CN111701254B

CN111701254B - Parallel acceleration display method for large-scale performance dynamic stage video

Info

Publication number: CN111701254B
Application number: CN202010828784.2A
Authority: CN
Inventors: 李立杰; 丁刚毅; 李鹏; 黄天羽; 唐明湘
Original assignee: Beijing Institute of Technology BIT
Current assignee: Beijing Institute of Technology BIT
Priority date: 2020-08-18
Filing date: 2020-08-18
Publication date: 2020-11-24
Anticipated expiration: 2040-08-18
Also published as: CN111701254A

Abstract

The invention relates to a parallel acceleration display method for a dynamic stage video of a large-scale performance, which comprises the following steps: firstly, designing a source video; secondly, parallelly starting a producer thread for each display time point according to the time sequence, and executing the following operations: s1, allocating a source memory and a target memory; s2, obtaining a dynamic stage model corresponding to the time point; s3, unfolding the display screen on the corresponding source image plane, and obtaining a segmentation area according to the corresponding relation between the unfolded geometric shape and the source image; s4, outputting all the display screen division areas to a target memory and outputting to a stage screen control video file; and thirdly, outputting the stage screen control video files containing all the time points to a display controller in sequence. The invention reduces the hardware complexity of dynamic stage management and improves the memory utilization efficiency and the data processing speed while realizing the accurate stage modeling video background.

Description

Parallel acceleration display method for large-scale performance dynamic stage video

Technical Field

The invention relates to a digital video display method of a dynamic stage display screen, in particular to a parallel acceleration display method of a dynamic stage video for large-scale performance.

Background

The stage provides space for the performance. Whether the expected effect can be achieved by one artistic performance or not is based on the stage. The modern stage, especially the multimedia dynamic stage, creates more development space for stage art in the limited stage, and provides more changes and choices for directors and stage beauty personnel.

The development of the lifting platform is a representative of the mechanical development of the stage, and the lifting platform is commonly applied to the modern stage technology. From the first few lifting platforms to the application of the present large-scale lifting platforms, the lifting platforms become an important component of the stage from an auxiliary mechanical device of the stage. As the number of lift tables increases, the structures and usage of the lift tables also change. Especially in the three-dimensional dynamic multimedia stage, the top surface and the periphery of the lifting platform are provided with LED boards capable of playing videos. When the lifting platform is lifted and forms a static stage platform type, each video playing surface displays pictures or videos matched with programs, and therefore the lifting platform also becomes a part of the stage background.

In a large performance, the number of LED display screens mounted on the stage module is large, and often thousands of display screens of various sizes are included. If for every display screen individual design show content, will consume huge manpower and materials undoubtedly, increased stage designer's the design degree of difficulty moreover, also be difficult to guarantee the final synthetic effect of all display screens.

How to manage the editing and output of video files in a large performance so that all screens can be matched with each other to accurately display each video; how to reduce the hardware complexity for managing the LED screen; particularly, how to effectively manage the digital video mapping memory of the display screen and reduce the time required by digital mapping through parallel processing are all technical problems which need to be solved urgently. But no relevant description is found in the prior art.

Disclosure of Invention

The invention aims to provide a parallel accelerated display method for a large-scale performance dynamic stage video aiming at the defects of the prior art, provide accurate LED screen digital mapping aiming at a dynamic stage video background and improve the display speed through parallel processing.

In order to achieve the above purpose, the invention provides a parallel acceleration display method for a large-scale performance dynamic stage video, which comprises the following steps:

designing one or more source videos serving as a stage background according to the overall stage display effect, and decomposing each source video file into a video frame sequence;

setting a display time interval of a display screen, starting a producer thread for each display time point in parallel according to a time sequence under the condition that system resources allow, and executing the following operations:

s1, distributing source memory for the frame image and distributing target memory for the display screen; reading all the frame images of the time point into a source memory;

s2, obtaining a dynamic stage model corresponding to the time point, and obtaining the spatial position, orientation and size of each display screen in each stage module;

s3, determining a display source image corresponding to the stage module display screen, unfolding the display screen on the plane of the source image, and obtaining a corresponding segmentation area of each display screen in the source image according to the corresponding relation between the unfolded geometric shape of the display screen and the source image set by a user;

s4, circularly executing the following operations on all display screens:

s41, segmenting a corresponding segmentation area of the display screen from the corresponding source image;

s42, outputting the content of the corresponding divided area to a target memory of a display screen;

s43, outputting the target memory content to a stage screen control video file;

and thirdly, outputting the stage screen control video files containing all the time points to a display controller in sequence.

