherent limitation of this control method is that one operator can only control one gimbal at the same time, and must hold the handle on site. When multiple cameras and multiple angles need to be switched, for example, when the same subject is shot from different directions and angles, multiple operators must be arranged to control multiple gimbals. This solution not only increases the labor input, but also reduces the flexibility of shooting deployment. In a complex director environment, the coordinated movement between multiple gimbals, the avoidance of one gimbal entering the shooting picture of another camera (i.e. the risk of being exposed), and the assurance of clear focus and accurate exposure when switching, all of these picture quality achievements are highly dependent on the experience and real-time judgment of the operator. The inherent characteristics of manual coordination make it difficult for the system to achieve highly automated and standardized quality control, thereby increasing the risk of potential shooting errors and limiting the further improvement of overall shooting efficiency.

[0004] Therefore, the present application provides a multi-camera photography gimbal control system to solve the problems of the prior art. SUMMARY

[0005] In view of the shortcomings of the prior art, the present application provides a multi-camera photography gimbal control system to solve the problems of high labor input, low coordination efficiency, difficulty in automated guarantee of picture quality, and the risk of exposure in the prior art.

[0006] To achieve the above purpose, the present application realizes the following technical scheme: a multi-camera photography gimbal control system, comprising:

[0007] A central control host for managing and coordinating the operation of multiple gimbal sub-nodes;

[0008] At least one gimbal sub-node for executing the control instructions issued by the central control host;

[0009] A dual-channel communication link for data exchange between the central control host and at least one gimbal sub-node;

[0010] The central control host comprises:

[0011] a state machine director logic module for storing a preset system state set, an event set, and a logical transition relationship between states, and deciding a target state according to a received external event;

[0012] a predictive verification engine for continuously running in the background, acquiring attitude data of all gimbal sub-nodes in real time, and performing kinematics simulation and safety verification on all potential next state transitions in the state machine director logic module;

[0013] a dynamic path replanning module for working in cooperation with the predictive verification engine, and generating a corrected safe motion path when the predictive verification engine predicts that a default path of state transition has a risk of physical interference;

[0014] a state transition readiness verification module for being called before state transition execution to query an internal readiness state of a target gimbal sub-node.

[0015] Preferably, the dual-channel communication link comprises:

[0016] a control link for transmitting high-priority, low-latency control instructions, real-time attitude data of the gimbal sub-nodes, and state response codes between the central control host and the gimbal sub-nodes;

[0017] a data link for transmitting high-bandwidth media data streams between the central control host and the gimbal sub-nodes, the media data streams including real-time preview video streams or material files.

[0018] Preferably, the state machine director logic module is based on a deterministic finite state machine model, and the finite state machine model comprises:

[0019] a finite state set, wherein each state is determined by assigning one of an ON_AIR role, a PREVIEW role, a TRANSFER role, or an IDLE role to all the gimbal sub-nodes in the system;

[0020] a finite event set for triggering state changes;

[0021] a logical transition matrix for defining conversion rules from a current state to a target state.

[0022] Preferably, the predictive verification engine performs parallel simulation and pure kinematics geometry anti-collision calculation when performing safety verification.

[0023] Preferably, the pure kinematics geometry anti-collision calculation comprises:

[0024] A global unified three-dimensional world coordinate system is established, and the accurate three-dimensional positions of all the pan-tilt sub-nodes and the real-time attitudes of the camera lenses of the pan-tilt sub-nodes in space are mapped into the three-dimensional world coordinate system;

[0025] A set of pan-tilt sub-nodes assuming the role of ON_AIR and a set of pan-tilt sub-nodes assuming the role of TRANSFER in the state transition are identified;

[0026] The field of view range of each pan-tilt sub-node camera assuming the role of ON_AIR is defined as a field of view cone;

[0027] The space occupied by each pan-tilt sub-node assuming the role of TRANSFER in the movement process is defined as a movement envelope;

[0028] Collision detection is performed to determine whether the movement envelope geometrically intersects with the field of view cone, and if there is an intersection, it is determined that there is a risk of a slip.

[0029] Preferably, the field of view cone is determined by the accurate position, orientation and optical parameters of the lens of the pan-tilt sub-node camera assuming the role of ON_AIR in the three-dimensional world coordinate system; and the movement envelope is a boundary volume that moves along the simulated movement path of the pan-tilt sub-node assuming the role of TRANSFER and completely contains the physical entity of the pan-tilt sub-node.

[0030] Preferably, the modified safe movement path generated by the dynamic path re-planning module aims to generate a new movement path for the pan-tilt sub-node from the initial attitude to the target attitude, and the new movement path satisfies the no-slip constraint throughout the movement schedule.

