Mobile control method of intelligent podium and intelligent podium

Abstract

The present invention provides a mobile control method of an intelligent podium and an intelligent podium, and relates to the field of automatic control. The method comprises: inputting a facial image into a preset face recognition model to obtain target personnel information; when the target personnel information is non-inside personnel information, determining the distance between the non-inside personnel and the intelligent podium, and determining the non-inside personnel with the smallest distance as the target personnel; determining the target distance between the target personnel and the intelligent podium according to the focal length of the camera, the first distance and the second distance, and the target spacing between the first camera and the second camera, and determining the actual height according to the target distance, the first distance, the focal length of the camera, the target spacing, and the height of the second camera from the ground; and adjusting the actual height of the intelligent podium according to the actual height. In the case of facing multiple speakers, the present invention flexibly utilizes the selection tendency of whether it is an internal staff or a non-inside staff, and cooperates with the distance of the speaker from the podium to achieve accurate determination of the target personnel.

Description

Technical Field

[0001] The present invention relates to the field of automatic control, and in particular to a mobile control method of an intelligent podium and the intelligent podium. Background Art

[0002] Existing podiums mostly focus on meeting the basic needs of lectures and teaching, with fixed heights, positions, and functions, often limited. While these meet the most basic lecture needs, they also present numerous challenges. For example, height requirements vary, yet adjustments are difficult to make. On some stages, the podium may simply be a minor feature. Traditional podiums and some multimedia platforms are fixed and difficult to move, hindering both stage layout and flow.

[0003] Related technologies have attempted to successfully integrate artificial intelligence technology and semi-automation technology into podium products, but they still have obvious limitations in height adjustment, azimuth movement, etc., and are not widely used. For example, when there are multiple identification targets on the podium, it is impossible to determine which target task to use to adjust the height of the podium. For another example, when the speaker on the podium moves, the podium cannot follow the movement. At this time, the microphone or manuscript placed on the podium will limit the movement range of the speaker and reduce the speaker's speaking experience.

[0004] However, there is currently no technical solution that can solve the above technical problems, and there is no mobile control method and intelligent podium. Summary of the Invention

[0005] The present invention provides a mobile control method for an intelligent podium and an intelligent podium, which solves the recognition confusion problem when there are multiple faces on the podium and can adjust the movement of the intelligent podium according to the height of the target person.

[0006] In a first aspect, the present invention provides a method for controlling the movement of an intelligent podium, comprising:

[0007] Identify a facial image facing the direction of the smart podium, input the facial image into a preset face recognition model, and obtain target person information output by the preset face recognition model, where the target person information is internal person information and / or non-internal person information;

[0008] When it is determined that the target person information is non-inside person information, determining the distance between each non-inside person and the smart podium, and determining the non-inside person with the shortest distance as the target person;

[0009] Determine the target distance between the target person and the smart podium based on the camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera; determine the actual height of the target person based on the target distance, the first distance, the camera focal length, the target distance, and the height of the second camera from the ground; the first distance is determined based on the vertical distance from the top of the target person's head to the optical center of the first camera and the first camera; the second distance is determined based on the vertical distance from the top of the target person's head to the optical center of the second camera and the second camera; the first camera and the second camera are arranged in the vertical direction of the smart podium;

[0010] Adjust the actual height of the smart podium according to the actual height of the target person;

[0011] The preset face recognition model is determined after training based on different sample face images and face labels corresponding to each sample face image, wherein the face labels include insider labels and non-insider labels.

[0012] According to the mobile control method of the smart podium provided by the present invention, after obtaining the target person information output by the preset face recognition model, the method further includes:

[0013] When it is determined that the target person information includes insider information and non-insider information, and when it is determined that there is more than one insider information, determining the distance between each insider and the smart podium, and determining the insider with the shortest distance as the target person;

[0014] When it is determined that the target person information includes insider information and non-insider information, and when it is determined that there is only one insider information, the insider is determined to be the target person;

[0015] According to the target person, the actual height of the target person is determined from a preset height database; the preset height database is constructed based on the corresponding relationship between all internal personnel and the height of each internal person.

[0016] According to the mobile control method of the smart podium provided by the present invention, determining the target distance between the target person and the smart podium based on the camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera includes:

[0017]

[0018] Among them, d is the target distance between the target person and the smart podium, f is the focal length of the camera, L is the target distance between the first camera and the second camera, x is the first distance, and y is the second distance.

[0019] According to the mobile control method of the smart podium provided by the present invention, determining the actual height of the target person based on the target distance, the first distance, the camera focal length, the target spacing, and the height of the second camera from the ground includes:

[0020]

[0021] Among them, H is the actual height of the target person, x is the first distance, f is the focal length of the camera, d is the target distance between the target person and the smart podium, L is the target distance between the first camera and the second camera, and h is the height of the second camera from the ground.

