Multi-camera intelligent director method, system and device, storage medium and product

Abstract

The invention discloses a multi-camera intelligent director method, system and device, a storage medium and a product, and relates to the technical field of videos, the method is applied to an audio and video all-in-one machine, and the audio and video all-in-one machine is connected with at least one panoramic camera; the method comprises the following steps: acquiring first detection data and second detection data output by a panoramic camera; determining a current active spokesman according to the first detection data and the second detection data; target front face information of the active spokesman is determined in the first detection data and the second detection data; and determining a spokesman picture of the active spokesman in the first detection data and the second detection data according to the front face information. According to the invention, the conference all-in-one machine is connected with the panoramic camera, and the panoramic camera captures the pictures of the participants at a short distance, so that the front face of each participant can be captured through cascading of the cameras compared with a method that only one camera is adopted for shooting, and the distance of the panoramic camera is relatively short, so that the shooting efficiency is improved. And the definition of the captured front face picture is high.

Description

Technical Field

[0001] The present application relates to the field of video technology, and in particular to a multi-camera intelligent directing method, system, device, storage medium and product. Background Art

[0002] With the development of technology, audio and video technology has also been applied to the recording and live broadcast of modern conferences.

[0003] When recording or broadcasting a meeting, the meeting scene is usually captured through a camera. When someone is identified as speaking, the speaker's image is displayed in close-up to improve the immersive effect of the meeting. However, since the seats and orientations of the participants are usually not exactly the same during the meeting, it is difficult to capture the speaker's face when taking a close-up of the speaker. In addition, the distance between the camera and the face is usually far, and the close-up image quality is low. If a camera is used to take close-up shots of each participant, it will increase the hardware cost and be difficult to achieve.

[0004] Therefore, how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of a conference has become a problem that needs to be solved urgently in this field.

[0005] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The main purpose of this application is to provide a multi-camera intelligent directing method, system, device, storage medium and product, aiming to solve the technical problem of how to improve the image quality of close-up images of faces captured in the recording and live broadcast of a meeting.

[0007] To achieve the above-mentioned purpose, the present application proposes a multi-camera intelligent directing method, which is applied to an audio and video all-in-one machine, and the audio and video all-in-one machine is connected to at least one panoramic camera; the method comprises:

[0008] Acquire first detection data and second detection data output by the panoramic camera;

[0009] determining a current active speaker according to the first detection data and the second detection data;

[0010] determining target front face information of the active speaker in the first detection data and the second detection data;

[0011] The speaker screen of the active speaker is determined in the first detection data and the second detection data according to the frontal face information.

[0012] In one embodiment, the first detection data includes: a first detection picture and first sound source information, and the second detection data includes: a second detection picture and second sound source information;

[0013] The step of determining the current active speaker according to the first detection data and the second detection data comprises:

[0014] Tracking and identifying multiple faces in the first detection picture and the second detection picture to obtain face tracking results;

[0015] A current active speaker is determined among the multiple faces according to the first sound source information, the second sound source information and the face tracking result.

[0016] In one embodiment, the step of determining the current active speaker among the multiple faces according to the first sound source information, the second sound source information and the face tracking result comprises:

[0017] Determine the current target sound source location in the first sound source information and the second sound source information;

[0018] Identifying a target face at the target sound source location from the face tracking result;

[0019] When the sound accumulation at the target sound source location exceeds a preset first time, the target face is used as the current active speaker.

[0020] In one embodiment, the frontal face information includes: a frontal face picture, a target detection picture to which the frontal face picture belongs, and an identification code of the frontal face picture;

[0021] The step of determining the target front face information of the active speaker in the first detection data and the second detection data comprises:

[0022] Identify one or more front face images of the active speaker in the first detection image and the second detection image;

[0023] When there are multiple frontal face pictures, calculate the area ratio of each of the frontal face pictures in the target detection picture to which they belong, and update the frontal face picture with the largest area ratio as the current frontal face picture;

[0024] Determine a target detection screen to which the front face screen belongs, wherein the target detection screen is the first detection screen and / or the second detection screen;

[0025] An identification code is generated according to the position of the front face picture in the target detection picture.