[0031] Preferably, the dynamic path re-planning module uses a search-based algorithm, a sampling-based algorithm or an interpolation-based algorithm to generate the modified safe movement path.

[0032] Preferably, the state transition readiness verification module:

[0033] Sends a readiness state query instruction to the target pan-tilt sub-node through the control link to request the target pan-tilt sub-node to report the internal readiness state;

[0034] When the internal readiness state is not ready, the state transition is prevented and the polling is continued until the internal readiness state becomes ready.

[0035] Preferably, the internal readiness state includes at least one of a focus lock state, an exposure stability state, a white balance calibration state or a movement mechanism stability state.

[0036] The present application provides a multi-camera photography pan-tilt control system.

[0037] 1、The application sets up a predictive verification engine and a dynamic path replanning module, and performs kinematic simulation and geometric interference detection on all potential motion paths before actual state transition execution. Once the motion envelope and the view cone are detected to have a risk of penetration, the system generates a corrected safe path. This forward-looking safety verification mechanism eliminates potential shooting accidents in the decision-making stage, avoids broadcast accidents caused by gimbal motion blocking live pictures, and significantly improves the safety and reliability of multi-camera automated system operation.

[0038] 2、The application adds a state transition readiness verification module, which actively queries the internal readiness state of the target gimbal subnode before executing state switching. The system ensures that only when the camera's focus, exposure and other parameters reach a stable state, the picture switching or lens operation is performed. If the target position is not ready, the system automatically enters standby until it is ready, effectively avoiding switching to a blurred or improperly exposed picture, thereby ensuring the professionalism and technical quality of the final output picture.

[0039] 3、The application uses a dual-channel communication link design to separate high-priority control instructions and high-bandwidth media data streams physically or logically. The control link is dedicated to transmitting low-latency control instructions and gimbal attitude data, while the data link carries real-time video streams. Through this design, it is ensured that the transmission of critical control instructions will not be delayed or packet loss due to bandwidth occupation by large-capacity data streams, ensuring the real-time and stability of the entire system control.

[0040] 4、The application integrates complex director logic, motion safety verification, and position state checking into the central control host for automated processing. Through the state machine director logic module, the operator only needs to trigger high-level events to drive the entire multi-camera system, without the need for independent manual control of each gimbal. The system automatically completes path planning, penetration avoidance, and picture readiness confirmation, which were originally completed by multiple people, integrating traditional director, camera operator and other multi-post functions, making it possible for a single person to complete multi-camera shooting and production, thereby significantly reducing on-site manpower and improving overall shooting efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The figure is a schematic diagram of the system architecture of the application;

[0042] Figure 2 The figure is a schematic diagram of the method flow of the application.

[0043] Wherein, 100, a central control host; 200, a PTZ sub-node; 300, a dual-channel communication link; 310, a control link; 320, a data link; 400, a state machine director logic module; 500, a predictive verification engine; 600, a dynamic path re-planning module; 700, a state transition readiness verification module; 800, an action primitive library. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0045] Referring to the accompanying drawings Figure 1 The present application provides a multi-camera photography PTZ control system, which can include:

[0046] The central control host 100 and at least one PTZ sub-node 200 exchange data through a dual-channel communication link 300; the central control host 100 is used to generate and send control instructions, and the PTZ sub-node 200 is used to receive and execute the control instructions to drive the photography PTZ to complete the specified action.

[0047] The dual-channel communication link 300 is divided at the physical layer or the logical layer, including:

[0048] The control link 310 is used to transmit high-priority, low-delay control instructions, real-time PTZ attitude data and state response codes between the central control host 100 and the PTZ sub-node 200;

[0049] The data link 320 is used to transmit high-bandwidth media data streams, such as real-time preview video streams or material files, between the central control host 100 and the PTZ sub-node 200.

[0050] By separating the transmission channels of the control instructions and the media data streams, the real-time performance and reliability of the control instructions are ensured not to be disturbed by the high-bandwidth data transmission.

[0051] The central control host 100 internally integrates a plurality of logically working modules; in an embodiment, these modules can include:

[0052] The state machine director logic module 400 is used to store a preset system state set, an event set and a logical transition relationship (i.e. a logical transition matrix) between states, and to decide the target state according to the received external events (such as operator input).

[0053] A predictive verification engine 500 is used to continuously run in the background, acquiring real-time pose data of all PTZ child nodes 200, and performing kinematic simulation and safety verification on all possible next state transitions in the state machine director logic module 400.

[0054] A dynamic path replanning module 600 is used in conjunction with the predictive verification engine 500, and is activated to generate a corrected, interference-free safe motion path when it is predicted that the default path of state transition has a risk of physical interference (e.g. a “walk-through”).