According to a specific implementation manner of the embodiment of the invention, each producer thread is not responsible for writing in the stage screen control video file, and after the step S42 is executed, a buffer conversion completion flag is set to wait for the consumer thread to output the buffer; and additionally, starting a global consumer thread, requesting a target memory according to a time sequence, adding the target memory of the producer thread into the stage screen control video file if the conversion of the corresponding buffer area is completed, and releasing the buffer area.

According to a specific implementation manner of the embodiment of the present invention, the method further includes a step of editing the image of the divided area.

According to a specific implementation manner of the embodiment of the present invention, all the display screens are grouped according to the positions, the display controllers of each group of display screens are merged, and a target memory space is allocated to each group of display screens, in step S4, the image segmentation and memory copy operations are cyclically executed by taking the display screen groups as a unit, and the segmentation areas corresponding to each display screen in the group are copied to the target memory space according to the display screen arrangement sequence; the consumer thread requests the target memory according to time and display screen output order.

According to a specific implementation mode of the embodiment of the invention, for the multi-layer arrangement stage with a shielding relation, the shielded display screens are grouped without carrying out target memory copy operation; and in the process of outputting the target memory to the display controller, directly multiplexing the display screen which is not shielded at the most front position to group the target memory content.

According to a specific implementation manner of the embodiment of the invention, the output sequence of the display screen groups is arranged, so that the multi-layer arrangement stage with the shielding relation is output from front to back, and the common target memory space is distributed for the corresponding display screen groups; in step S4, it is first determined whether the display screen packet is occluded, and the occluded display screen packet is touched to directly multiplex the existing content sharing the target memory space.

According to a specific implementation manner of the embodiment of the invention, a different video file is generated for each display screen group; dividing each producer thread into a plurality of batches by taking display screen grouping as a unit, and starting to execute the p-th batch of the thread i when the p-1-th batch of the thread i is completely consumed and the p-th batch of the thread i-1 is also completely consumed; and if the consumer thread detects a buffer conversion completion mark in any producer thread, adding the thread target memory into the corresponding stage screen control video file, and releasing the buffer.

In another aspect, the present invention further provides an electronic device, including:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform one of the aforementioned methods of video-parallel accelerated large-performance motion stage display.

In another aspect, the present invention further provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned method for video parallel accelerated display of a large performance dynamic stage.

In another aspect, the present invention further provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to execute the aforementioned method for video parallel accelerated display of a large performance dynamic stage.

Advantageous effects

The parallel acceleration display method of the large-scale performance dynamic stage video, provided by the invention, provides a parallel acceleration LED screen video background digital mapping scheme aiming at the dynamic stage modeling, reduces the hardware complexity of dynamic stage management and improves the memory utilization efficiency and the data processing speed while realizing the accurate stage modeling video background.

Drawings

FIG. 1 is a flow chart of a parallel acceleration display method for a large-scale performance dynamic stage video, which is implemented by the present invention;

FIG. 2 is a schematic view of a dynamic stage;

FIG. 3 is a schematic view of a second dynamic stage;

fig. 4 is a schematic view of a third dynamic stage;

fig. 5 is a data flow diagram of a video parallel acceleration display method for a dynamic stage of a large performance implemented by the present invention;

FIG. 6 is a schematic diagram of parallel execution of batch-free tasks;

FIG. 7 is a diagram illustrating the direct parallel execution of batch tasks;

FIG. 8 is a schematic diagram of an improved parallel execution of batch tasks.

Detailed Description

Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

In the stereo dynamic multimedia stage used in large-scale performance, LED boards capable of playing videos are installed on the top surface and the periphery of the lifting platform. When the lifting platform is lifted and forms a static stage platform type, each video playing surface displays pictures or videos matched with programs, and therefore the lifting platform also becomes a part of the stage background. The invention provides a video parallel acceleration display method for a dynamic stage for large-scale performance, which aims at the dynamic stage used in the large-scale performance and controls the display content of an LED display screen in a coordinated manner and accelerates by using parallel processing.