[0026] In one embodiment, the step of determining the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information includes:

[0027] When the target detection picture is the first detection picture or the second detection picture, identifying a picture edge of the front face picture;

[0028] The first detection picture or the second detection picture is cropped according to preset picture parameters to obtain a speaker picture of the active speaker, wherein the speaker picture includes the picture edge.

[0029] In one embodiment, the step of determining the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information includes:

[0030] When the target detection screen is the first detection screen and the second detection screen, cropping the first detection screen in the first detection data according to the frontal face information to obtain a first cropped screen;

[0031] Sending the identification code in the target front face information to the panoramic camera, and receiving a second cropped image output by the panoramic camera in response to the identification code;

[0032] The first cropped image and the second cropped image are spliced ​​together to obtain a speaker image.

[0033] In addition, to achieve the above-mentioned purpose, the present application also proposes a multi-camera intelligent director system, wherein the system is arranged in an audio and video integrated machine, and the audio and video integrated machine is connected to at least one panoramic camera; the system comprises:

[0034] A picture acquisition module, used for acquiring first detection data and second detection data output by the panoramic camera;

[0035] a speaker determination module, configured to determine a current active speaker according to the first detection data and the second detection data;

[0036] a front face determination module, configured to determine target front face information of the active speaker in the first detection data and the second detection data;

[0037] A screen generation module is used to determine the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information.

[0038] In addition, to achieve the above-mentioned purpose, the present application also proposes a multi-camera intelligent directing device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the multi-camera intelligent directing method described above.

[0039] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the multi-camera intelligent directing method described above are implemented.

[0040] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the multi-camera intelligent directing method described above.

[0041] The present application provides a multi-camera intelligent directing method, in which an audio and video all-in-one machine is connected to at least one panoramic camera. During a meeting, the audio and video all-in-one machine receives second detection data collected by the panoramic camera while acquiring first detection data. The audio and video all-in-one machine then identifies the current active speaker in the detected data, determines the front face information of the active speaker in the detected data, and processes the images of the first detection data and the second detection data according to the front face information to obtain the speaker image.

[0042] In summary, the present application connects the all-in-one conference machine with a panoramic camera and uses the panoramic camera to capture the images of the participants at close range. Compared with the method of using only one camera for shooting, the present application can capture the front face of each participant through the cascade of cameras, and because the panoramic camera is at a closer distance, the captured front face image is also of higher clarity. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 A flowchart of the first embodiment of the multi-camera intelligent directing method of the present application is provided;

[0046] Figure 2This is a schematic diagram of a scenario involved in an embodiment of the multi-camera intelligent directing method of the present application;

[0047] Figure 3 This is a schematic diagram of the external process involved in the embodiment of the multi-camera intelligent directing method of this application;

[0048] Figure 4 This is a schematic diagram of the internal process involved in the embodiment of the multi-camera intelligent directing method of the present application;

[0049] Figure 5 This is a schematic diagram of information transmission involved in an embodiment of the multi-camera intelligent directing method of the present application;

[0050] Figure 6 This is a schematic diagram of the module structure of the multi-camera intelligent directing system according to an embodiment of the present application;

[0051] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the multi-camera intelligent directing method in the embodiment of the present application.

[0052] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0054] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0055] The main solution of the embodiment of the present application is: obtaining first detection data and second detection data output by the panoramic camera; determining the current active speaker based on the first detection data and the second detection data; determining the target frontal face information of the active speaker in the first detection data and the second detection data; and determining the speaker screen of the active speaker in the first detection data and the second detection data based on the frontal face information.

[0056] With the development of technology, audio and video technology has also been applied to the recording and live broadcast of modern conferences.

[0057] When recording or broadcasting a meeting, the meeting scene is usually captured through a camera. When someone is identified as speaking, the speaker's image is displayed in close-up to improve the immersive effect of the meeting. However, since the seats and orientations of the participants are usually not exactly the same during the meeting, it is difficult to capture the speaker's face when taking a close-up of the speaker. In addition, the distance between the camera and the face is usually far, and the close-up image quality is low. If a camera is used to take close-up shots of each participant, it will increase the hardware cost and be difficult to achieve.

[0058] Therefore, how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of a conference has become a problem that needs to be solved urgently in this field.