[0055] A state transition readiness verification module 700 is called before state transition execution, and is used to query the internal readiness state (e.g. whether the focus is locked) of the target PTZ child node 200, to determine whether to immediately execute the transition or enter automatic standby until the target is ready.

[0056] The PTZ child node 200 internally includes an action primitive library 800, which stores a plurality of standardized atomic action units (e.g. smooth panning, standby tracking, state transition, etc.), and the PTZ child node 200 invokes and executes the corresponding action primitive according to the role instruction sent by the central control host 100.

[0057] The system of the present application first performs initialization, the central control host 100 loads the state machine director logic module 400, and establishes a three-dimensional kinematic model of all PTZ child nodes 200 for use by the predictive verification engine 500.

[0058] During system operation, the predictive verification engine 500 continuously and in parallel simulates all possible next state transitions in the background; for any simulation, the engine invokes the dynamic path replanning module 600 to determine whether there is a risk of a “walk-through” by pure kinematic geometric calculation (e.g. frustum projection). If there is a risk, the dynamic path replanning module 600 generates a corrected safe path. All verified or corrected transition schemes are cached.

[0059] When an external event (e.g. operator button press) is triggered, the state machine director logic module 400 determines the target state; at this time, the system does not immediately execute, but first invokes the state transition readiness verification module 700 to query the readiness state of the target PTZ child node 200 through the control link 310.

[0060] If the target PTZ child node 200 is not ready, the state transition readiness verification module 700 will prevent state transition, causing the system to remain in the current state until the PTZ child node 200 feedbacks a readiness signal.

[0061] If the target PTZ child node 200 feeds back ready, the central control host 100 immediately extracts the safe transfer scheme (including the corrected safe path) previously calculated by the predictive verification engine 500 and the dynamic path re-planning module 600 from the cache and sends it as an instruction to the corresponding PTZ child node 200 through the control link 310; the action primitive library 800 of the PTZ child node 200 receives the instruction and executes the corresponding safe action; in this process, the preview picture of the PTZ child node 200 is returned back through the data link 320, and its real-time attitude is continuously fed back to the central control host 100 through the control link 310 for the predictive verification engine 500 to perform the next cycle of calculation.

[0062] Referring to the accompanying drawings Figure 2 , the present application provides a multi-camera shooting PTZ control method, comprising the following steps:

[0063] S100, system initialization and double-channel network construction step: deploy the central control host 100 and multiple PTZ child nodes 200, establish a physically or logically isolated double-channel communication link 300; the central control host 100 loads the state machine director logic module 400 and establishes a three-dimensional kinematics model based on the physical parameters of each PTZ child node 200.

[0064] S200, predictive verification and dynamic path re-planning step: before receiving an external state switching event, the predictive verification engine 500 continuously collects real-time attitude data of each PTZ child node 200 through the control link 310, and simulates all potential next state transition paths in the state machine director logic module 400 in parallel; the dynamic path re-planning module 600 performs interference detection based on geometric view cones for the simulated paths, and if it detects a line-of-sight obstruction or a physical collision risk, it generates a corrected safe motion path and stores it in cache.

[0065] S300, event response and target state decision step: the central control host 100 receives an external triggering event, and the state machine director logic module 400 determines the target state of the system according to the preset logical transition matrix.

[0066] S400, ready state check and automatic standby switching step: the state transition ready verification module 700 sends a state query instruction to the PTZ child node 200 involved in the target state through the control link 310; if the received feedback signal is not ready, the current state is maintained and waiting; if the received feedback signal is ready, the state execution is triggered.

[0067] S500, safe path execution and action primitive calling step: the central control host 100 calls the verified or corrected safe transfer scheme cached in step S200 and issues an instruction through the control link 310; the PTZ child node 200 completes the physical motion according to the instruction by calling the corresponding action unit in the action primitive library 800.

[0068] The module technical implementation details of the above systems will be described in detail below in combination with specific principles.

[0069] Referring to the accompanying Figure 1 The application provides a multi-camera photography cloud platform control system, which can include: a central control host 100, and at least one cloud platform sub-node 200.

[0070] The central control host 100 is the core processing unit of the system, responsible for the running of control logic, the decision of state, the planning and verification of motion path; the hardware composition of the central control host 100 includes but is not limited to: a high-performance processor, used for executing complex computing tasks, such as state machine logic, kinematics simulation and path planning algorithm; a storage unit, used for storing system configuration data, a three-dimensional model of the cloud platform, a preset state transition matrix, an action primitive library, and real-time running data and pre-computed safety path; a network interface module, used for establishing a wireless communication connection with the cloud platform sub-node 200; and an operator interactive interface, including a display, an input device (such as a physical button, a touch screen, etc.), for the operator to make system settings, input instructions and real-time monitoring.