Fig. 1 is a flowchart of parallel acceleration of large performance dynamic stage video display implemented according to an embodiment of the present invention. Fig. 2 and 3 show two dynamic stages, respectively. As shown in fig. 2, the large performance dynamic stage is composed of a plurality of stage modules, which may be cubes, cuboids, or special-shaped stage modules. Each module of the stage is connected with a slide rail, and the stage module is controlled to move in the front-back direction or the left-right direction or the up-down direction through the mechanical slide rail. By controlling the movement of the stage module, different stage models can be formed. In the stage shown in fig. 3, LED display screens capable of playing video are installed on the top and side surfaces of each stage module. Similar to the display card for controlling the output of the display, in order to control the display of the LED display screens, each LED display screen is connected with the display controller through a signal line, and the broadcasting control personnel control the display content of each LED display screen by outputting the display signal to the display controller.

Because the dynamic stage comprises a plurality of stage modules, the contents output by each stage module are matched to form the whole background of the stage. In a large performance, the number of LED display screens mounted on the stage module is large, and often thousands of display screens of various sizes are included. If for every display screen individual design show content, will consume huge manpower and materials undoubtedly, increased stage designer's the design degree of difficulty moreover, also be difficult to guarantee the final synthetic effect of all display screens. And the display information of thousands of high-precision display screens needs huge storage space, and mass video data needs to be processed to reach the high precision, low delay and high frame rate required by video display, so that the video display method is also a technical problem faced by large-scale performance dynamic stage video display.

In order to solve the above technical problem, the embodiment of the present invention realizes a parallel acceleration display method for a dynamic stage video of a large performance, as shown in fig. 1, including the following steps:

according to the space-time consistency-based large-scale performance dynamic stage video editing and displaying method provided by the embodiment, stage designers do not need to design display contents for each display screen independently, but consider the stage shapes formed by all stage modules as a whole, and design backgrounds according to the whole display effect of the stages. For example, if all stage modules constitute a flat large screen, the designer only needs to design one video file to be displayed on the large screen. Different positions of the complex stage may display different videos, for example, a top stereo stage displays a sky video, a ground stereo stage displays a forest video, and the like, and a designer needs to make a plurality of source videos for the stage background. The stage shown in fig. 3 is composed entirely of cubic stage modules, i.e., the display screens installed on the top and side surfaces of the stage modules may have five orientations, i.e., top, left, right, front, and back, and the display contents of the display screens oriented in the same direction are combined into a stage background of the orientation. The audience located at different positions on the stage sees different display screens, i.e., different stage backgrounds. Therefore, the designer needs to design a different source video for each orientation, i.e., five source videos for the stage shown in fig. 3. For a complex stage configuration, such as arranging a plurality of partial stereoscopic dynamic stage configurations shown in fig. 3 in a whole stage, it is necessary to design five source videos for each partial stereoscopic stage.

For stage designers, the design and installation of the display screen of the stage module are not required to be concerned, and only the stage model at a specific moment is required to be known and the source video serving as the overall background of the stage is designed.

After the designer designs the source video file, the technician needs to display the video file as a background on a complex dynamic stage. The video map is embodied as an image map at each specific instant of time, and thus in order to perform the video mapping, each source video file needs to be decomposed into a sequence of video frames.

Setting a display time interval of a display screen, and parallelly starting a producer thread for each display time point according to the time sequence under the condition that system resources allow:

the time interval of the video frames is not necessarily consistent with the display time interval of the stage display screen, so that the video frames may be sampled or differenced to obtain frame images corresponding to the display time point of each display screen. Because the video file data volume of the large-scale performance dynamic stage is huge, reading, converting and storing the video data consume a large amount of time and space resources, and the data processing process needs to be accelerated by parallel processing. The present embodiment starts multiple producer threads, each of which handles frame image conversion for one time point. The number of threads is determined by system resources, for example, each thread may correspond to a CPU core, and 10 producer threads are simultaneously turned on, that is, frame image conversion at 10 time points may be simultaneously processed. When the 1 st time point is processed, the occupied resources can be released, and the 11 th time point conversion is started to be processed. Each producer thread performs the following operations:

after obtaining the frame image as the overall background of the stage, stage technicians need to accurately map the source image to each LED display screen of the stage module. In order to perform accurate mapping, a specific stage model needs to be obtained first. Since each stage module in the dynamic stage moves with time, the position of each stage module at a specific moment of mapping needs to be acquired first, and spatial position, orientation and size data of each display screen installed on the stage module can be acquired from the position of the stage module and the shape and size data of the stage module.