[0059] In response to the above problems, the present application provides a multi-camera intelligent directing method. In the present application, an audio and video all-in-one machine is connected to at least one panoramic camera. During the meeting, the audio and video all-in-one machine receives second detection data collected by the panoramic camera while obtaining first detection data. Then, the audio and video all-in-one machine identifies the current active speaker in the detected data, and then determines the front face information of the active speaker in the detected data, and then processes the images of the first detection data and the second detection data according to the front face information to obtain the speaker image.

[0060] In summary, the present application connects the all-in-one conference machine with a panoramic camera and uses the panoramic camera to capture the images of the participants at close range. Compared with the method of using only one camera for shooting, the present application can capture the front face of each participant through the cascade of cameras, and because the panoramic camera is at a closer distance, the captured front face image is also of higher clarity.

[0061] In this embodiment, for ease of description, the following description is made using a multi-camera intelligent director device as the execution entity.

[0062] Based on this, the embodiment of the present application provides a multi-camera intelligent directing method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the multi-camera intelligent directing method of the present application.

[0063] In this embodiment, the method is applied to an audio and video integrated machine, and the audio and video integrated machine is connected to at least one panoramic camera; the multi-camera intelligent directing method includes steps S10 to S40:

[0064] Step S10, acquiring first detection data and second detection data output by the panoramic camera;

[0065] It should be noted that, in this embodiment, please refer to Figure 2 , Figure 2This is a schematic diagram of a scenario involved in an embodiment of the multi-camera intelligent directing method of the present application, such as Figure 2 As shown, the integrated audio and video machine is usually set in front of the conference room (the black square in the figure), and the panoramic camera is set on the conference table (the black oval square in the figure). The integrated audio and video machine can also cascade multiple panoramic cameras.

[0066] In this embodiment, when a meeting is in progress, the device will obtain first detection data from the internal sensor or processing module of the connected audio and video all-in-one machine. At the same time, the device will also receive second detection data from at least one panoramic camera. These data provide a broader and more detailed scene view, which helps to capture and analyze the on-site situation more comprehensively.

[0067] Step S20, determining a current active speaker according to the first detection data and the second detection data;

[0068] In this embodiment, the two sets of data obtained are comprehensively analyzed using a built-in algorithm or processing module. By comparing information such as sound intensity and image features, the device can automatically determine which person is the active speaker, that is, the person who is speaking or making important actions.

[0069] Further, in a feasible implementation manner, the first detection data includes: a first detection picture and first sound source information, and the second detection data includes: a second detection picture and second sound source information; the above-mentioned step S20 may include steps S21 to S22:

[0070] Step S21, tracking and identifying multiple faces in the first detection picture and the second detection picture to obtain face tracking results;

[0071] In this embodiment, after acquiring the first detection screen and the second detection screen, multiple faces are tracked and identified simultaneously in the first detection screen and the second detection screen, and the facial information of all characters in the scene is captured and updated in real time to form face tracking results. These results not only include basic information such as the position and size of the character's face, but also may include facial feature information, such as the position and shape of the eyes, nose, and mouth.

[0072] Step S22: determining a current active speaker among the multiple faces according to the first sound source information, the second sound source information and the face tracking result.

[0073] In this embodiment, after obtaining the face tracking results, the first sound source information and the second sound source information, as well as the facial information of each person in the face tracking results, are combined to determine the current active speaker. By comparing the voice features and facial information of different characters, the device can automatically determine which person is speaking or making important actions, and identify him as the active speaker.

[0074] Furthermore, in a feasible implementation manner, the above step S22 may include steps S211 to S213:

[0075] Step S211, determining the current target sound source location in the first sound source information and the second sound source information;

[0076] In this embodiment, the first sound source information and the second sound source information are comprehensively analyzed to determine the current target sound source location, and the sound intensity, direction, frequency and other characteristics are analyzed and compared to determine the specific location of the sound source. By combining the data of the audio and video all-in-one machine and the panoramic camera, the device can more accurately locate the sound source, providing strong support for subsequent face recognition and active speaker determination.

[0077] Step S212, identifying a target face at the location of the target sound source from the face tracking result;

[0078] In this embodiment, after obtaining the target sound source location, the device will identify the target face located at that location from the face tracking results, and compare and analyze the facial information of each person in the face tracking results to determine the face that matches the target sound source location.