[0071] The cloud platform sub-node 200 is the execution unit of the system, responsible for receiving and accurately executing instructions from the central control host 100, and real-time acquisition of its own state data; the hardware composition of each cloud platform sub-node 200 includes but is not limited to: a high-precision robot cloud platform system, which integrates multiple degrees of freedom servo motors and encoders, used for realizing accurate attitude control of the camera; a camera unit, including an imaging sensor, a variable focus lens and an automatic focusing mechanism, used for completing image acquisition; a local processing unit, used for analyzing instructions from the central control host 100, executing a local control loop and acquiring sensor data of the cloud platform and the camera; a local storage unit, used for storing high-resolution raw shooting materials; and a wireless communication unit, used for realizing data transmission with the central control host 100.

[0072] The central control host 100 and the at least one cloud platform sub-node 200 exchange information through a double-channel communication link 300; the double-channel communication link 300 is physically or logically isolated in technical implementation, to ensure the transmission characteristics of different types of information.

[0073] The control link 310 is the first link in the dual-channel communication link 300, mainly used for transmitting high-priority, low-delay, and low-data-volume control information; the control link 310 can work in the 2.4 GHz frequency band, and adopt a wireless communication protocol suitable for low-delay and high-reliability transmission, such as the IEEE-802.15.4 standard or a customized narrowband wireless protocol; the information transmitted through the link includes: the role instruction sent by the central control host 100 to the pan-tilt sub-node 200, such as executing the ON_AIR primitive or entering the PREVIEW mode; the ready state query instruction sent by the central control host 100 to the pan-tilt sub-node 200; the self-ready state signal fed back by the pan-tilt sub-node 200 to the central control host 100; and the real-time attitude data (such as the three-dimensional position, attitude angle, lens zoom and focus parameters of the camera) reported by the pan-tilt sub-node 200 at high frequency; the control link 310 is designed with redundancy and error control mechanism to ensure the reliable delivery of key control instructions.

[0074] The data link 320 is the second link in the dual-channel communication link 300, mainly used for transmitting high-bandwidth, relatively high-delay-tolerant data streams; the data link 320 can work in the 5.8 GHz frequency band, and adopt a high-speed wireless communication standard, such as IEEE-802.11ax (Wi-Fi6) or similar wideband wireless technology; the information transmitted through the link includes: the preview video stream collected by the pan-tilt sub-node 200 in real time and returned to the central control host 100 for the operator to watch and make rough composition decisions; and after the shooting task is completed, the pan-tilt sub-node 200 uploads the locally stored high-resolution raw shooting materials to the central control host 100 or the designated storage server in batches; the design of the data link 320 focuses on throughput to meet the needs of large-capacity data transmission.

[0075] The action primitive library 800 is pre-stored in the local storage unit of each pan-tilt sub-node 200, and internally contains a set of standardized defined atomic action units; these atomic action units are abstract descriptions of specific behaviors of the pan-tilt and its mounted camera, and each primitive encapsulates the corresponding kinematic characteristics, control parameters, and execution logic; when the central control host 100 issues a role instruction to the pan-tilt sub-node 200, it means that the pan-tilt sub-node 200 should call and execute a specific action primitive in the action primitive library 800.

[0076] S211: Define pan-tilt control variables; the control of the attitude and lens parameters of the pan-tilt sub-node 200 is performed through a unified control variable vector is described as follows:

[0077] ;

[0078] wherein, represents the attitude angle of the gimbal in three-dimensional space (e.g. yaw angle, pitch angle, roll angle), defining the pointing direction of the gimbal; represents the angular velocity of the gimbal attitude change; represents the optical parameters of the camera lens, such as represents the zoom focal length, represents the focus distance, represents the aperture size; each action primitive controls some or all of the variables according to its own function.

[0079] S212: define the ON_AIR primitive ( ); for performing preset camera movements in live broadcast or recording; the motion trajectory design of this primitive follows the principle of smoothness to avoid jerky feeling in the start-stop picture; its kinematic characteristics are mathematically described by smooth transition of position, velocity and acceleration, and S-shaped velocity curve can be used; the S-shaped velocity curve ensures the smoothness of the motion by limiting the jerk (i.e. jerk) in the motion process;

[0080] If represents the attitude angle of the gimbal in a certain degree of freedom in time , and represents the jerk of this degree of freedom, then the motion constraint of

[0081] ;

[0082] wherein is the preset maximum jerk limit value, which ensures smooth transition of the gimbal motion in the acceleration and deceleration stages; The control parameters of

[0083] S213: define the PREVIEW primitive ( ); for the gimbal to track the predetermined target or maintain a specific composition when no picture is played out; the motion characteristics of this primitive are relatively gentle in response speed, allowing a certain target tracking error, and can continuously adjust to keep the target in the picture. When is executed, the gimbal sub-node 200 continuously monitors the position information of the target and fine-tunes the attitude of the gimbal through an internal feedback control loop (e.g. PID controller); at the same time, this primitive also controls the camera lens to perform automatic focusing to ensure that the target is always clear; ​The control parameters of the IDLE primitive include the static target pose or standby pose information.

