CN103428482A - Processing method and device of image data - Google Patents
Processing method and device of image data Download PDFInfo
- Publication number
- CN103428482A CN103428482A CN2012101501643A CN201210150164A CN103428482A CN 103428482 A CN103428482 A CN 103428482A CN 2012101501643 A CN2012101501643 A CN 2012101501643A CN 201210150164 A CN201210150164 A CN 201210150164A CN 103428482 A CN103428482 A CN 103428482A
- Authority
- CN
- China
- Prior art keywords
- bounds
- image
- view data
- described image
- filling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses a processing method and device of image data. The processing method comprises the steps that the out-of-bound information of images of a reference frame matching block is determined; the image data of the image part which is not out of bound are acquired from an external memory according to the out-of-bound information; the image data of the image part which is not out of bound are used for filling the images of an image out-of-bound part. According to the processing method and device of the image data, the image data of the image part which is not out of bound are only acquired from the external memory, the data amount of the image data acquired from the external memory is reduced, and therefore the bandwidth requirement for an image processing system of motion compensation in the process of image data processing is lowered.
Description
Technical field
The present invention relates to the communications field, in particular to a kind of processing method and device of view data.
Background technology
Existing video encoding and decoding standard is that to take image macro or piece be that basic algorithm is realized.In motion compensation, take image macro or piece as the basis find best matching blocks in reference frame image, be used for reducing the spatial domain redundancy in video sequence.In the motion compensation calculations of the more violent image of motion, the motion vector obtained likely points to outside image, and partial data reality does not exist, and needs, with the boundary pixel of image, the pixel outside image is done to filling (padding) and process in motion compensation.
In correlation technique, the filling of the outer data of image is processed to the two kinds of modes that comprise.A kind of is software mode, and software copies the pixel value of image boundary to image outer part, and during motion compensation, reading out data does not need to do other processing, and this mode processing speed is slower, reduces decoding efficiency.Another mode is hardware handles, by independent packing module, the data outside image is replaced with boundary pixel, and this mode need to expend certain hardware resource, and control flow is more complicated also.
In correlation technique, the processing speed of view data reduces the problem of decoding efficiency slowly; And the processing of view data expends the problem of hardware resource and control flow complexity, effective solution is not yet proposed at present.
Summary of the invention
For in image processing, having relatively high expectations of data bandwidth being affected the problem of the efficiency of video decode; And the processing of view data expends the problem of hardware resource and control flow complexity, the invention provides a kind of processing method and device of view data, at least one of to address the above problem.
According to an aspect of the present invention, provide a kind of processing method of view data, having comprised: the out-of-bounds information of determining the image of reference frame match block; Obtain the not view data of out-of-bounds part of described image according to described out-of-bounds information from external memory storage; Utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part.
Preferably, according to described out-of-bounds information, obtain the not view data of out-of-bounds part of described image from external memory storage, comprising: according to described out-of-bounds information, directly from described external memory storage, read the not view data of out-of-bounds part of described image.
Preferably, according to described out-of-bounds information, obtain the not view data of out-of-bounds part of described image from external memory storage, comprising: send the DMA request that carries described out-of-bounds information to outside dma module; Receive the not view data of out-of-bounds part of described image that described outside DMA reads from described external memory storage according to described out-of-bounds information.
Preferably, in the situation that described image out-of-bounds are partly horizontal direction, utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part, comprise: obtaining described image not during the view data of out-of-bounds parts from external memory storage, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part; The described image that will carry out after the horizontal process filling is stored to internal storage.
Preferably, in the situation that described image out-of-bounds are partly vertical direction, utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part, comprising: by the described image that obtains not the view data of out-of-bounds part be stored to internal storage; The view data of utilizing described image boundary member is in the vertical direction carried out the vertical direction filling to the image of described image out-of-bounds part, and the described figure that will carry out after the vertical direction filling is stored to described internal storage.
Preferably, in the situation that described image out-of-bounds part comprises vertical direction and horizontal direction simultaneously, utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part, comprise: obtaining described image not during the view data of out-of-bounds parts from external memory storage, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part, and the described image that will carry out after the horizontal process filling is stored to internal storage; The view data that the described image boundary member in the vertical direction after the level filling is carried out in utilization is carried out the vertical direction filling to the image of described image out-of-bounds part, and the described figure that will carry out after the vertical direction filling is stored to described internal storage.
