A kind of live all-in-one of medical video image
Technical field
The present invention relates to the transmission of video images technical field for tele-medicine, be specifically related to a kind of live all-in-one of medical video image.
Background technology
In China, the application of the video capture devices such as all kinds of video frequency collection cards, DVR makes tele-medicine efficiency be greatly improved.What many medical treatment R&D teams let slip no opportunity seeks various technological means.For realizing the various medical signal high-efficiency transfer such as medical imaging, video, the compatibility of all kinds of video capture devices, efficiently the most therefore become focus of attention.
In the middle of practical application, we are it can be seen that the video output formats of all kinds of armarium is multifarious, and convention video collecting device is more tends to nonstandard form.When in the face of hospital all kinds of various, during the different armarium of brand, conventional way is to inquire the engineer of this producer, understands video output interface and signal format, the parameter etc. of this equipment.Need to be equipped with special collecting device for this equipment.And when hospital needs the signal receiving miscellaneous equipment, equipment before this is likely to mate.Thus need the work before again repeating, again buy new equipment.Such result one is the waste of fund, and two is owing to these equipment are not owing to customizing production, can only be scattered be placed in by the service of connection devices, it is impossible to fixed installation, not only occupy space, and cause the probability of damage to be also greatly improved.
It addition, in practical application, it has been found that the video image used in tele-medicine has the advantages that field (back of the body) scape is fixed, picture is the least, also there is definition height, stability height and the requirement of low latency.The current live broadcast system towards consumer level market and its coding, coding/decoding method can not be well adapted for the characteristics of demand of tele-medicine.There is the factors such as poor definition, incompatible, easy decay in transmission and format transformation so that in fuzzy, the shake etc. that signal becomes at the image after long-distance transmissions, video quality, strong influence uses.Especially under the more severe communication environments such as wireless network during transmission, often as the reasons such as packet loss cause picture stuck or fuzzy, it is impossible to reach the level required for tele-medicine.
Summary of the invention
The technical problem to be solved in the present invention is: propose a kind of live all-in-one of medical video image processed and solve all kinds of video capture device compatibling problem based on high-definition video signal.
The present invention solves that the technical scheme that above-mentioned technical problem proposes is: a kind of live all-in-one of medical video image, including rack, be provided with mainboard and the power module for power supply in described rack, also include the video management server being positioned at outside rack;
Be sequentially provided with on described mainboard by I/O bus be connected video acquisition unit, video processing unit, video compression coding unit;
Described video acquisition unit has several video signal access interface, and described video acquisition unit is for being connected to the video output terminals of existing various armarium and carrying out multichannel instant raw video signal collection;
Described video processing unit includes video pre-filtering module and HD-SDI transcoding module, and described video processing unit is suitable to become instant for the described multichannel raw video signal transformation of ownership after the pre-treatment the high definition video standard signal of HD-SDI form;
Described video compression coding unit includes video compress arithmetical unit and video encoder, described video compression coding unit is suitable to the high definition video standard signal of the HD-SDI form after being processed by described video processing unit and is compressed and encodes, described video compression coding unit, H.265 coding standard is;
Described video management server is adapted to respond to the access demand of user and sends encoded video stream media to client.
Further, described video pre-filtering module includes that the video resolution sequentially forming signal connection adjusts submodule, video frame rate adjusts submodule, code check adjusts submodule.
Further, described video management server includes access control management module, cache module, stream media network sending module and locally displayed module.
In order to enable to provide relatively stable and higher resolution video under severe communication environment, further, video compress includes that primitive frame memory module, reference frame handling module, reference frame cache module, agglomerate take off module, agglomerate cache module, agglomerate interframe movement trend cache module, present frame reconstructed module, the second present frame reconstructed module, agglomerate motion compensation calculations module, agglomerate interframe deformation prediction module and agglomerate deformation compensation calculation module arithmetical unit.