[0079] Step S213, when the sound accumulation at the target sound source location exceeds a preset first time, the target face is used as the current active speaker.

[0080] In this embodiment, after the target face is determined, it is further determined whether the sound at the location continues for more than a preset first time. If the accumulated sound time exceeds this threshold, the device will use the target face as the current active speaker.

[0081] As an example, when it is detected that someone continues to speak for more than 2-5 seconds, the person will be treated as an active speaker. If it does not exceed two seconds, he will not be treated as a valid speaker. After the active speaker stops speaking for more than 35 seconds, his image will be deleted from the conference screen.

[0082] By combining the precise positioning of sound source information with the analysis of face tracking results, this method can more accurately determine the current active speaker. It not only improves the accuracy and stability of recognition, but also effectively avoids misjudgment and missed judgment. In addition, this method further enhances the reliability and robustness of recognition by introducing the judgment condition of sound accumulation time. This intelligent directing method that integrates multiple information sources and judgment conditions provides more efficient and accurate technical support for recording and monitoring at conferences or events.

[0083] Step S30, determining the target front face information of the active speaker in the first detection data and the second detection data;

[0084] In this embodiment, once the active speaker is determined, the device further searches for the target front face information of the speaker in the first detection data and the second detection data.

[0085] Further, in a feasible implementation manner, the frontal face information includes: a frontal face picture, a target detection picture to which the frontal face picture belongs, and an identification code of the frontal face picture; the above-mentioned step S30 may include steps S31 to S34:

[0086] Step S31, identifying one or more front face images of the active speaker in the first detection image and the second detection image;

[0087] In this embodiment, one or more frontal faces of active speakers are searched and identified in the first detection screen and the second detection screen, and facial features of active speakers are captured and extracted in real time to form frontal faces. These frontal faces may exist in different detection screens, so further screening and optimization are required.

[0088] Step S32, when there are multiple frontal face pictures, calculating the area ratio of each of the frontal face pictures in the target detection picture to which they belong, and updating the frontal face picture with the largest area ratio as the current frontal face picture;

[0089] In this embodiment, when the device identifies multiple frontal face images, it calculates the area ratio of each frontal face image in the target detection image to which it belongs. This area ratio reflects the clarity of the frontal face image after it is enlarged. By comparing the area ratios of different frontal face images, the device can automatically select the frontal face image with the largest area ratio and update it as the current frontal face image. This process helps to ensure that the selected frontal face image is both clear and complete, and can accurately reflect the facial features of the active speaker.

[0090] Step S33, determining the target detection screen to which the front face screen belongs, wherein the target detection screen is the first detection screen and / or the second detection screen;

[0091] In this embodiment, after determining the current front face picture, it is necessary to further determine the target detection picture to which the front face picture belongs. The target detection picture may be the first detection picture, the second detection picture, or both detection pictures. By determining the target detection picture, the device can understand the source and location of the front face picture, providing important information for subsequent processing and display.

[0092] Step S34, generating an identification code according to the position of the front face image in the target detection image.

[0093] In this embodiment, an identification code is generated according to the position of the front face picture in the target detection picture. This identification code is usually a unique identifier used to distinguish different front face pictures and detection pictures. By generating the identification code, the device can more conveniently manage and track the relevant information of the front face picture, which facilitates subsequent processing and analysis.

[0094] Step S40: determining the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information.

[0095] In this embodiment, based on the determined frontal face information, the device will filter out the best picture containing the frontal face from the first detection data and the second detection data.

[0096] Furthermore, in a feasible implementation manner, the above step S40 may include steps S41 to S42:

[0097] Step S41, when the target detection picture is the first detection picture or the second detection picture, identifying the picture edge of the front face picture;

[0098] In this embodiment, when the target detection screen is determined to be the first detection screen or the second detection screen, the device will further identify the edge of the front face screen in the screen, and automatically identify and extract the outline and boundary of the front face screen. By identifying the edge of the screen, the device can more accurately determine the position and size of the front face screen in the overall screen, providing an important basis for subsequent screen cropping.

[0099] Step S42 , cropping the first detection picture or the second detection picture according to preset picture parameters to obtain a speaker picture of the active speaker, wherein the speaker picture includes the picture edge.