[0084] S214: define the TRANSFER primitive ); for the gimbal to smoothly transition from a starting pose to a target pose ; the path generation of this primitive can be based on simple linear interpolation (e.g. in joint space or Cartesian space), but in order to meet the needs of dynamic path replanning, its path definition is plastic; It can be connected by a series of discrete key pose points, or described by a parametric curve (e.g. Bezier curve or B-spline curve); the central control host 100 will attach a specific path point sequence or curve parameters when issuing ; The control parameters of the TRANSFER primitive include the starting pose , the target pose , the total motion time, the maximum motion speed limit, and the key intermediate path points or curve parameters generated by the dynamic path replanning module 600.

[0085] S215: define the IDLE primitive ; for the gimbal to maintain a pre-defined static pose during task intervals or system idle, such as parking in a safe location; the kinematic feature of this primitive is static state, or moving to a certain fixed standby position at a pre-defined speed; The control parameters of the IDLE primitive include the static target pose or standby pose information.

[0086] Each action primitive is instantiated as an executable program or function in the gimbal sub-node 200, and when receiving the call instruction and parameters from the central control host 100, it can drive the gimbal to perform the corresponding motion.

[0087] The state machine director logic module 400 is deployed on the processor of the central control host 100 and loads the configuration from its storage unit; the state machine director logic module 400 abstracts the entire multi-camera collaborative shooting process into a deterministic finite state machine (DFA) for defining the behavior mode of the system under different shooting scenes and the switching logic between scenes.

[0088] S221: define the finite state machine model; the finite state machine can be formally described by a four-tuple:

[0089] ;

[0090] Wherein: It is a finite set of states that includes all predefined shooting states in the system; It is a finite set of events that includes all external inputs that can trigger a change in state; It is a state transition function, which defines the rules for the system to transition from one state to another; It is the initial state of the system, the first preset state that the system enters after startup.

[0091] S222: Define the set of states Each state in the system Corresponding to a specific shooting scene; a state By providing all of the system Each gimbal sub-node 200 is assigned a unique role for definition; the role is the action primitive that the gimbal sub-node 200 needs to invoke.

[0092] ;

[0093] in, Indicates the state Next, the Each cloud platform sub-node has 200 assigned roles; the set of roles is... When a gimbal child node is assigned the role of When it executes The primitive is responsible for the main output of the current screen; when a gimbal child node is assigned the role of When it executes The original text describes a gimbal sub-node responsible for tracking or preparing the next shooting target; when a gimbal sub-node is assigned the role of... When it executes The original phrase means "in a standby or reset state".

[0094] When a gimbal sub-node is assigned the role of At that time, this means in the state Below, the gimbal sub-node needs to execute a preset or dynamically planned motion path, typically used to achieve complex camera movement effects, such as continuous movement from an initial pose to a target pose, or shooting along a curved path; this motion path is defined by the corresponding... Primitives are defined, and their duration can be related to the state. Duration matching. For example, in a state defined as "host close-up". In the context, the role of gimbal sub-node A Possibly The role of gimbal sub-node B Possibly (Preparing to film the guests), while the roles of other gimbal sub-nodes C. may be ; in a state defined as "scene switching motion" , the role of the PTZ sub-node D may be , which will perform a pre-planned motion path to transition the shooting perspective.

[0095] S223: define a set of events ; each event in the system corresponds to an external trigger signal that can be recognized by the central control host 100; the source of the trigger signal can include: the physical button operation of the operator through the interactive interface (such as pressing the "switch to guest position" button); the expiration of the internal timer of the system; or the logical signal generated by other intelligent analysis modules (such as speech recognition or image recognition modules) (such as detecting that the host has finished speaking).

[0096] S224: define a logical transfer matrix ; state transition function The specific implementation in the system is a logical transfer matrix ; the matrix is stored in the storage unit of the central control host 100, and its data structure can be a two-dimensional array or a hash table; defines the core director logic of the system, which maps a combination of (current state, input event) to a determined target state.

[0097] ;

[0098] where, is the current state of the system, is the received external event, is the target state obtained by query; when an event occurs, the state machine director logic module 400 queries to determine the next state to which it should be transferred, and passes the target state information to the subsequent processing module. The matrix defines all expected state flow paths that meet the director's intentions.