Preferably, will carry out after described image after the horizontal direction filling is stored to internal storage, described method also comprises: the view data that will carry out the described image boundary member in the vertical direction after the level filling is stored to inner buffer; Utilization is carried out before described image boundary image data in the vertical direction after the level filling carries out the vertical direction filling to the image of described image out-of-bounds part, and described method also comprises: from described inner buffer, read described view data of carrying out the described image boundary member in the vertical direction the level filling.
Preferably, determine the out-of-bounds information of out-of-bounds data of the view data of reference frame match block, comprising: according to the described image out-of-bounds of the location positioning of the motion vector obtained and described reference frame match block; Determine the out-of-bounds positional information of described image, wherein, described positional information comprises: horizontal out-of-bounds and/or vertical out-of-bounds; Determine the not position of view data in described external memory storage of out-of-bounds part of described image.
Preferably, before determining the out-of-bounds information of image of reference frame match block, also comprise: the block size information that obtains described reference frame match block.
According to another aspect of the present invention, provide a kind of processing unit of view data, having comprised: determination module, for the out-of-bounds information of the image of determining the reference frame match block; Acquisition module, for obtaining the not view data of out-of-bounds part of described image according to described out-of-bounds information from external memory storage; Packing module, for utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part.
Preferably, described acquisition module obtains the not view data of out-of-bounds part of described image in such a way from described external memory storage: according to described out-of-bounds information, directly from described external memory storage, read the not view data of out-of-bounds part of described image.
Preferably, described acquisition module comprises: transmitting element, for send the DMA request that carries described out-of-bounds information to outside dma module; Receiving element, for receiving the not view data of out-of-bounds part of described image that described outside DMA reads from described external memory storage according to described out-of-bounds information.
Preferably, described packing module comprises: the first filler cells, for obtaining described image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part; The first memory cell, be stored to internal storage for the described image that will carry out after the level filling.
Preferably, described packing module comprises: the second memory cell, for the described image by obtaining not the view data of out-of-bounds parts be stored to internal storage; The second filler cells, carry out the vertical direction filling for the view data of utilizing described image boundary member in the vertical direction to the image of described image out-of-bounds part; Described the second memory cell, also be stored to described internal storage for the described figure that will carry out after the vertical direction filling.
Preferably, described packing module comprises: horizontal filler cells, for obtaining described image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part; The 3rd memory cell, be stored to internal storage for the described image that will carry out after the level filling; Vertical filler cells, the view data of carrying out the described image boundary member in the vertical direction after the level filling for utilization is carried out the vertical direction filling to the image of described image out-of-bounds part; Described the 3rd memory cell, also be stored to described internal storage for the described figure that will carry out after the vertical direction filling.
Preferably, described packing module also comprises: the 4th memory cell is stored to inner buffer for the described image boundary image data in the vertical direction that will carry out after the level filling; Reading unit, for reading described described image boundary image data in the vertical direction of carrying out the level filling from described inner buffer.
By the present invention, after determining the out-of-bounds information of image of reference frame match block, according to out-of-bounds information, obtain the not view data of out-of-bounds part of image from external memory storage, and utilize image not the view data blank map of out-of-bounds part as the image of out-of-bounds part.Only from external memory storage, obtain the not view data of out-of-bounds part of image, reduced to obtain the data volume of view data in the external memory storage, thereby reduced the bandwidth demand of motion compensation to image processing system in the view data processing.