in order to establish a corresponding relationship between the source image and the stage module display screen, after the spatial position and the size of each display screen are obtained, the display screen needs to be unfolded on the corresponding source image plane. It should be noted that even a plurality of display screens installed on the same stage module surface have different orientations, so that the corresponding source images are different. In the dynamic stage as shown in fig. 3, 5 display screens on each stage module surface correspond to 5 source images. Therefore, before unfolding, the source image to be displayed corresponding to each display screen needs to be determined. This correspondence may be automatically determined by a program, for example in a dynamic stage as shown in fig. 3, from the orientation of each display screen the corresponding source image to be displayed may be determined. In some dynamic stages, a user needs to designate a corresponding source image to be displayed. For example, fig. 4 shows a cylindrical arrangement of stage modules. A plurality of columns are included in the overall stage, each column being made up of an arrangement of stage modules as shown in fig. 4. The user may specify that each cylinder displays a particular image.

After the source image to be displayed corresponding to each display screen is established, the display screen needs to be unfolded on the source image plane to be displayed. The specific deployment strategy is set by the user according to stage characteristics and stage design. For example, for the dynamic stage shown in fig. 3, the expansion can be performed by projection, i.e. each display screen is projected on the source image plane to be displayed. For the cylindrical stage shown in fig. 4, all display screens parallel to the cylindrical surface can be tiled and unfolded, the adjacent display screens are connected seamlessly, a rectangle is obtained after the display screens are unfolded, and then source images which are designed by stage designers and need to be displayed on the cylindrical surface are mapped on the rectangle.

After the stage module display screen is unfolded on the plane of the source image to be displayed, the corresponding relation between the unfolded geometric shape of the display screen and the source image is also required to be set. For example, in a dynamic stage as shown in fig. 3, the resulting unfolded geometry may not be a rectangle after projecting all the display screens onto the corresponding source image planes. The user can set the maximum rectangle which can be formed by the projection of the display screen to correspond to the source image according to the movement range of the display screen. Or the user may obtain a minimum rectangle containing all the display screen projections at each particular moment during the display screen movement and set this rectangle to correspond to the source image. Since the unfolded geometric shape is not a complete rectangle, the designer needs to consider that the motion of the stage module may cause the partial content of the source image to be lost in the design process. In the dynamic stage as shown in fig. 4, the rectangle obtained by spreading all the display screens parallel to the cylindrical surface is corresponding to the source image to be displayed on the cylindrical surface, so that the effect of wrapping the designed source image on the cylindrical surface can be obtained.

And obtaining the corresponding segmentation area of each display screen in the source image according to the corresponding relation between the expanded geometric shape of the display screen and the source image.

S4, circularly executing the following operations on all display screens:

s43, outputting the target memory content to a stage screen control video file;

After the corresponding division area of each display screen in the source image is obtained in step S3, in step S4, the content to be displayed of each display screen needs to be divided from the source image and transmitted to the display screen for output. Similar to the display card controlling the output of the display, each LED display screen of the dynamic stage also needs to be connected to the display controller through a signal line. And outputting the contents to be displayed of each display screen to the display controller by the stage technicians, and obtaining the contents to be output from the display controller through the signal lines and outputting the contents by the display screens.

Fig. 5 shows a data flow diagram of a source image segmentation conversion according to a specific implementation manner of the embodiment of the invention. As shown in fig. 5, when performing dynamic stage digital display mapping, frame images are first read into a source memory, and the content of the source memory remains unchanged during processing and display of all display screens. After the corresponding segmentation area of each display screen in the source image is obtained, the offset of each pixel point in the area relative to the original point of the image can be calculated by the segmentation area, and therefore the source memory address corresponding to the pixel point can be obtained. And combining the source memory address units corresponding to all the pixel points in the partition region to form a source memory address space corresponding to the partition region. The source memory address space corresponding to the partition region may be a block of continuous address space in the source memory, as shown in partition region 2 in fig. 5; it may also be some discrete address space in the source memory, as shown by partition 1 in fig. 5. After the address space of the source memory of the partition area corresponding to each display screen is obtained, the content of the address space needs to be copied to the target memory according to the sequence of the pixel points of the display screen, so that the content of the target memory is consistent with the content of the stage display screen in space. And finally, outputting the content of the target memory to a display controller.