[0100] In this embodiment, after identifying the edge of the front face picture, the device will crop the first detection picture or the second detection picture according to the preset picture parameters. These picture parameters may include the width, height, ratio, etc. of the picture, which are used to determine the size and shape of the speaker picture after cropping. During the cropping process, the device will ensure that the cropped picture contains the edge of the front face picture to ensure the integrity and accuracy of the picture. Through this process, the device can generate a clear, complete speaker picture that focuses on the active speaker.

[0101] Further, in another feasible implementation manner, the above step S40 may include steps S43 to S45:

[0102] Step S43, when the target detection picture is the first detection picture and the second detection picture, cropping the first detection picture in the first detection data according to the frontal face information to obtain a first cropped picture;

[0103] In this embodiment, when the target detection screen is included in both the first detection screen and the second detection screen (or, depending on the context, it can be understood that when the front face of the active speaker appears in the field of view of the two detection cameras at the same time), the device will crop the first detection screen in the first detection data according to the front face information. The cropping process will be performed based on the position and size of the front face screen in the first detection screen and the preset screen parameters to ensure that the cropped first cropped screen can clearly show the facial features of the active speaker.

[0104] Step S44, sending the identification code in the target front face information to the panoramic camera, and receiving the second cropped image output by the panoramic camera in response to the identification code;

[0105] In this embodiment, the identification code in the target face information is sent to the panoramic camera. After receiving the identification code, the panoramic camera will quickly locate and cut out the second cropped image corresponding to the identification code according to its internal image processing algorithm and database. This second cropped image may contain broader background information or show the full picture of the active speaker from another angle. The response speed and cropping accuracy of the panoramic camera are crucial to achieve real-time, high-quality speaker image stitching.

[0106] Step S45: splicing the first cropped image and the second cropped image to obtain a speaker image.

[0107] In this embodiment, finally, the device will splice the first cropped screen with the second cropped screen. The splicing process will take into account the perspective, light, color and other factors of the two screens to ensure that the spliced ​​speaker screen remains visually coherent and consistent. Through this step, the device can generate a speaker screen that contains both clear facial features and a broad background, providing the audience with a richer and more three-dimensional visual experience. Exemplarily, when multiple screens are synthesized, the host position can trigger the screen synthesis process with a new frame of the local screen. During synthesis, if the new frame of the sub-screen of the slave position has not arrived, the old frame is not avoided, and it is necessary to wait for up to 1 frame. The reception of the new frame from the slave position needs to be cached using a queue. When the new frame arrives, if the old frame has not been used, the new frame is cached first, and up to two frames are cached to avoid unnecessary delays. After receiving the encoding packet, it is necessary to determine whether a new encoding channel needs to be created or destroyed based on the channel number, etc. Since it takes a long time to create a codec channel, if you create and destroy a codec channel in the same thread that receives the encoded packet, packets may be lost. Therefore, the process of creating and destroying the codec and sending the encoded packet to the encoder needs to be sent to a thread pool, which is equivalent to caching the received encoded packets. If there is no new frame, the oldest frame will be used. If there is no frame for a period of time, it is considered that there is no data in this channel and it will be closed directly.

[0108] For example, in one example, please refer to Figure 3 and Figure 4 , Figure 3 This is a schematic diagram of the external process involved in the embodiment of the multi-camera intelligent directing method of this application. Figure 4 This is a schematic diagram of the internal flow involved in the embodiment of the multi-camera intelligent directing method of this application. During a meeting, a 360 camera is directly powered through the PSE network port of the Videobar. After the connection is successful, the Videobar turns on the intelligent directing function, and the intelligent directing function of the 360 ​​camera is turned on simultaneously. When cascading 2 or more 360 ​​cameras, the second 360 camera needs to be powered by an external power adapter; the 360 ​​cameras are connected through the upper and lower cascade ports, and the Videobar is selected as the host to decide the layout of the directing screen;

[0109] The 360 ​​camera sends all its detection data to VideoBar, including DOA information;

[0110] VideoBar selects the front face information of the active speaker from all the camera positions, and after comparison, deduplication, and DOA information calculation, selects the front face information of the speaker with the highest clarity; obtains a data with uuid layout information and publishes it to each camera position;

[0111] After receiving the layout information, the camera compares it with the local uuid to find the encoded data that needs to be cut out, and sends the code stream to the host (usually VideoBar);

[0112] The host receives the code stream, decodes it and combines it with its own cutout image, and outputs the image via USB.