The accompanying drawing explanation
Accompanying drawing described herein is used to provide a further understanding of the present invention, forms the application's a part, and schematic description and description of the present invention the present invention does not form inappropriate limitation of the present invention for explaining.In the accompanying drawings:
Fig. 1 is the flow chart according to the processing method of the view data of the embodiment of the present invention;
Fig. 2 is the schematic diagram according to the image coboundary out-of-bounds of the embodiment of the present invention;
Fig. 3 is the schematic diagram according to the image lower boundary out-of-bounds of the embodiment of the present invention;
Fig. 4 is the schematic diagram according to the processing unit of the view data of the embodiment of the present invention;
Fig. 5 is the schematic diagram according to the processing unit of a kind of preferred view data of the embodiment of the present invention;
Fig. 6 is the schematic diagram according to a kind of preferred packing module of the embodiment of the present invention;
Fig. 7 is the schematic diagram according to the another kind of preferred packing module of the embodiment of the present invention;
Fig. 8 is the schematic diagram according to another preferred packing module of the embodiment of the present invention;
Fig. 9 is the schematic diagram according to another preferred packing module of the embodiment of the present invention;
Figure 10 is the schematic diagram according to the motion compensation unit of the embodiment of the present invention;
Figure 11 is the schematic flow sheet according to the method for out-of-bounds pixel filling in the motion compensation of the embodiment of the present invention;
Figure 12 is the schematic diagram according to PADDING in the image coboundary out-of-bounds situation of the embodiment of the present invention;
Figure 13 is the schematic diagram according to PADDING in the image lower boundary out-of-bounds situation of the embodiment of the present invention.
Embodiment
Hereinafter with reference to accompanying drawing, also describe the present invention in detail in conjunction with the embodiments.It should be noted that, in the situation that do not conflict, embodiment and the feature in embodiment in the application can combine mutually.
Embodiment mono-
According to the embodiment of the present invention, a kind of processing method of view data is provided, reduced the bandwidth demand of motion compensation to image processing system during view data is processed.
Fig. 1 is the flow chart according to the processing method of the view data of the embodiment of the present invention, and as shown in Figure 1, the method mainly comprises following step (step S102-step S106):
Step S102, determine the out-of-bounds information of the image of reference frame match block;
Step S104 obtains the not view data of out-of-bounds part of image according to out-of-bounds information from external memory storage;
Step S106, utilize image not the view data blank map of out-of-bounds parts as the image of out-of-bounds part.
Pass through the embodiment of the present invention, after determining the out-of-bounds information of image of reference frame match block, obtain the not view data of out-of-bounds parts of image according to out-of-bounds information from external memory storage, and utilize image not the view data blank map of out-of-bounds part as the image of out-of-bounds part.Only from external memory storage, obtain the not view data of out-of-bounds part of image, reduced to obtain the data volume of view data in the external memory storage, thereby reduced the bandwidth demand of motion compensation to image processing system in the view data processing.
In a preferred implementation of the embodiment of the present invention, before determining the out-of-bounds information of image of reference frame match block, can also obtain the block size information of reference frame match block, divide to support piece in various video standards, the size of the image block of support is from 4x4 to 16x16 etc.
In embodiments of the present invention, while determining the out-of-bounds information of out-of-bounds data of view data of reference frame match block, can be according to the location positioning image out-of-bounds of the motion vector obtained and reference frame match block, further can determine the out-of-bounds positional information of image, (this position positional information can comprise horizontal out-of-bounds and/or vertical out-of-bounds, the image out-of-bounds can comprise: the out-of-bounds of horizontal direction, the out-of-bounds of vertical direction, and the out-of-bounds of horizontal direction and vertical direction), can also determine the not view data position in memory externally of out-of-bounds parts of image.
After determining the image out-of-bounds of reference frame match block, the image that can from external memory storage, obtain the reference frame match block is the view data of out-of-bounds part not, and the view data of image out-of-bounds part is carried out to motion compensation process.In a preferred implementation of the embodiment of the present invention, obtain image not during the view data of out-of-bounds parts from external memory storage according to out-of-bounds information, motion compensating system can be according to the direct reading images view data of out-of-bounds part not from external memory storage of out-of-bounds information; Perhaps, motion compensating system sends the DMA request that carries out-of-bounds information to outside dma module, and receives the not view data of out-of-bounds part of image that outside DMA reads from external memory storage according to out-of-bounds information.By the embodiment of the present invention, realized that neatly motion compensating system obtains the not view data of out-of-bounds part of image from external memory storage.
In the situation that the image out-of-bounds are partly horizontal direction, when utilizing image the view data blank map of out-of-bounds parts not being as the image of out-of-bounds part, can be directly obtaining image not during the view data of out-of-bounds part from external memory storage, the view data of utilizing image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of image out-of-bounds part, and the image that will carry out after the horizontal process filling is stored to internal storage.For example, during the horizontal direction out-of-bounds, can judge the number of the pixel of out-of-bounds, then find the pixel on the border of image, with the pixel pixel filling that image is outer on this border.