[0022] According to the mobile control method of the smart podium provided by the present invention, before determining the target distance between the target person and the smart podium based on the camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera, the method further includes:

[0023] When it is determined that the target person deviates from the center of the smart podium, determining an image distance of the target person deviating from the center;

[0024] Determine the calibrated camera focal length according to the image distance and the camera focal length, and determine the target distance between the target person and the smart podium according to the calibrated camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera;

[0025] The image distance of the target person deviating from the center is determined based on the distance between the first intersection and the second intersection. The first intersection is the intersection of the center line of the smart podium and the vertical line of the focal position of any camera, and the second intersection is the intersection of the deviation line formed by the target person and any camera and the vertical line of the focal position of the camera.

[0026] According to the mobile control method of the smart podium provided by the present invention, determining the focal length of the calibrated camera based on the image distance and the focal length of the camera includes:

[0027]

[0028] Wherein, z is the image distance of the target person off-center, f is the focal length of the camera, and f′ is the focal length of the camera after calibration.

[0029] According to the movement control method of the smart podium provided by the present invention, after adjusting the actual height of the smart podium, the method further includes:

[0030] Obtaining the real-time coordinates of the target person in a preset three-dimensional coordinate system, where the preset three-dimensional coordinate system is constructed based on the camera's center position along the smart podium as the Y-axis and the width direction of the smart podium as the X-axis;

[0031] When the target person is within the normal range of activity, the smart podium remains stationary;

[0032] The normal range of activity is:

[0033]

[0034] Wherein, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, and W is the preset vertical movement length.

[0035] According to the mobile control method of the smart podium provided by the present invention, after obtaining the real-time coordinates of the target person in the preset three-dimensional coordinate system, the method further includes:

[0036] When the target person is located in the activity range on both sides of the middle, the smart podium is driven to move laterally, and the activity range on both sides of the middle includes the left middle activity range and the right middle activity range;

[0037] The range of movement on both sides of the middle part is:

[0038]

[0039] Among them, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, W is the preset vertical movement length, M is the preset front activity length, and D is the depth of the smart podium.

[0040] According to the mobile control method of the intelligent podium provided by the present invention, the activity range of the target person includes the left front activity range, the right front activity range, the left middle activity range, the right middle activity range, the left rear activity range, the right rear activity range, the left invalid area range, the right invalid area range and the normal activity range;

[0041] When the target person is located in the left front activity range, the right front activity range, the left rear activity range, or the right rear activity range, the smart podium is driven to turn and move;

[0042] When the target person is located in the left invalid area or the right invalid area, the smart podium is adjusted to a preset initial state.

[0043] In a second aspect, a smart podium is provided, comprising: a podium body, a motor-controlled moving device, a binocular camera, and a control processor of the smart podium;

[0044] The podium body is fixedly connected to the motor-controlled moving device; the binocular camera includes a first camera and a second camera, which are respectively arranged at different heights of the podium body along the vertical direction;

[0045] The control processor is fixed on the podium body and is communicatively connected to the motor-controlled mobile device and the binocular camera; it also includes a memory and a program or instruction stored in the memory and capable of running on the control processor of the smart podium. When the program or instruction is executed by the control processor of the smart podium, the mobile control method of the smart podium is executed.

[0046] The present invention provides a mobile control method and a smart podium. By automatically identifying and locking the speaker's facial image, the smart podium can quickly respond and adjust to the optimal height and position suitable for the speaker without manual adjustment, greatly improving the convenience of the speech. The present invention solves the problem of flexibly utilizing the selection tendency of whether the speaker is an internal person or a non-internal person in combination with the distance between the speaker and the podium to achieve accurate determination of the target person. At the same time, the automatic adjustment function of the smart podium can ensure that the speaker always remains within the optimal viewing angle and range during the speech, thereby improving the overall effect of the speech. The present invention not only supports the identification of internal persons, but also can identify non-internal persons, and intelligently adjust the height and position of the podium according to the identification results. This flexibility enables the smart podium to adapt to different usage scenarios and user needs, thereby enhancing the user experience.