[0113] Externally, the final output image of the audio and video integrated machine is obtained by cropping the front face images obtained by each panoramic camera.

[0114] Internally, the panoramic camera also includes AI detection content. After detecting the speaker, the detection information is sent to the AI ​​application. The AI ​​application also receives the sound source positioning information output from the audio positioning module through the chip, and then obtains the speaker's front face information based on the DOA (sound source positioning) information and the front face image information, and then sends it to the AI ​​application of the main audio and video all-in-one machine. The AI ​​application of the main audio and video all-in-one machine simultaneously receives the monitoring information and sound source positioning information detected by itself, and also obtains the speaker's front face information based on the DOA (sound source positioning) information and the front face image information, and then performs UUID / IP / CameraID layout and cutout information on the front face information. The main device performs cutout based on the information and publishes the information at the same time. After receiving the information, the panoramic camera performs cutout and encoding of the face, and outputs the bit stream to the audio and video all-in-one machine. The audio and video all-in-one machine synthesizes the cutout and finally outputs the synthesized image through USB.

[0115] During data transfer, please refer to Figure 5 , Figure 5 This is a schematic diagram of information transmission involved in the embodiment of the multi-camera intelligent directing method of the present application, such as Figure 5 As shown in the figure, through the distributed data distribution method, a node as a publisher publishes a message to a topic, and other nodes as subscribers subscribe to this topic to receive messages. This publish / subscribe method allows decoupled communication, so that different nodes can flexibly exchange data. Each camera position performs AI detection of speakers locally. After detection, the non-host device transmits the detection data to the AI ​​application, and then the AI ​​application adds the new monitoring data with uui d to the data and publishes it as topic 1. The host's AI application receives the detection data and subscribes to topic 1. It processes the detection data to obtain layout information and publishes the layout information as topic 2. The media processing modules of the three devices subscribe to topic 2. The slave device adds its own information to the layout information and sends the cropped code stream to the host's media processing module for further processing.

[0116] Specifically, when directing, the presentation strategy for close-up images of active speakers is as follows:

[0117] (1) When the speaker is discussing with the on-site personnel, the front face is captured by the 360 ​​camera, and the 360 ​​camera provides a close-up picture; when the speaker is discussing with the remote participants, the face is facing the front large screen, and the front face is captured by the VideoBar (audio and video integrated machine), and the VideoBar provides a close-up picture;

[0118] (2) If the face is not facing completely straight on either side, determine which side it is facing and select the one that is more consistent with the straight face direction. If both sides are profile faces and it is impossible to determine, select the close-up on the 360 ​​camera, because the 360 ​​camera is usually closer to the person, the face accounts for a larger proportion, and the cutout is clearer.

[0119] (3) Whether standing or sitting, participants of different heights should keep the head and eye level of the presenter consistent;

[0120] (4) The active speaker moves slightly, and the picture is tracked and adjusted in each independent small picture;

[0121] When there are only two cameras, the slave camera directly uses H.264 according to the encoding algorithm and the frame rate requirement of 30 frames, and then encodes according to the resolution of the sub-pictures required for synthesis. The number of encoders needs to be created according to the number of sub-pictures of the local camera. The decoding end creates different decoders when it receives different new frames, and the synthesis module completes the decoding and synthesis processing.

[0122] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the multi-camera intelligent directing method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0123] This application also provides a multi-camera intelligent broadcasting system, please refer to Figure 6 , the system is arranged in an audio and video integrated machine, and the audio and video integrated machine is connected to at least one panoramic camera; the multi-camera intelligent directing system includes:

[0124] The image acquisition module 10 is used to acquire the first detection data and the second detection data output by the panoramic camera;

[0125] A speaker determination module 20, configured to determine a current active speaker according to the first detection data and the second detection data;

[0126] A front face determination module 30, configured to determine target front face information of the active speaker in the first detection data and the second detection data;

[0127] The screen generation module 40 is used to determine the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information.