In the situation that the image out-of-bounds are partly vertical direction, when utilizing image the view data blank map of out-of-bounds parts not being as the image of out-of-bounds part, can by the image that obtains not the view data of out-of-bounds part be stored to internal storage, and the view data of utilizing image boundary member in the vertical direction carries out the vertical direction filling to the image of image out-of-bounds part, the figure that will carry out after the vertical direction filling is stored to internal storage.
In the situation that image out-of-bounds parts comprises vertical direction and horizontal direction simultaneously, when utilizing image the view data blank map of out-of-bounds part is not as the image of out-of-bounds part, can first carry out the level filling, after vertically fill.In a preferred enforcement side of the embodiment of the present invention, can obtain image not during the view data of out-of-bounds parts from external memory storage, the view data of utilizing image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of image out-of-bounds part, and the image that will carry out after the horizontal process filling is stored to internal storage; Then, utilize the view data carry out the image boundary member in the vertical direction after the level filling to carry out the vertical direction filling to the image of image out-of-bounds part, the figure that will carry out after the vertical direction filling is stored to internal storage.
After carrying out the level filling, the view data of image boundary member in the vertical direction will change, when carrying out the filling of vertical direction, need in memory internally to read the view data of carrying out the image boundary member in the vertical direction after the level filling, for fear of above-mentioned read to process affect handling property, in a preferred implementation of the embodiment of the present invention, to carry out after image after the horizontal process filling is stored to internal storage, the view data of carrying out the image boundary member in the vertical direction after the level filling can be stored to inner buffer.The view data of the boundary member in the vertical direction of the image after the level filling is carried out in utilization is carried out vertical direction while filling to the image of image out-of-bounds parts, can be first internally buffer memory read the image boundary image data in the vertical direction of carrying out after the level filling.
In specific implementation process, the filling of image pixel can be divided into directions X and Y-direction, that is, and and two kinds of horizontal direction and vertical direction.Wherein, the filling of directions X can by image not the pixel the most left or rightmost circle of out-of-bounds parts carry out copying of horizontal direction, then be stored in internal random memory (Random Access Memory, referred to as RAM); The filling of Y-direction be by image not out-of-bounds parts go up most or bottom the pixel on boundary carry out copying of vertical direction.The image of Y-direction is filled with respect to the image of directions X and is filled comparatively complexity, the one-row pixels on the border of Y-direction can be carried out to mark, pixel after directions X is filled is put in buffer memory (buffer), by the data stuffing in buffer for image of vertical direction out-of-bounds part simultaneously.Concrete, can computed image in the horizontal direction with the scope of the out-of-bounds pixel of vertical direction, then the reference frame image pixel of out-of-bounds does not read in motion compensating system from external memory storage, first carry out the filling of horizontal direction, and by the picture element caching of coboundary or lower boundary in inner register, then the out-of-bounds pixel is filled to processing, the view data of buffer memory is stored into to the zone of not filling up of internal RAM.
For example, the size of match block is X*Y, and the size in the image of the match block of searching is a*b, and the number of pixels of out-of-bounds so in the horizontal direction is that (X-a) is individual, and the out-of-bounds number of pixels on vertical direction is that (Y-b) is individual.Can utilize two-dimentional dma module that the pixel in image (not the view data of out-of-bounds) is read from external memory storage, read altogether b capable, every row a pixel.If match block is the top out-of-bounds, as shown in Figure 2, the line number of out-of-bounds be (Y-b) OK, while carrying out motion compensation, in the time of reading out data by the 0th walk to Y-b that the capable data of b-1 are put into internal RAM walk to Y-1 capable in, simultaneously by the deposit data of the 0th row in internal register, and by the data stuffing of the 0th row to the 0th of RAM walk to Y-b-1 capable in.If the images match piece is following out-of-bounds, as shown in Figure 3, the line number of out-of-bounds be (Y-b) OK, while carrying out motion compensation, by the 0th walk to the capable data of b-1 be put into the 0th of internal RAM walk to b-1 capable in, then the b that capable view data is filled into RAM by b-1 to walk to Y-1 capable.