[0047] The present invention not only realizes automatic adjustment of height and position, but also ensures that the speaker always remains within the optimal shooting range through the automatic positioning and following function. This comprehensive intelligent function enables the smart podium to adapt to more usage scenarios and needs, improving its adaptability and flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 It is a flow chart of the mobile control method of the intelligent podium provided by the present invention;

[0050] Figure 2 This is a schematic diagram of binocular height recognition provided by the present invention;

[0051] Figure 3This is a schematic diagram of a situation where a target deviates from the center line of a camera provided by the present invention;

[0052] Figure 4 It is a schematic diagram of the platform coordinate system provided by the present invention;

[0053] Figure 5 This is a schematic diagram of angle correction provided by the present invention;

[0054] Figure 6 It is a structural diagram of the intelligent podium provided by the present invention. DETAILED DESCRIPTION

[0055] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0056] The present invention discloses a multifunctional, integrated intelligent podium designed using computer vision and automatic control technologies. The podium features automated height adjustment and automatic follow-up movement, enhancing the effectiveness and convenience of presentations. For automated height adjustment, the system employs binocular cameras and a convolutional neural network to perform convolution and pooling operations on input data, extracting features for learning and classification. The system then compares the features with a face database for facial recognition, enabling low-cost height calculations in multiple scenarios without requiring full-body height measurements. A motor-controlled screw-type lifting mechanism also enables segmented height adjustment. For automatic follow-up movement, the system employs a motor-controlled moving device and binocular cameras to establish a coordinate system with the podium itself as a reference and locate the user. It then recommends movement methods based on the coordinates' location. The system calculates the podium's angular deviation during steering based on the linear speed and operating time of the power wheels during differential operation. Finally, the two power wheels operate in opposite directions to achieve angle correction. The podium itself moves to maintain the user's coordinates within the appropriate area, achieving highly stable 360-degree automatic follow-up movement and podium posture correction.

[0057] Figure 1 : is a flow chart of the mobile control method of the intelligent podium provided by the present invention, the mobile control method of the intelligent podium includes:

[0058] Step 101: Identify a facial image facing the smart podium, input the facial image into a preset face recognition model, and obtain target person information output by the preset face recognition model, where the target person information is internal personnel information and / or non-internal personnel information;

[0059] Step 102: If the target person information is determined to be non-inside person information, determine the distance between each non-inside person and the smart podium, and determine the non-inside person with the shortest distance as the target person;

[0060] Step 103: Determine the target distance between the target person and the smart podium based on the camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera. Determine the actual height of the target person based on the target distance, the first distance, the camera focal length, the target distance, and the height of the second camera from the ground. The first distance is determined based on the vertical distance from the top of the target person's head to the optical center of the first camera and the optical center of the first camera. The second distance is determined based on the vertical distance from the top of the target person's head to the optical center of the second camera and the optical center of the second camera. The first camera and the second camera are arranged in the vertical direction of the smart podium.

[0061] Step 104: Adjust the actual height of the smart podium according to the actual height of the target person;

[0062] The preset face recognition model is determined after training based on different sample face images and face labels corresponding to each sample face image, wherein the face labels include insider labels and non-insider labels.

[0063] Optionally, the present invention consists of two main parts, including an automatic height adjustment control method and an automatic positioning and following control method. In the automatic height adjustment control method, a visual system is used to judge and lock the target person, and an innovative height calculation method based on a binocular camera is used to quickly obtain height data. According to the height data, a feedback control system is used in conjunction with the visual system to automatically adjust the podium height; in the automatic positioning and following control method, a visual system is used for coordinate positioning, and a following algorithm is used to drive the podium based on the positioning results.

[0064] In step 101, a facial image of a person facing the smart podium is first obtained. After this image is captured by a camera installed on the podium, it is input into a preset facial recognition model for analysis. The facial recognition model uses a machine learning algorithm to detect and identify facial features and match them with previously trained samples. Specifically, the preset facial recognition model is determined after training based on different sample facial images and the facial labels corresponding to each sample facial image. The facial labels include internal person labels and non-inside person labels. In an optional implementation, assume that a speaker is facing the podium, and the camera captures a clear facial image. Through model analysis, it is confirmed that the speaker is an internal person.

[0065] Optionally, the facial image facing the direction of the smart podium can be a facial image of a speaker, or can be multiple facial images facing the direction of the smart podium. In this case,

[0066] A preset face recognition model is used to obtain all target person information output by the preset face recognition model, including internal person information and / or non-internal person information.

[0067] Optionally, when the smart podium is operating, in default mode, automatic adjustment begins as soon as a face is detected. This mode only involves face detection. If the podium is connected to a facial database, adjustment will only begin after the podium's visual module recognizes an internal user. In this mode, after detecting a face, the smart podium captures the facial image and extracts the feature vector. It then uses a convolutional neural network (CNN) for recognition and compares it with the connected database to complete the identity verification process. Smart podiums need to adapt to different working environments, and lighting and occlusion are detrimental to facial recognition. Using convolutional neural networks can improve recognition accuracy and stability. Furthermore, facial recognition involves user privacy to a certain extent. In database-connected mode, facial information is not directly stored on the podium, meaning there are no information management issues. However, the management of facial information collected in the default mode requires proper storage and regular cleaning to prevent the leakage of user facial images and protect user privacy.