Further, the primitive frame image being pre-stored with under pure background scene in described primitive frame memory module;
Reference frame handling module captures a two field picture after the N frame of each interval as reference frame, and the span of N is 0-60;
Agglomerate takes off module for taking off out the agglomerate image with actual content in reference frame image and the information of this lumpy pattern being input in agglomerate cache module temporary;
Agglomerate interframe movement prediction module meets the lumpy pattern of lumpy pattern feature in reference frame for gathering the image after being positioned at reference frame searching wherein, and then calculate this lumpy pattern relatively in reference frame time motion vector, and this motion vector is sent in present frame reconstructed module as agglomerate motion prediction vectors;
Present frame reconstructed module is used for constructing virtual current frame image, containing virtual lumpy pattern in virtual current frame image;
Agglomerate motion compensation calculations module is used for gathering actual current frame image, and inspection that actual current frame image and virtual present frame are compared, calculate in virtual lumpy pattern and actual current frame image the alternate position spike between lumpy pattern and combine agglomerate motion prediction vectors and calculate agglomerate motion compensation difference;
Agglomerate interframe deformation prediction module meets the lumpy pattern of lumpy pattern feature in reference frame for gathering the image after being positioned at reference frame searching wherein, and then calculate this lumpy pattern relatively in reference frame time agglomerate deformation tendency parameter, and described agglomerate deformation tendency parameter is exported to the second present frame reconstructed module, described second present frame reconstructed module is used for constructing virtual current frame image, containing virtual lumpy pattern in virtual current frame image, the agglomerate deformation compensation calculation actual current frame image of module collection is also compared with lumpy pattern in aforementioned virtual current frame image, calculate the agglomerate deformation about lumpy pattern and compensate residual value.
Further, described lumpy pattern information, agglomerate motion compensation difference and agglomerate deformation compensate after residual value is encoded by described video encoder and send the cache module to streaming media server, and described primitive frame image is also sent described cache module by primitive frame memory module.
The invention has the beneficial effects as follows:
First the live all-in-one of medical video image in the present invention solves after how the various standards of various armarium output in hospital or non-standard video being acquired and to be processed and is used for live technical problem so that at a terminal, user can watch that multichannel has the video pictures of different content.User can carry out momentarily " movement " by 4G network at mobile phone terminal and watch, and improves Consumer's Experience.
Next also solves and how to meet medical video transmission stability requirement and to can still provide for the technical problem of transmission of high-quality video under conditions of more severe network communications environment, medical video signals collecting trans-coding system original creation in the present invention develops a kind of compressed encoding unit, this compressed encoding unit utilizes medical video to have the advantages that scene is fixing and the least, by first intercepting " agglomerate " of motion, then on clean background frames, it is superimposed with the means of lumpy pattern, reaches reduction transmission of video data volume.
Separately it is worth mentioning that: use identical motion prediction and deformation prediction module at local side with decoding end, after local side meets the adjustment of real image to predicting the outcome, adjustment parameter is transferred to decoding end.It is predicted the outcome by this adjustment parameter and is adjusted by decoding end so that it is the image of final display is identical with the image collected of local side.Transmission of video data volume (only needing transmission to adjust parameter) is further reduced on the premise of ensure that video quality.
Accompanying drawing explanation
Below in conjunction with the accompanying drawings the live all-in-one of medical video image of the present invention is described further.
Fig. 1 is the structured flowchart of the medical treatment live all-in-one of video image in the present invention;
Fig. 2 is the workflow diagram of the medical treatment live all-in-one of video image in the present invention;
Fig. 3 is logical structure and the operation principle schematic diagram of video compress arithmetical unit;
Fig. 4 is that agglomerate takes off the module decomposing schematic representation to reference frame image;
Fig. 5 is logical structure and the video reduction principle schematic diagram of client.
Detailed description of the invention
Embodiment
According to Fig. 1, Fig. 2 and Fig. 3, the live all-in-one of video image in the present invention, including rack, it is provided with mainboard and the power module for power supply in rack, also includes the video management server being positioned at outside rack.
Video management server includes access control management module, cache module, stream media network sending module and locally displayed module.Video management server is adapted to respond to the access demand of user and sends encoded video stream media to client.
Be sequentially provided with on mainboard by I/O bus be connected video acquisition unit, video processing unit, video compression coding unit.