Image has frame and two kinds of modes, and a two field picture can be comprised of field, top and field, the end, and two field picture is decoding separately, finally forms a frame image data.In embodiments of the present invention, can support piece in various video standards to divide, the size of the image block of support from 4x4 to 16x16 etc., is supported field mode simultaneously.
In out-of-bounds situation in vertical direction, can under different piece is divided, identify the one-row pixels on border, and this row pixel is stored in inner buffer, while carrying out motion compensation, the border data line of storing in inner buffer is filled in data RAM.
According to the embodiment of the present invention, also provide a kind of processing unit of view data, in order to realize the processing method of the view data that the above embodiment of the present invention provides.
Fig. 4 is the schematic diagram according to the processing unit of the view data of the embodiment of the present invention, and as shown in Figure 4, this device mainly comprises: determination module 10, acquisition module 20 and packing module 30.Wherein, determination module 10, for the out-of-bounds information of the image of determining the reference frame match block; Acquisition module 20, be coupled with determination module 10, for according to out-of-bounds information, from external memory storage, obtaining the not view data of out-of-bounds part of image; Packing module 30, be coupled with acquisition module 20, for utilize image not the view data blank map of out-of-bounds parts as the image of out-of-bounds part.
Pass through the embodiment of the present invention, after the out-of-bounds information of the image of determination module 10 definite reference frame match block, acquisition module 20 obtains the not view data of out-of-bounds parts of image according to out-of-bounds information from external memory storage, and by packing module 30 utilize image not the view data blank map of out-of-bounds part as the image of out-of-bounds part.Acquisition module 20 only obtains the not view data of out-of-bounds part of image from external memory storage, has reduced to obtain the data volume of view data in the external memory storage, thereby has reduced the bandwidth demand of motion compensation to image processing system in the view data processing.
In a preferred implementation of the embodiment of the present invention, acquisition module 20 obtains image not during the view data of out-of-bounds parts according to out-of-bounds information from external memory storage, can be according to the direct reading images view data of out-of-bounds part not from external memory storage of out-of-bounds information.Perhaps, in device as shown in Figure 5, acquisition module 20 can comprise: transmitting element 202, for send the DMA request that carries out-of-bounds information to outside dma module; Receiving element 204, be coupled with transmitting element 202, for receive outside DMA according to out-of-bounds information from the image that reads of the external memory storage view data of out-of-bounds part not.By the embodiment of the present invention, realized that neatly acquisition module 20 obtains the not view data of out-of-bounds part of image from external memory storage.
In embodiments of the present invention, while determining the out-of-bounds information of out-of-bounds data of view data of reference frame match block, can be according to the location positioning image out-of-bounds of the motion vector obtained and reference frame match block, further can determine the out-of-bounds positional information of image, this position positional information can comprise horizontal out-of-bounds and/or vertical out-of-bounds, can also determine the not view data position in memory externally of out-of-bounds part of image.The out-of-bounds position can comprise: horizontal out-of-bounds and/or vertical out-of-bounds.
In the situation that the image out-of-bounds are partly horizontal direction, when packing module 30 utilizes image the view data blank map of out-of-bounds parts is not as the image of out-of-bounds part, can be directly obtaining image not during view data partly of out-of-bounds from external memory storage.And the view data of being utilized image boundary member in the horizontal direction by packing module 30 carries out the horizontal direction filling to the image of image out-of-bounds part, and the image that will carry out after the horizontal process filling is stored to internal storage.Therefore, in device as shown in Figure 6, packing module 30 can comprise: the first filler cells 302, for obtaining image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of image out-of-bounds part; The first memory cell 304, be coupled with the first filler cells 302, for the described image that will carry out after the level filling, is stored to internal storage.