[0099] The predictive verification engine 500 is one of the core logic modules of the central control host 100, and its function is to continuously run in the background to perform pre-motion simulation and geometric safety verification on the possible state transitions of the system; This mechanism ensures that all state switches have been evaluated for potential risks before execution, and provides input for subsequent dynamic path re-planning.

[0100] S311: Parallel simulation mechanism; predictive verification engine 500 continuously monitors the current system state of state machine broadcast logic module 400. The engine does not wait for external events to trigger it, but actively and in parallel simulates all potential future state transitions; specifically, it traverses the logical transition matrix. All from Triggered events and according to Determine all possible target states For each such potential transfer The engine will simulate in a virtual environment, simulating the actions that all gimbal sub-nodes 200 will perform during this transfer; this parallel simulation mechanism allows the system to grasp all possible risks before the actual switch occurs.

[0101] S312: Pure kinematic geometry anti-obfuscation calculation; This calculation aims to detect whether, during any potential state transition, a gimbal sub-node in the camera's field of view will be occluded by other moving gimbal sub-nodes, i.e., an "obfuscation" will occur.

[0102] S3121: Establish a three-dimensional coordinate system A globally unified three-dimensional world coordinate system is established within the central control unit 100. Precise three-dimensional position of all 200 gimbal sub-nodes in the system. And the real-time attitude (position and orientation) of its camera lens in space are precisely mapped to In the middle; these precise geometric data are the foundation for subsequent anti-fake calculations; real-time attitude data Each PTZ sub-node 200 reports to the central control host 100 at high speed via control link 310.

[0103] S3122: Identify ON_AIR and TRANSFER gimbals; for each state transition to be simulated... The predictive validation engine 500 first determines the target state. The definition identifies those who will bear The set of child nodes of the character's gimbal and will bear The set of child nodes of the character's gimbal ;in, The gimbal in the middle is the camera position that needs to maintain the integrity of the image. The gimbal in the image is a camera position where its movement could cause continuity errors.

[0104] S3123: Define the field of view cone ;for Each gimbal sub-node The field of view of the camera mounted on it can be abstracted as a field of view cone in three-dimensional space The field of view cone is determined by the exact position, orientation and optical parameters (such as focal length, field of view angle) of the camera lens in ; the zoom parameter of the camera lens determines the horizontal field of view angle and the vertical field of view angle ; The field of view cone can be described as a truncated pyramid defined by the camera optical center vertex and the clipping planes (near plane and far plane).

[0105] S3124: Define the motion envelope For each PTZ child node in , its physical shape occupies a space during the motion process, which is abstracted as a three-dimensional physical envelope ; the envelope is a boundary volume, such as an axis-aligned bounding box (AABB), an oriented bounding box (OBB), or a sphere, which moves along the simulated motion path in over time , and completely contains the physical entity of the PTZ child node , including the PTZ body, the camera and its accessories.

[0106] S3125: Collision detection; the predictive verification engine 500 traverses each PTZ child node in and each PTZ child node in , and continuously or discretely checks whether the motion envelope of intersects the field of view cone of during the entire motion time course ; if there is a time point such that the following conditions are met, it is determined that there is a risk of breaking through:

[0107] ;

[0108] where is the time variable, is the motion duration; is the motion envelope of the PTZ child node at time ; is the field of view cone of the PTZ child node ​the view frustum of ; ; ; ; ;

[0109] ; ;

[0110] ; ; ; ; ; ; ; ; ; ; ; ; ; ;

[0111] ; ; ; ; ; ; ; ;

[0112] ; ;

[0113] ; ; ;

[0114] ; ; ;

[0115] ; is a new motion path generated after optimization; is a candidate path in time ; is a cost function used to evaluate the quality of a path, which can include the total time of the path, the smoothness of the path (e.g. the integral of the third derivative of the path, i.e. the jerk ), or the degree of deviation from the original path; is the total time of the motion; is the motion envelope along the motion; is the field of view cone of ;represents the constraint that the intersection of the motion envelope and the field of view cone must be an empty set throughout the motion time course.

[0116] S323: specific implementation of the path generation algorithm; to solve the optimization problem described in S322 above, the present application can employ a variety of specific algorithms.

[0117] In one embodiment, the algorithm is a search-based algorithm, such as A* algorithm; the central control host 100 discretizes the pose motion space (e.g. a two-dimensional or three-dimensional space composed of the yaw angle and the pitch angle) of the PTZ sub-node , and constructs a configuration space graph; in this graph, all the pose configuration points or regions that are occluded are marked as “obstacles” or “high-cost regions”; the A* algorithm then searches for a path in this graph from to with the minimum total cost and without passing through the obstacle regions.