Under an ideal condition, at each display time point, the source image is converted and output to the display screen in real time, so that the dynamic video background can be seen on the display screen of the dynamic stage. However, in practical situations, since a large-scale motion stage involves hundreds of high-resolution display screens, the volume of video files is huge, and it takes a long time to convert, copy and output each frame of source image. Therefore, if the conversion and output of the video frame image are performed at each specific time of display, the frame rate required for video output cannot be achieved, that is, the effect of real-time video background display cannot be achieved.

According to a specific implementation manner of the embodiment of the present invention, at each display time point, the target memory content is added to the stage screen control video file in step S43; and outputting the stage screen control video file containing all the time points to a display controller when the display is needed. Namely, completing the conversion and storage of the source video before the performance; when the performance is carried out, the processed stage screen control video file is directly output to the display controller, so that the video background display effect of the dynamic stage of the large-scale performance can be realized.

When converting a source video file into a stage screen control video file, it is most important to keep the temporal and spatial relationship between the stage screen control video file and the display screen display consistent: the time consistency means that each stage screen control video file is formed by arranging display contents at different display time points in sequence, and the space consistency means that each stage screen control video file corresponds to a specific display screen in a stage space and only contains the display contents in the corresponding display screen. The stage screen control video file is completely different from the source video, and the stage screen control video file is messy and difficult to understand if the stage screen control video file is directly played on a common computer, but if the stage screen control video file is output to a display controller of a dynamic stage, the output of the display screen can be correctly controlled, and the combination of a plurality of display screens can display correct video backgrounds.

The stage screen control video file may be one or a plurality of files. For a simple stage, the number of screens is small, the video data volume is small, and one stage screen can meet the requirement for controlling a video file; however, for a large complex stage, the amount of video data is very large, and if only one video file is used for controlling, it is difficult to achieve the processing speed and frame rate required for smooth video playing, so that a plurality of final stage screen control video files are usually required to be formed, and each file controls a partial screen.

Because a plurality of parallel producer threads exist in the system, if the plurality of threads are responsible for the stage screen to control the writing of the video file, the correct time sequence of the file writing is difficult to ensure, so that the frame time sequence error occurs in the final display screen video.

Different from the parallel opening of multiple producer buffers, in order to ensure that the output video file writes data in a correct sequence, in this embodiment, only one consumer thread is opened, and a target memory is requested according to a time sequence, that is, the target memory at the 1 st time point is requested first, and if the conversion of the producer thread corresponding to the 1 st time point is completed, the target memory corresponding to the producer thread is added to the stage screen control video file, and the buffer is released. The consumer thread then requests the target memory at the 2 nd time point. If the producer thread corresponding to the 2 nd time point has not completed the conversion, the consumer thread will wait. Therefore, the data of the point 2 can be written after the data of the point 1 is written by the consumer thread, and the output video file is ensured to be written with the data according to the correct sequence.

And step two, one or more stage screen control video files are obtained, so that the stage screen control video files can be directly and sequentially output to the display controller during performance, and the stage video background during performance is realized.

As shown in fig. 5, in the display process of the display screen, it is often necessary to perform transformation processing on the content of the source memory, that is, the output content of the display screen is not a simple copy of the source image, and at this time, the processor needs to perform corresponding transformation on the content of the division area of the source memory, such as image editing, such as rotation, color matching, scaling, and the like. For example, if the divided region corresponding to the source image is not consistent with the resolution of the display screen, the divided region copied to the target memory space needs to be scaled, so that the scaled divided region is consistent with the resolution of the display screen.

According to a specific implementation manner of the embodiment of the present invention, all display screens are grouped according to the positions, the display controllers of each group of display screens are merged, in step one, a target memory space is allocated to each group of display screens, in step S4, image segmentation and memory copy operations are cyclically performed by taking the group of display screens as a unit, and a segmentation region corresponding to each display screen in the group is copied to the target memory space according to the arrangement order of the display screens.