In the situation that the image out-of-bounds are partly vertical direction, when packing module 30 utilizes image the view data blank map of out-of-bounds parts is not as the image of out-of-bounds part, can by the image that obtains not the view data of out-of-bounds part be stored to internal storage, and the view data of utilizing image boundary member in the vertical direction carries out the vertical direction filling to the image of image out-of-bounds part, the figure that will carry out after the vertical direction filling is stored to internal storage.Therefore, in device as shown in Figure 7, packing module 30 can comprise: the second memory cell 306, for by obtain image not the view data of out-of-bounds parts be stored to internal storage; The second filler cells 308, be coupled with the second memory cell 306, for the view data of utilizing image boundary member in the vertical direction, the image of image out-of-bounds part carried out to the vertical direction filling.And the figure that will be carried out after the vertical direction filling by the second memory cell 306 is stored to internal storage.
In the situation that image out-of-bounds parts comprises vertical direction and horizontal direction simultaneously, when utilizing image the view data blank map of out-of-bounds part is not as the image of out-of-bounds part, can first carry out the level filling, after vertically fill.In a preferred enforcement side of the embodiment of the present invention, as shown in Figure 8, packing module 30 can comprise: horizontal filler cells 310, for obtaining image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of image out-of-bounds part; The 3rd memory cell 312, be stored to internal storage for the image that will carry out after the level filling; Vertical filler cells 314, the view data of carrying out the image boundary member in the vertical direction after the level filling for utilization is carried out the vertical direction filling to the image of image out-of-bounds part.And the figure that will be carried out after the vertical direction filling by the 3rd memory cell 312 is stored to internal storage.
After carrying out the level filling, the view data of image boundary member in the vertical direction will change, when carrying out the filling of vertical direction, need in memory internally to read the view data of carrying out the image boundary member in the vertical direction after the level filling, for fear of above-mentioned read to process affect handling property, in a preferred implementation of the embodiment of the present invention, as shown in Figure 9, packing module 30 can also comprise: the 4th memory cell 316, with horizontal filler cells 310, be coupled, be stored to inner buffer for the image boundary image data in the vertical direction that will carry out after the level filling, reading unit 318, be coupled with vertical filler cells 314, reads the image boundary image data in the vertical direction of carrying out after the level filling for buffer memory internally.
In a preferred implementation of the embodiment of the present invention, before determining the out-of-bounds information of image of reference frame match block, can also obtain the block size information of reference frame match block, divide to support piece in various video standards, the size of the image block of support is from 4x4 to 16x16 etc.
Embodiment bis-
According to the embodiment of the present invention, the method for out-of-bounds pixel filling in a kind of motion compensation is provided, can be applied in video encoding and decoding standard.If motion compensating module is when the reference frame match block view data out-of-bounds of reading, motion compensation unit is the valid data in a reading images from outside memory, then motion compensation unit inner with image in pixel carry out padding to outside, effectively reduce the bandwidth demand to outside memory.
In motion compensation unit as shown in figure 10, mainly comprise: pretreatment module 40, dma module 50 and PADDING module 60.Wherein, pretreatment module 40 is equivalent to above-mentioned determination module 10, for motion vector and the macro block position according to input, and whether out-of-bounds of computed image, and the situation of the number of pixels of out-of-bounds when the image out-of-bounds and other out-of-bounds.Dma module (inside) 50 is equivalent to above-mentioned acquisition module 20, the DMA request is initiated to outside dma module in memory location for the externally memory of the boundary pixel according to calculating, by outside dma module, the data in image is moved to motion compensation unit from outside dma module.PADDING module 60 is equivalent to above-mentioned packing module 30, in reading out data according to the out-of-bounds situation of horizontal direction, carry out the filling of horizontal direction simultaneously, the filling of horizontal direction and the placement of data are carried out simultaneously.After the data provisional capital in image is placed in internal RAM, if there are vertical out-of-bounds, PADDING module 60 is also for carrying out the data padding of vertical row.
Image has frame and two kinds of modes, and a two field picture can be comprised of field, top and field, the end, and two field picture is decoding separately, finally forms a frame image data.In embodiments of the present invention, the size of the image block that motion compensation unit is supported from 4x4 to 16x16 etc..Block size information can obtain by video in code stream, according to the block size of knowing, carries out the processing corresponding with block size.In the invention process, also support field mode.
In out-of-bounds situation in vertical direction, can under dividing, different piece identify the one-row pixels on border, and this row pixel is stored in data b uffer, while carrying out motion compensation, the border data line of buffer memory in data b uffer is filled in data RAM.