[0118] In another embodiment, the algorithm is a sampling-based algorithm, such as the RRT algorithm, which randomly samples in the configuration space and constructs a search tree until a path connecting and that satisfies the constraint condition is found.

[0119] In yet another embodiment, the algorithm is an interpolation-based algorithm, such as using a Bezier curve or a B-spline to parameterize the motion path ; the dynamic path replanning module 600 adjusts the control points of the curve through an optimization algorithm (e.g. gradient descent method) so that the generated curve can both satisfy the no-clip constraint in S322 (by imposing a high penalty when ) and minimize the cost function (e.g. maintain the smoothness of the curve).

[0120] The dynamic path replanning module 600 generates Ultimately, it was packaged into a new, revised version. Original language ( ) parameters (e.g., path point sequence or curve control points).

[0121] S324: Generate and cache the safe transition matrix The dynamic path replanning module 600 will calculate the safe path. (Alternatively, if no risk is detected in S321, the original method is used.) ), and the transfer Action primitives of all other gimbal child nodes involved (e.g.) or These can be combined to form a complete and securely verified state transition scheme.

[0122] This scheme is stored in a dedicated cache area within the central control host's 100 storage unit, called the secure transfer matrix. ; Data structures and The indexing method remains consistent, that is, through (current state) Input event Use the query function to perform the query.

[0123] ;

[0124] in, It is the target state. It is the first The securely verified action primitives that each cloud platform sub-node should execute during this transfer ( , , or the revised version ).

[0125] The predictive validation engine 500 runs continuously in the background, and when it detects... When changes occur, the data will be cleared or updated. The cache in the middle, and recalculate all from the new A safe transfer plan for departure, ensuring The system always caches all next steps for safe transfer based on the current system state.

[0126] The state transition readiness verification module 700, deployed on the processor of the central control host 100, is a key innovative module for ensuring the quality of the broadcast footage. This module determines the target state from the state machine broadcast logic module 400. It is activated after the actual execution of the state transfer. Its function is to actively query the internal readiness status of the target gimbal sub-node 200 and decide whether to immediately execute the transfer or enter standby based on the feedback.

[0127] S411: Ready-to-go check workflow; when the state machine broadcast logic module 400 is based on events... and current state The target state has been determined. At that time, the state transition readiness verification module 700 is invoked. This module first identifies the target state. China was designated as or The character's gimbal sub-nodes are identified, and these gimbal sub-nodes are marked as target camera positions.

[0128] The state transition readiness verification module 700 sends a specific readiness status query command CMD_CHECK_READY to each target station via control link 310; this command requests the PTZ sub-node 200 to report the readiness status of its internal system. .

[0129] After receiving the CMD_CHECK_READY command, the PTZ sub-node 200 will have its local processing unit check its critical subsystems and aggregate them into a comprehensive system readiness status. ; It contains one or more of the following sub-states:

[0130] S4111: Focus Lock Status; Whether the camera's autofocus system has successfully locked onto the target and reached the focus distance required for clear imaging.

[0131] S4112: Exposure stability status; whether the camera's aperture, shutter speed, ISO and other exposure parameters have been adjusted to appropriate values ​​and remain stable to avoid overexposure or underexposure of the image.

[0132] S4113: White balance calibration status; whether the camera's white balance has been accurately calibrated according to ambient light conditions.

[0133] S4114: Stable state of motion mechanism; whether the mechanical movement of the gimbal has stopped and is in a stable state, or is about to complete a smooth movement without obvious shaking or drift.

[0134] S4115: Internal self-test status; whether other key sensors, communication modules or storage units of the PTZ sub-node 200 are operating normally.

[0135] The PTZ sub-node 200 integrates its system readiness status. (For example, the Boolean value TRUE represents "Ready" or FALSE represents "Not-Ready") The state transition readiness verification module 700 of the central control host 100 is fed back through the control link 310.

[0136] S412: Automatic standby logic; the state transition readiness verification module 700 receives data from all target positions. After receiving feedback, a summary and judgment are made.

[0137] If all target positions If both values ​​are "Ready", the state transition readiness verification module 700 considers the system ready for a state transition; it will notify the state machine directing logic module 400 that execution can continue, and will retrieve the data from the safe transition matrix. Extracting targets The secure transfer solution is sent to the corresponding PTZ sub-node 200.

[0138] If any target position If the state is "Not-Ready", the state transition readiness verification module 700 will block the current state transition operation; at this time, the system remains in the current state. The state transition readiness verification module 700 then enters a loop query mode, which continuously sends the CMD_CHECK_READY instruction to the unready target station through the control link 310 at a preset frequency (e.g., every 100 milliseconds).