One difficulty faced by large performance dynamic stages is the management of a large number of display screen display controllers. The simplest case is to set one display controller for each display screen, but this requires separately setting a management program and a process for each display screen, for example, allocating a target memory space for each display screen, copying the display controller, and then performing a management process for the next display screen. By adopting the mode, the target memory space corresponding to the display screen can be recycled after being used up, the target memory occupies small space, but the system hardware and wiring are complex, and frequent switching in each display screen control flow is needed, so that the efficiency is low. Another way is to set a display controller for all the display screens, as shown in fig. 5, so that the hardware structure and management procedure of the system are simple, but due to the huge number of display screens involved in a large performance, an excessively large target memory space needs to be allocated; in addition, when the target memory is copied to the display controller, a long time is consumed for each copying due to too large space, and the real-time display requirement required by a large-scale performance cannot be met.

In order to solve the problem of managing a large number of display screen display controllers, a method for grouping the display screens is adopted according to a specific implementation mode of the embodiment of the invention. For convenience of hardware and wiring, the grouping principle is based on where the display is located. Although the display screens are dynamically changed in the performance process, the display screens generally move in a local range, the display screens close to each other are grouped, and the display screens in the same group share one display controller, so that the convenience and the simplicity in wiring can be brought, and the hardware management of the display controller is simplified. After the display screens are grouped, each group is used as a basic unit for allocating a target memory space and performing display controller replication, which brings balance of time and space efficiency. When the target memory space is allocated, for example, the target memory space of 5 display screen groups can be allocated simultaneously according to specific hardware resources, when a certain group is processed, the occupied memory is released, and the memory resources are acquired by other groups to perform corresponding data processing. When memory allocation is performed according to display screen grouping, the corresponding partition area of each display screen in the group needs to be copied to a target memory space according to the display screen arrangement sequence, and finally, the stored content in the target memory space is output to the display controller of the group of display screens.

As shown in fig. 3, for a large-scale performance stereoscopic dynamic stage, the stage modules are divided into many levels, and often have a shielding relationship. For the viewer, only the display screen at the front row can be seen, and the display screen at the back row is blocked, so that the viewer can not see the display screen at the front row. However, if the shielded display screen does not contain any display signal, the shielded display screen becomes a black screen, and the black screen may be displayed in the movement process of the display screen, so that the stage background effect is greatly influenced. One better way to do this is for the back row of displays to multiplex the display information from the front row of displays so that the best stage background effect can be achieved. In the display process of the display screen, the consumption of space and time resources is mainly embodied by distributing a target memory for the display screen, determining the display content of the display screen according to the partition area and writing the display content into the target memory. According to a specific implementation manner of the embodiment of the invention, when the display information of the front-row display screen is multiplexed by the rear-row display screen, the target memory copy operation is not performed; and in the process of outputting the target memory to the display controller, directly multiplexing the display screen which is not shielded at the most front position to group the target memory content. Therefore, when the appropriate display information is set for the rear display screen, the time and space resources of the system are greatly saved.

For the display screens with the shielding relation, the display information of the front display screen can be reused by the rear display screen, and the display efficiency can be improved in the case. Therefore, in order to improve the overall display efficiency of the dynamic stage, according to a specific implementation manner of the embodiment of the present invention, the output order of the display screen groups is arranged, so that the multi-layer arrangement stage having the shielding relationship is output in the order from front to back, and the common target memory space is allocated to the corresponding display screen groups. In step S3, it is first determined whether the display screen packet is occluded, and the occluded display screen packet is touched to directly multiplex the existing content sharing the target memory space. The output sequence of the display screen groups is arranged in advance before the display screen is output, so that all the shielded display screen groups can multiplex the front-row display information. Because the shielding conditions in the large-scale performance dynamic stage are very much, the digital display efficiency of the stage can be greatly improved.