[0068] In step 102, in this embodiment, the target person must first be identified as a non-inside person, and the number of non-inside persons is greater than 1. At this time, the distance between each non-inside person and the smart podium will be calculated, and the non-inside person with the smallest distance will be determined as the target person. For example, if there are two speakers with distances of 3 meters and 5 meters from the podium respectively, the system will select the speaker with a distance of 3 meters as the target person. This step ensures that the system responds to the nearest target, optimizes the movement and adjustment strategy of the podium, improves the recognition strategy and improves the recognition accuracy.

[0069] Optionally, after obtaining the target person information output by the preset face recognition model, the method further includes:

[0070] When it is determined that the target person information includes insider information and non-insider information, and when it is determined that there is more than one insider information, determining the distance between each insider and the smart podium, and determining the insider with the shortest distance as the target person;

[0071] When it is determined that the target person information includes insider information and non-insider information, and when it is determined that there is only one insider information, the insider is determined to be the target person;

[0072] According to the target person, the actual height of the target person is determined from a preset height database; the preset height database is constructed based on the corresponding relationship between all internal personnel and the height of each internal person.

[0073] Optionally, when the target person information contains multiple insiders and at least one non-inside person, the non-inside persons are first eliminated, and then among all the insiders, the distance between each insider and the podium is identified, and the insider with the shortest distance is determined as the target person; if it is determined that the target person information includes insider information and non-inside person information, and it is determined that there is only one insider information, then the insider is preferentially locked as the target person, and then when it is determined that the target person is an insider, the preset height database can be directly used to realize the determination of the actual height of the target person. This application constructs the preset height database in advance according to the corresponding relationship between the heights of all insiders and each insider, so that the actual height only needs to be calculated when the non-inside person determines his height. When the insider determines his height, the calculation step is omitted and the height data is directly obtained, thereby improving the execution response speed and lifting speed of the smart podium. In the subsequent operation of the podium, as long as the locked person does not exit the use range, the podium will not readjust the height due to detection of other people.

[0074] In step 103, based on the focal length of the camera and the distance information to the target person, the target distance between the target person and the podium is determined. Then, in combination with the geometric relationship, the first distance (determined based on the vertical distance from the top of the target person's head to the first camera and the optical center of the first camera) and the second distance (determined based on the vertical distance from the top of the target person's head to the second camera and the optical center of the second camera) are used to calculate the actual height of the target person. Step 103 improves the adaptability of the smart podium through precise calculation, ensures that the podium height can accurately correspond to the height of each target person, and enhances the user experience.

[0075] Figure 2 This is a schematic diagram of binocular height recognition provided by the present invention, such as Figure 2 As shown, the first camera and the second camera are arranged in the vertical direction of the smart podium, and the target distance between the target person and the smart podium is determined according to the camera focal length, the first distance and the second distance, and the target distance between the first camera and the second camera, including:

[0076] (1)

[0077] Among them, d is the target distance between the target person and the smart podium, f is the focal length of the camera, L is the target distance between the first camera and the second camera, x is the first distance, and y is the second distance.

[0078] Optionally, determining the actual height of the target person according to the target distance, the first distance, the focal length of the camera, the target spacing, and the height of the second camera from the ground includes:

[0079] (2)

[0080] Among them, H is the actual height of the target person, x is the first distance, f is the focal length of the camera, d is the target distance between the target person and the smart podium, L is the target distance between the first camera and the second camera, and h is the height of the second camera from the ground.

[0081] The premise of the present invention's response control is to identify the user's highest point, that is, the top of the user's head. During the calculation process, it is necessary to analyze the known quantities such as the distance from the highest point to the center of the image by receiving data transmitted by the visual module. The present invention intends to measure the approximate height of the target through a binocular system. The binocular cameras are installed on the same vertical line as the podium. The target is located in the center of the camera. The height h of camera 2 from the ground can be directly read from the height variable. The spacing L between the cameras is fixed and is a known quantity. The focal length f of the camera is known. The distances x and y from the top of the target's head to the optical centers of the two cameras are obtained by image analysis. In the figure, the camera imaging plane is rotated 180°. Formula (1) can be listed to solve the distance d between the target and the podium. The target's height can be substituted into formula (2) to calculate the height H.

[0082] Figure 3 is a schematic diagram of a situation where a target deviates from the center line of a camera provided by the present invention. Before determining the target distance between the target person and the smart podium based on the camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera, the method further includes:

[0083] When it is determined that the target person deviates from the center of the smart podium, determining an image distance of the target person deviating from the center;

[0084] Determine the calibrated camera focal length according to the image distance and the camera focal length, and determine the target distance between the target person and the smart podium according to the calibrated camera focal length, the first distance, the second distance, and the target distance between the first camera and the second camera;

[0085] The image distance of the target person deviating from the center is determined based on the distance between the first intersection and the second intersection. The first intersection is the intersection of the center line of the smart podium and the vertical line of the focal position of any camera, and the second intersection is the intersection of the deviation line formed by the target person and any camera and the vertical line of the focal position of the camera.