[0128] Optionally, the first detection data includes: a first detection picture and first sound source information, and the second detection data includes: a second detection picture and second sound source information; the speaker determination module 20 is further used to:

[0129] Tracking and identifying multiple faces in the first detection picture and the second detection picture to obtain face tracking results;

[0130] A current active speaker is determined among the multiple faces according to the first sound source information, the second sound source information and the face tracking result.

[0131] Optionally, the speaker determination module 20 is further configured to:

[0132] Determine the current target sound source location in the first sound source information and the second sound source information;

[0133] Identifying a target face at the target sound source location from the face tracking result;

[0134] When the sound accumulation at the target sound source location exceeds a preset first time, the target face is used as the current active speaker.

[0135] Optionally, the frontal face information includes: a frontal face picture, a target detection picture to which the frontal face picture belongs, and an identification code of the frontal face picture; the frontal face determination module 30 is further used for:

[0136] Identify one or more front face images of the active speaker in the first detection image and the second detection image;

[0137] When there are multiple frontal face pictures, calculate the area ratio of each of the frontal face pictures in the target detection picture to which they belong, and update the frontal face picture with the largest area ratio as the current frontal face picture;

[0138] Determine a target detection screen to which the front face screen belongs, wherein the target detection screen is the first detection screen and / or the second detection screen;

[0139] An identification code is generated according to the position of the front face picture in the target detection picture.

[0140] Optionally, the screen generating module 40 is further used for:

[0141] When the target detection picture is the first detection picture or the second detection picture, identifying a picture edge of the front face picture;

[0142] The first detection picture or the second detection picture is cropped according to preset picture parameters to obtain a speaker picture of the active speaker, wherein the speaker picture includes the picture edge.

[0143] Optionally, the screen generating module 40 is further used for:

[0144] When the target detection screen is the first detection screen and the second detection screen, cropping the first detection screen in the first detection data according to the frontal face information to obtain a first cropped screen;

[0145] Sending the identification code in the target front face information to the panoramic camera, and receiving a second cropped image output by the panoramic camera in response to the identification code;

[0146] The first cropped image and the second cropped image are spliced ​​together to obtain a speaker image.

[0147] The multi-camera intelligent broadcasting system provided by the present application adopts the multi-camera intelligent broadcasting method in the above-mentioned embodiment, and can solve the technical problem of how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of the conference. Compared with the prior art, the beneficial effects of the multi-camera intelligent broadcasting system provided by the present application are the same as the beneficial effects of the multi-camera intelligent broadcasting method provided by the above-mentioned embodiment, and other technical features of the multi-camera intelligent broadcasting system are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0148] The present application provides a multi-camera intelligent directing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the multi-camera intelligent directing method in the above-mentioned embodiment one.

[0149] Reference below Figure 7 , which shows a structural schematic diagram of a multi-camera intelligent director device suitable for implementing an embodiment of the present application. Figure 7 The multi-camera intelligent director device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0150] like Figure 7As shown, the multi-camera intelligent broadcasting device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the multi-camera intelligent broadcasting device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the multi-camera intelligent broadcasting device to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a multi-camera intelligent broadcasting device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or provided alternatively.

[0151] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0152] The multi-camera intelligent broadcasting device provided by the present application adopts the multi-camera intelligent broadcasting method in the above embodiment, which can solve the technical problem of how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of the meeting. Compared with the prior art, the beneficial effects of the multi-camera intelligent broadcasting device provided by the present application are the same as the beneficial effects of the multi-camera intelligent broadcasting method provided by the above embodiment, and other technical features in the multi-camera intelligent broadcasting device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0153] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0154] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0155] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the multi-camera intelligent directing method in the above-mentioned embodiment.

[0156] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0157] The above-mentioned computer-readable storage medium may be included in the multi-camera intelligent broadcast director device; or it may exist independently without being assembled into the multi-camera intelligent broadcast director device.

[0158] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the multi-camera intelligent broadcast director device, the multi-camera intelligent broadcast director device: obtains first detection data and second detection data output by the panoramic camera; determines the current active speaker according to the first detection data and the second detection data; determines the target front face information of the active speaker in the first detection data and the second detection data; and determines the speaker screen of the active speaker in the first detection data and the second detection data according to the front face information.