While carrying out motion compensation process, can go out according to the position calculation of motion vector and current block in the horizontal direction the out-of-bounds number of pixels with vertical direction, determine it is the situation of upper out-of-bounds, lower out-of-bounds, left out-of-bounds or right out-of-bounds simultaneously.And calculate the memory location of the externally memory of boundary pixel.Obtain the reference frame data by inner DMA.If left out-of-bounds or right out-of-bounds, will carry out the padding operation of the out-of-bounds pixel of horizontal direction.If left out-of-bounds, carry out padding by leftmost pixel, and the pixel data after padding is put in RAM or buffer.If right out-of-bounds, need that rightmost pixel is done to padding and process.Upper out-of-bounds or lower out-of-bounds need to be the padding of the out-of-bounds pixel of vertical direction.Now need the line number according to out-of-bounds on vertical direction, determine a line view data on border on vertical direction, and by data placement in inner buffer.Data in other image are put in internal RAM.On top, in the padding process on boundary, relate to the wrap operation of RAM addr.Address need to rap around to the address location of out-of-bounds the data stuffing in buffer is entered.
Below in conjunction with Figure 11, the method for out-of-bounds pixel filling in the motion compensation of the embodiment of the present invention is described, Figure 11 shows the handling process of the method for reference frame match block out-of-bounds pixel p adding in motion compensation.
As shown in figure 11, at first the whether out-of-bounds of reference block data are obtained in judgement, if not out-of-bounds, the data that will read from outside directly are put into internal RAM according to the order read, the interpolating module that Ti is follow-up Gong Give.If out-of-bounds, do different processing according to different situations.
In the situation that only have the left side or the right out-of-bounds, only need to be according to the number of pixels of out-of-bounds, the pixel of finding out border is carried out level and is copied, and is put in inner RAM and gets final product.In the situation that other, need to carry out the padding operation of vertical direction, if have the left side or the right out-of-bounds, also need first to carry out the padding of level, the data stuffing after horizontal padding for the part of out-of-bounds not, if the pixel on border can also be carried out buffer memory, the outer data of image directly use the data stuffing of buffer memory buffer in RAM.
The padding of coboundary and lower boundary processes different.8 ~ 15 row as shown in figure 12 are pixels in image, and 0 ~ 7 behavior image exterior pixel, need to carry out padding, and the data of 0 ~ 7 row part need to be used the data stuffing of eighth row, and the order of filling is as shown in the numbering on the left side in 12 figure.As shown in figure 12, the pixel column in the blank map picture at first, one-row pixels that simultaneously will border is preserved, and then will fill the address wraparound, and the data stuffing to the of eighth row 0 is walked to the 7th row.As shown in figure 13, the following pixel out-of-bounds of images match piece, 0 ~ 7 row data are pixel in image, and eighth row is the pixel column on border, and the address of filling is consistent with the data line number, but need to be by the eighth row data buffer storage, and then order is filled in RAM.
By the embodiment of the present invention, can support 4*4 in the various video standards piece dividing mode to 16*16, while carrying out motion compensation, the pixel data in a reading images, not extra read unwanted data, effectively saved bandwidth.Only add the resources seldom such as data line buffer, taking of hardware is fewer, with respect to correlation technique, on efficiency, has greatly improved.
From above description, can find out, the present invention has realized following technique effect: after determining the out-of-bounds information of image of reference frame match block, obtain the not view data of out-of-bounds parts of image according to out-of-bounds information from external memory storage, and utilize image not the view data blank map of out-of-bounds part as the image of out-of-bounds part.While carrying out motion compensation, the pixel data in a reading images, not extra read unwanted data, effectively saved bandwidth.Simultaneously, only add the resources seldom such as data line buffer, taking of hardware is fewer, with respect to correlation technique, on efficiency, has greatly improved.
Obviously, those skilled in the art should be understood that, above-mentioned each module of the present invention or each step can realize with general calculation element, they can concentrate on single calculation element, perhaps be distributed on the network that a plurality of calculation elements form, alternatively, they can be realized with the executable program code of calculation element, thereby, they can be stored in storage device and be carried out by calculation element, and in some cases, can carry out step shown or that describe with the order be different from herein, perhaps they are made into respectively to each integrated circuit modules, perhaps a plurality of modules in them or step being made into to the single integrated circuit module realizes.Like this, the present invention is not restricted to any specific hardware and software combination.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (16)
1. the processing method of a view data, is characterized in that, comprising:
Determine the out-of-bounds information of the image of reference frame match block;
Obtain the not view data of out-of-bounds part of described image according to described out-of-bounds information from external memory storage;
Utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part.