[0086] Optional, since in reality the speaker is not necessarily in front of the podium, the simulation situation is as follows Figure 3 As shown, if the target is not on the center line, the data calculated by the above algorithm is incorrect and needs to be corrected. Since the error mainly exists in the substituted focal length f, the focal length of the camera after calibration is determined based on the image distance and the camera focal length, including:

[0087] (3)

[0088] Wherein, z is the image distance of the target person off-center, f is the focal length of the camera, and f′ is the focal length of the camera after calibration.

[0089] Furthermore, we obtain the corrected f′, where z is the distance of the speaker from the center of the image. The value of z is given by image analysis, just like the variables x and y mentioned above. Substitute f′ into f in equation (2) and calculate H. The rest of the process is the same.

[0090] In step 104, the present invention adjusts the actual height of the smart podium based on the actual height of the target person. After calculating the actual height of the target person, the most suitable height of the podium can be obtained. The operation of the lifting device can be controlled by continuously reading the current podium height and comparing it with the target height. The height adjustment of the present invention does not need to be too precise. In actual use, the change in the lecture experience brought about by adjustments within 1 cm can be almost considered zero. The greater the precision, the greater the development cost. Therefore, the automatic height adjustment method of the smart podium is designed to be "segmented", that is, each 1 cm is a level, and the lifting device drives the podium to adjust to the range where the target height falls.

[0091] Optionally, the present invention intends to use a motor-controlled screw lifting device to convert rotational motion into linear motion while transmitting power to the vertical direction to achieve height adjustment. Compared with hydraulic presses, it does not have a large amount of oil and is easier to adapt to various environments, and fault maintenance is also more convenient. Through programmed control of the motor, the rotation direction of the motor is changed to achieve raising or lowering of the podium, and the appropriate motor rotation speed and time are controlled to achieve gear adjustment. The power provided by the motor is stable, which can ensure that the lifting device runs smoothly when adjusting the height, and is not prone to shaking or jittering, which is beneficial to protecting the equipment, has high reliability and durability, and can operate stably for a long time.

[0092] Furthermore, to meet dynamic demands during use, the Smart Podium can automatically fine-tune itself based on the speaker's posture. For example, a user's posture often changes during use, and the resulting height fluctuation is generally within 5cm. The Smart Podium repeats the height calculation process every 30 seconds and compares it with the original data. Any deviation exceeding 5cm is considered an error, and the system rechecks after 10 seconds. Changes within 5cm are adjusted based on the latest height data.

[0093] The present invention provides a mobile control method and a smart podium. By automatically identifying and locking the speaker's facial image, the smart podium can quickly respond and adjust to the optimal height and position suitable for the speaker without manual adjustment, greatly improving the convenience of the speech. The present invention solves the problem of flexibly utilizing the selection tendency of whether the speaker is an internal person or a non-internal person in combination with the distance between the speaker and the podium to achieve accurate determination of the target person. At the same time, the automatic adjustment function of the smart podium can ensure that the speaker always remains within the optimal viewing angle and range during the speech, thereby improving the overall effect of the speech. The present invention not only supports the identification of internal persons, but also can identify non-internal persons, and intelligently adjust the height and position of the podium according to the identification results. This flexibility enables the smart podium to adapt to different usage scenarios and user needs, thereby enhancing the user experience.

[0094] The present invention not only realizes automatic adjustment of height and position, but also ensures that the speaker always remains within the optimal shooting range through the automatic positioning and following function. This comprehensive intelligent function enables the smart podium to adapt to more usage scenarios and needs, improving its adaptability and flexibility.

[0095] Figure 4 This is a schematic diagram of the platform coordinate system provided by the present invention, such as Figure 4 As shown in the figure, the smart podium realizes user positioning based on the visual system. The intermediate quantity obtained by height calculation includes the distance between the user and the podium and the distance from the center line. The former is y and the latter is x. A two-dimensional coordinate system with the podium as the reference can be constructed. Once the user's coordinates (x, y) are obtained, the positioning is completed.

[0096] Optionally, after adjusting the actual height of the smart podium, the method further includes:

[0097] Obtaining the real-time coordinates of the target person in a preset three-dimensional coordinate system, where the preset three-dimensional coordinate system is constructed based on the camera's center position along the smart podium as the Y-axis and the width direction of the smart podium as the X-axis;

[0098] When the target person is within the normal range of activity, the smart podium remains stationary;

[0099] The normal range of activity is:

[0100] (4)

[0101] Wherein, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, and W is the preset vertical movement length.