[0139] This loop will continue until feedback is received from the target location. The state transition readiness verification module 700 automatically triggers once all target stations have returned a "Ready" signal, allowing the state transition to proceed. The corresponding secure transfer plan is retrieved and sent.

[0140] This automatic standby logic ensures that the camera only switches shots or moves when all key camera positions are in optimal shooting condition. This effectively avoids quality problems such as blurry images, inaccurate exposure, or unstable movement caused by camera positions not being ready, thus improving the overall professionalism and safety of the broadcast.

[0141] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-position camera cloud control system, characterized in that, The application relates to a control system for a multi-rotatable camera platform, comprising: a central control host for managing and coordinating the operation of multiple pan-tilt sub-nodes; at least one pan-tilt sub-node for executing control instructions issued by the central control host; a dual-channel communication link for data exchange between the central control host and the at least one pan-tilt sub-node; the central control host comprises: a state machine director logic module for storing a preset system state set, an event set and a logical transition relationship between states, and deciding a target state according to a received external event; a predictive verification engine for continuously running in the background, acquiring real-time attitude data of all pan-tilt sub-nodes, and performing kinematic simulation and safety verification on all potential next state transitions in the state machine director logic module; a dynamic path re-planning module for working with the predictive verification engine, and generating a corrected safe motion path when the predictive verification engine predicts that a default path of state transition has a physical interference risk; a state transition readiness verification module for being called before state transition execution to query an internal readiness state of a target pan-tilt sub-node.

2. The multi-position photography cloud platform control system according to claim 1, wherein, the dual-channel communication link comprises: a control link for transmitting high-priority, low-delay control instructions, real-time attitude data of the pan-tilt sub-nodes and state response codes between the central control host and the pan-tilt sub-nodes; a data link for transmitting high-bandwidth media data streams between the central control host and the pan-tilt sub-nodes, the media data streams containing real-time preview video streams or material files.

3. The multi-position photography cloud platform control system according to claim 1, wherein, the state machine director logic module is based on a deterministic finite state machine model, and the finite state machine model comprises: a limited state set, wherein each state is determined by assigning one of ON_AIR, PREVIEW, TRANSFER or IDLE roles to all pan-tilt sub-nodes in the system; a limited event set for triggering state change; a logical transition matrix for defining conversion rules from a current state to a target state.

4. The multi-position photography cloud platform control system according to claim 1, wherein, the predictive verification engine performs parallel simulation and pure kinematic geometry anti-collision calculation when performing safety verification.

5. The multi-position photography cloud platform control system according to claim 4, wherein, the pure kinematic geometry anti-collision calculation comprises: establishing a global unified three-dimensional world coordinate system, and mapping accurate three-dimensional positions of all pan-tilt sub-nodes and real-time attitudes of camera lenses of the pan-tilt sub-nodes in space into the three-dimensional world coordinate system; identifying a set of pan-tilt sub-nodes assuming ON_AIR roles and a set of pan-tilt sub-nodes assuming TRANSFER roles in state transition; defining a field of view range of each pan-tilt sub-node camera assuming ON_AIR roles as a field of view cone; defining a space occupied by a physical shape of each pan-tilt sub-node assuming TRANSFER roles in a motion process as a motion envelope; performing collision detection to judge whether the motion envelope geometrically intersects with the field of view cone, and determining that there is a risk of anti-collision if the two intersect.

6. The multi-position photography cloud platform control system according to claim 5, wherein, The field of view pyramid is determined by the accurate position, orientation and optical parameters of the lens of the camera node assuming the role of ON_AIR in the three-dimensional world coordinate system; the motion envelope is a boundary volume that moves along the simulated motion path of the pan-tilt node assuming the role of TRANSFER and completely contains the physical entity of the pan-tilt node.

7. The multi-position photography cloud control system according to claim 1, wherein, The modified safe motion path generated by the dynamic path re-planning module aims to generate a new motion path for the pan-tilt node from the initial pose to the target pose, and the new motion path satisfies the no-through constraint throughout the motion schedule.

8. The multi-position photography cloud control system according to claim 7, wherein, The dynamic path re-planning module generates the modified safe motion path by using a search-based algorithm, a sampling-based algorithm or an interpolation-based algorithm.

9. The multi-position photography cloud control system according to claim 2, wherein, The state transition readiness verification module: sends a readiness state query instruction to the target pan-tilt node through the control link to request the target pan-tilt node to report the internal readiness state; when the internal readiness state is not ready, prevents state transition and continues to poll until the internal readiness state becomes ready.

10. The multi-position photography cloud platform control system according to claim 9, wherein, The internal readiness state includes at least one of the focus lock state, the exposure stability state, the white balance calibration state or the motion mechanism stability state.

Images