The invention improves the speed of video data processing in a parallel mode. When a large number of display screens are grouped, and multiplexing under the shielding condition is considered, each parallel producer thread processes a frame image of one time point, but the frame image is processed and copied by taking the display screen as a unit, namely, each producer thread is divided into a plurality of batches. Because the batches are used as data processing units, the memory multiplexing under the shielding condition is facilitated. In the same producer thread, display screens of different batches share the target memory space, and when the situation that the memory can be directly reused is found, the image processing and copying time of the batch can be saved. However, when the batch is not divided, in order to ensure that the output video file is in the correct sequence, the consumer thread adopts the strategy of completely writing the data at the point 1 and then writing the data at the point 2, as shown in fig. 6. When each producer thread is divided into a plurality of batches, if the write strategy of the consumer thread is adopted again, the problem of time efficiency reduction exists, as shown in fig. 7. In the figure, each thread task can be divided into three steps of reading video, processing images and writing video, and the image processing time is long, but the image processing can be parallel. The reading and writing of the video can be performed only in series because the reading and writing of the disk are required. As can be seen from fig. 6 and 7, to complete the writing of the i +1 task, i.e. to generate the i +1 frame of the video, it is necessary to ensure that the i frame processing of the video is completed, so the writing operation of the first batch of the i +1 task is delayed until the i task is completely completed. By analogy, the start of each task can be performed only after the previous task is completely completed, and the parallel efficiency is greatly reduced. In fig. 7, after dividing into 3 batches, only 2 tasks are completed within t time. More than 4 tasks may have been done before.

The parallel processing strategy of the present embodiment is improved for the above case: since each batch corresponds to different display screen groups, an independent display controller can be arranged for each display screen group, and a different video file can be generated for each display screen group. When the written files contained in different batches are different due to batch division, the output of each video file is ensured to be carried out according to the frame sequence number. In order to realize target memory multiplexing, in the same producer thread, batches need to be executed in sequence, but the i +1 task execution does not need to wait for the completion of all the batches of the i task, and all the batches before p of the i +1 task can be executed only by ensuring the completion of the batch write operation of the p of the i task. As shown in fig. 8, after the batch parallel algorithm is improved, the read-write time of the disk is fully utilized, and the parallel efficiency is greatly improved.

at least one processor; and the number of the first and second groups,

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in the embodiments of the present disclosure may be implemented by software or hardware. Where the name of an element does not constitute a limitation on the element itself.

It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof.

The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A parallel acceleration display method for a dynamic stage video of a large-scale performance is characterized by comprising the following steps:

s4, circularly executing the following operations on all display screens:

each producer thread is not responsible for writing in the stage screen control video file, and after the step S42 is executed, a buffer area conversion completion mark is set to wait for the consumer thread to output the buffer area; additionally starting a global consumer thread, requesting a target memory according to a time sequence, adding the target memory of the producer thread into the stage screen control video file if the conversion of the corresponding buffer area is completed, and releasing the buffer area;

thirdly, outputting the stage screen control video files containing all the time points to a display controller in sequence;

grouping all the display screens according to the positions, merging the display controllers of each group of display screens, distributing a target memory space for each group of display screens, circularly executing image segmentation and memory copy operation by taking the display screen grouping as a unit in step S4, and copying the corresponding segmentation area of each display screen in the group to the target memory space according to the display screen arrangement sequence; the consumer thread requests a target memory according to time and a display screen output sequence;

generating a different video file for each display screen group; dividing each producer thread into a plurality of batches by taking display screen grouping as a unit, and starting to execute the p-th batch of the thread i when the p-1-th batch of the thread i is completely consumed and the p-th batch of the thread i-1 is also completely consumed; and if the consumer thread detects a buffer conversion completion mark in any producer thread, adding the thread target memory into the corresponding stage screen control video file, and releasing the buffer.

2. The parallel acceleration display method for the large-scale performance dynamic stage videos according to claim 1, wherein for a multi-layer stage with a shielding relationship, the shielded display screens are grouped without performing target memory copy operation; in the process of outputting the target memory to the display controller, directly multiplexing the display screen which is not shielded at the most front position to group the target memory content; arranging the output sequence of the display screen groups to output the multi-layer arrangement stage with the shielding relation from front to back, and distributing the common target memory space for the corresponding display screen groups; in step S4, it is first determined whether the display screen packet is occluded, and the occluded display screen packet is touched to directly multiplex the existing content sharing the target memory space.

3. The method as claimed in claim 2, further comprising the step of editing the image of the segmented area.

4. An electronic device, characterized in that the electronic device comprises:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method of large performance dynamic stage video parallel accelerated display as claimed in any one of the preceding claims 1 to 3.

5. A non-transitory computer readable storage medium storing computer instructions for causing a computer to execute a large performance dynamic stage video parallel accelerated display method according to any one of the preceding claims 1 to 3.