[0102] Optionally, after positioning is completed, it is determined how to drive the mobile device according to the coordinates. Figure 4 , the y>0 plane is divided into eight areas in the coordinate system. The activity range of the target person includes the left front activity range, the right front activity range, the left middle activity range, the right middle activity range, the left rear activity range, the right rear activity range, the left invalid area range, the right invalid area range and the normal activity range, where θ is the maximum shooting angle of the camera.

[0103] If the user is within the normal range of activity, the platform considers that there is no need to adjust the posture and position. At this time, the coordinates should satisfy formula (4). After obtaining the real-time coordinates of the target person in the preset three-dimensional coordinate system, the method further includes:

[0104] When the target person is located in the activity range on both sides of the middle, the smart podium is driven to move laterally, and the activity range on both sides of the middle includes the left middle activity range and the right middle activity range;

[0105] The range of movement on both sides of the middle part is:

[0106] (5)

[0107] Among them, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, W is the preset vertical movement length, M is the preset front activity length, and D is the depth of the smart podium.

[0108] Optionally, the left and right horizontal movement areas can achieve horizontal lateral movement of the podium. For example, if the coordinates satisfy Equation (5), since it is in the "horizontal right movement" area, the moving device drives the active wheels to work in the same direction and speed to achieve horizontal right movement. During the movement process, the user's new position is continuously located until the user coordinates are within the normal range of movement and the movement stops. The same applies to the "horizontal left movement" area.

[0109] Optionally, the target person's activity range includes a left front activity range, a right front activity range, a left middle activity range, a right middle activity range, a left rear activity range, a right rear activity range, a left invalid area range, a right invalid area range, and a normal activity range;

[0110] When the target person is within the left front activity range, right front activity range, left rear activity range, or right rear activity range, drive the intelligent podium to turn and move;

[0111] When the target person is within the left invalid area range or right invalid area range, adjust the intelligent podium to the preset initial state.

[0112] Figure 5 This is the schematic diagram of angle correction provided by the present invention. In the area involving turning, taking "left rear movement" as an example, after positioning that the user is within the left rear activity range, drive the two driving wheels to work with differential speed in the same direction. To achieve backward movement, the rotational speed of the front wheels should be greater than that of the rear wheels. Continuously position the user's coordinates during the movement. If the user is at the edge of the invalid area, reduce the rotational speed of the front wheels, aiming to always keep the user within the recognizable range. Once the user's coordinate y < W, enter the "horizontal left shift" area, and only horizontally shift to make the user enter the normal activity range. Different from the previous two movement parts, these four situations after turning require angle correction after the user enters the normal range. When the two wheels work with differential speed in the same direction, assume the linear velocity of the fast wheel is v1, the linear velocity of the slow wheel is v2, and the deviation angle is β. Work in this mode for time t. It can be obtained from the formula:

[0113] (6)

[0114] Obtain the deviation angle β. After the user enters the normal area, the two wheels work in the opposite direction at speed v0 for time t0, so that:

[0115] (7)

[0116] Thus, eliminate the angle deviation caused by turning movement and correct the posture of the podium.

[0117] When the target person is within the left invalid area range or right invalid area range, adjust the intelligent podium to the preset initial state. The invalid area is the area that the camera cannot detect. If the user is in this area, the vision system cannot recognize and lock, and the mobile device will not be started.

[0118] Computer vision technology and automatic control technology are the core technologies for the intelligent podium to achieve automatic height adjustment and positioning tracking. The realization of the automatic height adjustment function is divided into three steps: target recognition, height calculation, and control adjustment. The target recognition link is the key to starting the whole function. As long as a human face is detected, the automatic adjustment is started. Computer vision technology uses binocular cameras and image processing algorithms to collect facial images and extract feature vectors, and uses convolutional neural networks for recognition, thus starting the automatic adjustment.

[0119] The measurement of height data of the present invention involves the use of a binocular camera system to calculate the height of the target by measuring the distance from the top of the target's head to the camera and the known distance between the cameras, combined with the focal length and installation height of the camera. Automatic control technology focuses on control and adjustment to ensure the precise coordination of motor control and mechanical transmission, and to achieve height fine-tuning through a screw lifting device. The implementation of the positioning and tracking function is divided into two steps: coordinate positioning and real-time tracking. Based on the intermediate quantity obtained by the height calculation module in computer vision technology, the two-dimensional coordinates of the speaker relative to the podium are obtained. Automatic control technology is responsible for the decision-making of movement strategy: controlling the active wheel or power wheel to achieve horizontal movement or steering of the podium; adjusting the wheel speed to achieve precise movement control and angle correction to eliminate deviation; and adjusting the position and posture of the podium in real time through a feedback mechanism to ensure that the speaker always remains within the optimal shooting range.