Video acquisition unit has several video signal access interface, and video acquisition unit is for being connected to the video output terminals of existing various armarium and carrying out multichannel instant raw video signal collection.Video acquisition unit includes acquisition module and output module.Acquisition module is the hardware capture card of buying in the present embodiment, multichannel input interface it is provided with on hardware capture card, hardware capture card is connected on the video output terminals of various armarium by coaxial cable, optical fiber cable or other wireline cable, for carrying out the video-on-demand signal collection of multichannel.Output module is to add the process chip being located on hardware capture card and the output port being connected with this process chip, for by by acquisition module instantaneous acquiring to every road raw video signal after sequential arranges, be transferred to video processing unit, transmission Shi Mei road raw video signal is continuous print frame.
Video processing unit includes video pre-filtering module and HD-SDI transcoding module, and video processing unit is suitable to become instant for the multichannel raw video signal transformation of ownership after the pre-treatment the high definition video standard signal of HD-SDI form.Can be as preferably: video pre-filtering module includes that the video resolution sequentially forming signal connection adjusts submodule, video frame rate adjusts submodule, code check adjusts submodule.The resolution of every two field picture of every road raw video signal is all adjusted to 1280 pixels and takes advantage of 720 pixels by video resolution adjusting module.When the resolution of the two field picture of primary signal is less than this resolution, carry out interpolation calculation so that it is expand to 1280 pixels and take advantage of 720 pixels, when the resolution of the two field picture of primary signal is more than this resolution, carry out resolution compression so that it is be reduced to 1280 pixels and take advantage of 720 pixels.Video frame rate adjusting module is for being adjusted the frame per second of the raw video signal after resolution adjustment so that it is having the frame per second of 50HZ-60HZ, code check adjusting module is adjusted the preset value of 1Mps-2Mps to the code check parameter of whole raw video signal.
Video compression coding unit includes video compress arithmetical unit and video encoder, video compression coding unit is suitable to the high definition video standard signal of the HD-SDI form after being processed by video processing unit and is compressed and encodes, video compression coding unit, H.265 coding standard is.
In order to enable to provide relatively stable and higher resolution video under severe communication environment, video compress includes that primitive frame memory module, reference frame handling module, reference frame cache module, agglomerate take off module, agglomerate cache module, agglomerate interframe movement trend cache module, present frame reconstructed module, the second present frame reconstructed module, agglomerate motion compensation calculations module, agglomerate interframe deformation prediction module and agglomerate deformation compensation calculation module arithmetical unit.
Wherein, primitive frame memory module is pre-stored with the primitive frame image under pure background scene.
Reference frame handling module captures a two field picture as reference frame after the N frame of each interval, the span of N is 0-60, reference frame handling module captures after can being preset to be spaced N frame, can also be preset to capture continuously, when network environment preferably or to video quality requirement higher time can be arranged to capture continuously or N value becomes little.It is to capture continuously in the present embodiment, reference frame handling module captures the second two field picture, and (" the second frame " is only for convenience of in conjunction with the accompanying drawings, in a particular application, " the second frame " reality is X+1 frame, lower same) as reference frame, the present embodiment is preset the reference frame captured and is contained within the substantial image (such as the parameter curve in electrocardiogram indicator or the actual scene in operating room) beyond background, view network transmission environment and to the demand of video quality depending on, the reference frame after crawl is deposited in reference frame cache module standby.
Agglomerate takes off module for taking off out the agglomerate image with actual content in reference frame image and the information of this lumpy pattern being input in agglomerate cache module temporary.When taking off as shown in Figure 4, first reference frame image and primitive frame image are resolved into the some macro blocks of various level and some microliths according to colourity and the brightness of image, then overlapping contrast removes repeatable block, retain difference block, the set of the difference block retained is defined as agglomerate, and lumpy pattern information includes the relevant parameters such as aforementioned differences block coordinate in reference frame, colourity, brightness.Supply lumpy pattern information temporary storage in agglomerate cache module to call.As shown in Figure 4, in the present embodiment, agglomerate is five-pointed star case.
Agglomerate interframe movement prediction module collection the 3rd two field picture also finds the lumpy pattern meeting aforementioned lumpy pattern information characteristics wherein, and then calculate agglomerate in the 3rd frame relative to motion vector during at the second frame, and this motion vector is sent in present frame reconstructed module as agglomerate motion prediction vectors, for constructing virtual current frame image (the 4th frame), containing virtual agglomerate in virtual current frame image.