[0159] The computer program code for performing the operation of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0160] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0161] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0162] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned multi-camera intelligent directing method, and can solve the technical problem of how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of a conference. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the multi-camera intelligent directing method provided by the above-mentioned embodiment, and will not be repeated here.

[0163] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the multi-camera intelligent directing method as described above.

[0164] The computer program product provided by the present application can solve the technical problem of how to improve the image quality of close-up images of human faces captured in the recording and live broadcast of a conference. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the multi-camera intelligent directing method provided by the above embodiment, and will not be repeated here.

[0165] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A multi-camera intelligent directing method, characterized in that: The method is applied to an integrated audio and video machine, and the integrated audio and video machine is connected to at least one panoramic camera; the method comprises: Acquire first detection data and second detection data output by the panoramic camera; determining a current active speaker according to the first detection data and the second detection data; determining target front face information of the active speaker in the first detection data and the second detection data; The speaker screen of the active speaker is determined in the first detection data and the second detection data according to the frontal face information.

2. The multi-camera intelligent directing method according to claim 1, characterized in that: The first detection data includes: a first detection picture and first sound source information, and the second detection data includes: a second detection picture and second sound source information; The step of determining the current active speaker according to the first detection data and the second detection data comprises: Tracking and identifying multiple faces in the first detection picture and the second detection picture to obtain face tracking results; A current active speaker is determined among the multiple faces according to the first sound source information, the second sound source information and the face tracking result.

3. The multi-camera intelligent directing method according to claim 2, characterized in that: The step of determining the current active speaker among the multiple faces according to the first sound source information, the second sound source information and the face tracking result comprises: Determine the current target sound source location in the first sound source information and the second sound source information; Identifying a target face at the target sound source location from the face tracking result; When the sound accumulation at the target sound source location exceeds a preset first time, the target face is used as the current active speaker.

4. The multi-camera intelligent directing method according to claim 3, characterized in that: The front face information includes: a front face picture, a target detection picture to which the front face picture belongs, and an identification code of the front face picture; The step of determining the target front face information of the active speaker in the first detection data and the second detection data comprises: Identify one or more front face images of the active speaker in the first detection image and the second detection image; When there are multiple frontal face pictures, calculate the area ratio of each of the frontal face pictures in the target detection picture to which they belong, and update the frontal face picture with the largest area ratio as the current frontal face picture; Determine a target detection screen to which the front face screen belongs, wherein the target detection screen is the first detection screen and / or the second detection screen; An identification code is generated according to the position of the front face picture in the target detection picture.

5. The multi-camera intelligent directing method according to claim 4, characterized in that: The step of determining the speaker screen of the active speaker in the first detection data and the second detection data according to the front face information comprises: When the target detection picture is the first detection picture or the second detection picture, identifying a picture edge of the front face picture; The first detection picture or the second detection picture is cropped according to preset picture parameters to obtain a speaker picture of the active speaker, wherein the speaker picture includes the picture edge.

6. The multi-camera intelligent directing method according to claim 4, characterized in that: The step of determining the speaker screen of the active speaker in the first detection data and the second detection data according to the front face information comprises: When the target detection screen is the first detection screen and the second detection screen, cropping the first detection screen in the first detection data according to the frontal face information to obtain a first cropped screen; Sending the identification code in the target front face information to the panoramic camera, and receiving a second cropped image outputted by the panoramic camera in response to the identification code; The first cropped image and the second cropped image are spliced ​​together to obtain a speaker image.

7. A multi-camera intelligent broadcasting system, characterized in that: The system is arranged on an audio and video integrated machine, and the audio and video integrated machine is connected to at least one panoramic camera; the system comprises: A picture acquisition module, used to acquire first detection data and second detection data output by the panoramic camera; a speaker determination module, configured to determine a current active speaker according to the first detection data and the second detection data; a front face determination module, configured to determine target front face information of the active speaker in the first detection data and the second detection data; A screen generation module is used to determine the speaker screen of the active speaker in the first detection data and the second detection data according to the frontal face information.

8. A multi-camera intelligent broadcasting device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the multi-camera intelligent directing method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the multi-camera intelligent directing method as described in any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the multi-camera intelligent directing method as described in any one of claims 1 to 6 are implemented.

Images