[0120] This invention provides a method for controlling the movement of an intelligent podium. This method addresses the problem that existing podiums are fixed and limited in height and position, resulting in inconsistent height requirements for different groups of people and an inability to adjust them in a timely manner. This problem is also unfriendly to scene layout and process flow. This invention adds height adjustment and free movement functions to the traditional podium. Manual height adjustment is achieved by controlling a segmented motor-controlled screw lifting device through a mobile phone or interactive interface. Facial recognition is used to confirm the target activation function.

[0121] The present invention uses a binocular camera to identify the highest point of the target, calculates the height based on the focal length and the fixed height of the podium, and confirms the height range of the podium. The height is adjusted within the corresponding range through a screw lifting device controlled by a segmented motor, realizing the automatic height adjustment function. Through the "six-wheel" design, the motor controls the forward and reverse rotation and the speed to realize the mobility of the podium. The motor is controlled through the podium touch screen or mobile phone to realize manual control movement. The intermediate data of the height calculation is used to confirm the user coordinates. The specified area is divided into eight blocks according to the maximum shooting angle of the camera. Based on the position of different blocks, the podium is moved differently to ensure that the user is in the normal block. The deviation angle β calculated by the wheel speed is used to eliminate the rotation deviation and realize positioning and following movement. Compared with the existing technology, the use of computer vision technology realizes the height adjustment and free movement functions, increases the flexibility of the podium, solves the existing problems and makes the podium adaptable to more scenarios.

[0122] Figure 6 Schematic diagram of the structure of the intelligent podium provided by the present invention, the intelligent podium comprises: a podium body 1, a motor-controlled mobile device 2, a binocular camera 3 and a control processor 4 of the intelligent podium;

[0123] The podium body 1 is fixedly connected to the motor-controlled mobile device 2; the binocular camera 3 includes a first camera and a second camera, which are respectively arranged at different heights of the podium body 1 along the vertical direction;

[0124] The control processor 4 is fixed on the podium body 1 and is communicatively connected with the motor-controlled mobile device 2 and the binocular camera 3; it also includes a memory and a program or instruction stored in the memory and capable of running on the control processor 4 of the smart podium. When the program or instruction is executed by the control processor 4 of the smart podium, the mobile control method of the smart podium is executed.

[0125] The present invention provides a smart podium that automatically recognizes and locks onto the speaker's facial image, allowing the smart podium to quickly respond and adjust to the optimal height and position for the speaker without manual adjustment, greatly improving the convenience of the speech. The present invention solves the problem of flexibly utilizing the selection preference of internal or non-internal personnel in the case of multiple speakers, and accurately determining the target person in combination with the distance between the speaker and the podium. At the same time, the automatic adjustment function of the smart podium can ensure that the speaker always remains within the optimal viewing angle and range during the speech, thereby improving the overall effect of the speech. The present invention not only supports the recognition of internal personnel, but also non-internal personnel, and intelligently adjusts the height and position of the podium based on the recognition results. This flexibility enables the smart podium to adapt to different usage scenarios and user needs, thereby enhancing the user experience.

[0126] This invention not only enables automatic height and position adjustment, but also ensures that the speaker always remains within the optimal shooting range through automatic positioning and tracking. This comprehensive intelligent functionality enables the smart podium to adapt to a wider range of usage scenarios and needs, improving its adaptability and flexibility. This invention can optimize the speaking experience for different audiences and scenarios, and achieves automatic height adjustment and automatic position tracking functions.

[0127] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A mobile control method for an intelligent podium, characterized in that: include: Identify a facial image facing the direction of the smart podium, input the facial image into a preset face recognition model, and obtain target person information output by the preset face recognition model, wherein the target person information is internal person information and / or non-internal person information; In the case where it is determined that the target person information is non-inside person information, determining the distance between each non-inside person and the smart podium, and determining the non-inside person with the shortest distance as the target person; The target distance between the target person and the smart podium is determined according to the focal length of the camera, the first distance, the second distance, and the target distance between the first camera and the second camera. The actual height of the target person is determined according to the target distance, the first distance, the focal length of the camera, the target distance, and the height of the second camera from the ground. The first distance is determined according to the distance from the top of the target person's head to the first camera and the optical center of the first camera in the vertical direction. The second distance is determined according to the distance from the top of the target person's head to the second camera and the optical center of the second camera in the vertical direction. The first camera and the second camera are arranged in the vertical direction of the smart podium. According to the actual height of the target person, adjust the actual height of the smart podium; The preset face recognition model is determined after training based on different sample face images and face labels corresponding to each sample face image, wherein the face labels include internal staff labels and non-internal staff labels; Determining the target distance between the target person and the smart podium according to the focal length of the camera, the first distance, the second distance, and the target distance between the first camera and the second camera includes: ； Among them, d is the target distance between the target person and the smart podium, f is the focal length of the camera, L is the target distance between the first camera and the second camera, x is the first distance, and y is the second distance.