Agglomerate motion compensation calculations module gathers actual current frame image, and inspection that actual current frame image and virtual present frame are compared, calculate in virtual agglomerate and actual current frame image the alternate position spike between agglomerate and combine agglomerate motion prediction vectors and calculate agglomerate motion compensation difference.
Agglomerate interframe deformation prediction module collection the 3rd two field picture also finds the lumpy pattern meeting aforementioned lumpy pattern information characteristics wherein, calculate in the 3rd frame lumpy pattern relative to the agglomerate deformation tendency parameter of the lumpy pattern information in agglomerate cache module, when calculating agglomerate deformation tendency parameter, investigate the chroma luminance parameter of each microlith in the deformation parameter of agglomerate pattern edge profile and lumpy pattern.
Second present frame reconstructed module is used for constructing virtual current frame image, and the second present frame reconstructed module is according to the shape of lumpy pattern, colourity and brightness in deformation tendency parameter prediction reconstruct present frame.The agglomerate deformation compensation calculation actual current frame image of module collection is also compared with lumpy pattern in aforementioned virtual current frame image, calculates the agglomerate deformation about lumpy pattern and compensates residual value.
Lumpy pattern information, agglomerate motion compensation difference and agglomerate deformation compensate after residual value is encoded by video encoder and send the cache module to streaming media server, and primitive frame image is also sent cache module by primitive frame memory module.
The video data being buffered in cache module, for responding the access request of terminal use and judging whether the authority of user mates with this request, if coupling, further is sent to user by stream media network sending module by the access control management module in video management server.
By module locally displayed in video management server, in the present embodiment, locally displayed module includes liquid crystal display, buffer location and decoding unit.Can be in local preview video image information by locally displayed module.It is positioned in the client of network opposite side and also there is identical display device, buffer location and decoding unit.
As it is shown in figure 5, decoding unit includes decoder module, the second primitive frame memory module, the second agglomerate cache module, agglomerate motion compensation feedback module, agglomerate deformation Compensation Feedback module, the second agglomerate interframe movement prediction module, agglomerate interframe movement trend cache module, the second agglomerate interframe deformation prediction module, agglomerate interframe deformation tendency cache module and the 3rd present frame reconstructed module.
Second primitive frame memory module reads primitive frame image from buffer location, and in this, as first two field picture during video playback, decoder module reads lumpy pattern information, agglomerate motion compensation difference and agglomerate deformation from buffer location and compensates residual value and be separately sent to agglomerate cache module, agglomerate motion compensation feedback module and agglomerate deformation Compensation Feedback module after the decoding.
The lumpy pattern information received, agglomerate motion compensation difference and agglomerate deformation compensation residual value is loaded into present frame reconstructed module and is used for reconstructing current frame image by the second agglomerate cache module, agglomerate motion compensation feedback module and agglomerate deformation Compensation Feedback module respectively.
Second agglomerate interframe movement prediction module and the agglomerate motion prediction vectors of the second agglomerate interframe deformation prediction module and agglomerate deformation tendency parameter result of calculation load to the 3rd present frame reconstructed module the most respectively for reconstructing current frame image.Agglomerate motion prediction vectors and agglomerate deformation tendency parameter are also separately input to agglomerate interframe movement trend cache module and agglomerate interframe deformation tendency cache module by the second agglomerate interframe movement prediction module and the second agglomerate interframe deformation prediction module, in case occur packet loss, frame losing time can before frame trend continue reconstruct image.
Second agglomerate interframe movement prediction module and the second agglomerate interframe deformation prediction module the most constantly gather the two field picture reconstructing out, are output as sky at initial light with when cannot calculate motion or deformation tendency after gathering primitive frame image and the second two field picture.
The 3rd present frame reconstructed module in decoding unit first reads the primitive frame image in the second primitive frame memory module when reconstruct, read the lumpy pattern information in the second agglomerate cache module again, compensate residual value and carry out the picture of present frame by combining agglomerate motion prediction vectors and agglomerate deformation tendency and agglomerate motion compensation difference and agglomerate deformation and reconstruct and export.
The present invention is not limited to above-described embodiment, and the technical scheme of each embodiment above-mentioned of the present invention can form new technical scheme with combined crosswise each other, and the technical scheme that the most all employing equivalents are formed all falls within the protection domain of application claims.