2. The mobile control method of the intelligent podium according to claim 1, characterized in that: After obtaining the target person information output by the preset face recognition model, the method further includes: When it is determined that the target person information includes the insider information and the non-insider information, and when it is determined that there is more than one insider information, the distance between each insider and the smart platform is determined, and the insider with the shortest distance is determined as the target person; When it is determined that the target person information includes the insider information and the non-insider information, and when it is determined that the insider information is one, determining the insider as the target person; According to the target person, the actual height of the target person is determined from a preset height database; the preset height database is constructed based on the corresponding relationship between all internal personnel and the height of each internal person.

3. The mobile control method of the intelligent podium according to claim 1, characterized in that: Determining the actual height of the target person according to the target distance, the first distance, the focal length of the camera, the target spacing, and the height of the second camera from the ground includes: ； Among them, H is the actual height of the target person, x is the first distance, f is the focal length of the camera, d is the target distance between the target person and the smart podium, L is the target distance between the first camera and the second camera, and h is the height of the second camera from the ground.

4. The mobile control method of the intelligent podium according to claim 1, characterized in that: Before determining the target distance between the target person and the smart podium according to the focal length of the camera, the first distance, the second distance, and the target distance between the first camera and the second camera, the method further includes: In the case where it is determined that the target person deviates from the center of the smart podium, determining an image distance of the target person deviating from the center; Determine the focal length of the calibrated camera according to the image distance and the focal length of the camera, and determine the target distance between the target person and the smart podium according to the focal length of the calibrated camera, the first distance and the second distance, and the target distance between the first camera and the second camera; The image distance of the target person deviating from the center is determined based on the distance between the first intersection and the second intersection, the first intersection being the intersection of the center line of the smart podium and the vertical line of any camera focal position, and the second intersection being the intersection of the deviation line formed by the target person and any camera and the vertical line of the camera focal position.

5. The mobile control method of the intelligent podium according to claim 4, characterized in that: The determining the focal length of the camera after calibration according to the image distance and the focal length of the camera includes: ； Wherein, z is the image distance of the target person deviating from the center, f is the focal length of the camera, and f′ is the focal length of the camera after calibration.

6. The mobile control method of the intelligent podium according to claim 1, characterized in that: After adjusting the actual height of the intelligent podium, the method further includes: Acquire the real-time coordinates of the target person in a preset three-dimensional coordinate system, where the preset three-dimensional coordinate system is constructed based on the center position of the camera along the smart podium as the Y axis and the width direction of the smart podium as the X axis; When the target person is within the normal activity range, the smart podium remains motionless; The normal range of activities is: ； Among them, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, and W is the preset vertical movement length.

7. The mobile control method of the intelligent podium according to claim 6, characterized in that: After obtaining the real-time coordinates of the target person in the preset three-dimensional coordinate system, the method further includes: When the target person is located in the activity range on both sides of the middle, the smart podium is driven to move laterally, and the activity range on both sides of the middle includes the left middle activity range and the right middle activity range; The range of activities on both sides of the middle part is: ； Among them, x is the horizontal coordinate of the target person, y is the vertical coordinate of the target person, L is the preset horizontal movement length, W is the preset vertical movement length, M is the preset front activity length, and D is the depth of the smart podium.

8. The mobile control method of the intelligent podium according to claim 7, characterized in that: The activity range of the target person includes the left front activity range, the right front activity range, the left middle activity range, the right middle activity range, the left rear activity range, the right rear activity range, the left invalid area range, the right invalid area range and the normal activity range; When the target person is located in the left front activity range, the right front activity range, the left rear activity range, or the right rear activity range, driving the smart podium to turn and move; When the target person is located in the left invalid area or the right invalid area, the smart podium is adjusted to a preset initial state.

9. An intelligent podium, characterized in that: include: The podium body, the motor-controlled mobile device, the binocular camera and the control processor of the smart podium; The platform body is fixedly connected to the motor-controlled moving device; the binocular camera includes a first camera and a second camera, which are arranged at different heights of the platform body along the vertical direction; The control processor is fixed on the podium body and is communicatively connected with the motor-controlled mobile device and the binocular camera; it also includes a memory and a program or instruction stored in the memory and executable on the control processor of the smart podium. When the program or instruction is executed by the control processor of the smart podium, the mobile control method of the smart podium as described in any one of claims 1 to 8 is executed.

Images