2. method according to claim 1, is characterized in that, according to described out-of-bounds information, obtains the not view data of out-of-bounds part of described image from external memory storage, comprising:
Directly from described external memory storage, read the not view data of out-of-bounds part of described image according to described out-of-bounds information.
3. method according to claim 1, is characterized in that, according to described out-of-bounds information, obtains the not view data of out-of-bounds part of described image from external memory storage, comprising:
Send the DMA request that carries described out-of-bounds information to outside dma module;
Receive the not view data of out-of-bounds part of described image that described outside DMA reads from described external memory storage according to described out-of-bounds information.
4. according to the method in any one of claims 1 to 3, it is characterized in that, in the situation that described image out-of-bounds are partly horizontal direction, utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part, comprising:
Obtaining described image not during the view data of out-of-bounds parts from external memory storage, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part;
The described image that will carry out after the horizontal process filling is stored to internal storage.
5. according to the method in any one of claims 1 to 3, it is characterized in that, in the situation that described image out-of-bounds are partly vertical direction, utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part, comprising:
By the described image that obtains not the view data of out-of-bounds part be stored to internal storage;
The view data of utilizing described image boundary member is in the vertical direction carried out the vertical direction filling to the image of described image out-of-bounds part, and the described figure that will carry out after the vertical direction filling is stored to described internal storage.
6. according to the method in any one of claims 1 to 3, it is characterized in that, in the situation that described image out-of-bounds part comprises vertical direction and horizontal direction simultaneously, utilize described image not the view data of out-of-bounds part fill the image of described image out-of-bounds part, comprising:
Obtaining described image not during the view data of out-of-bounds parts from external memory storage, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part, and the described image that will carry out after the horizontal process filling is stored to internal storage;
The view data that the described image boundary member in the vertical direction after the level filling is carried out in utilization is carried out the vertical direction filling to the image of described image out-of-bounds part, and the described figure that will carry out after the vertical direction filling is stored to described internal storage.
7. method according to claim 6, is characterized in that,
To carry out after described image after the horizontal direction filling is stored to internal storage, described method also comprises: the view data that will carry out the described image boundary member in the vertical direction after the level filling is stored to inner buffer;
Utilization is carried out before described image boundary image data in the vertical direction after the level filling carries out the vertical direction filling to the image of described image out-of-bounds part, and described method also comprises: from described inner buffer, read described view data of carrying out the described image boundary member in the vertical direction the level filling.
8. according to the method in any one of claims 1 to 3, it is characterized in that, determine the out-of-bounds information of out-of-bounds data of the view data of reference frame match block, comprising:
According to the described image out-of-bounds of the location positioning of the motion vector obtained and described reference frame match block;
Determine the out-of-bounds positional information of described image, wherein, described positional information comprises: horizontal out-of-bounds and/or vertical out-of-bounds;
Determine the not position of view data in described external memory storage of out-of-bounds part of described image.
9. according to the method in any one of claims 1 to 3, it is characterized in that, before determining the out-of-bounds information of image of reference frame match block, also comprise:
Obtain the block size information of described reference frame match block.
10. the processing unit of a view data, is characterized in that, comprising:
Determination module, for the out-of-bounds information of the image of determining the reference frame match block;
Acquisition module, for obtaining the not view data of out-of-bounds part of described image according to described out-of-bounds information from external memory storage;
Packing module, for utilize described image not the view data of out-of-bounds parts fill the image of described image out-of-bounds part.
11. device according to claim 10, is characterized in that, described acquisition module obtains the not view data of out-of-bounds part of described image in such a way from described external memory storage:
Directly from described external memory storage, read the not view data of out-of-bounds part of described image according to described out-of-bounds information.
12. device according to claim 10, is characterized in that, described acquisition module comprises:
Transmitting element, for sending the DMA request that carries described out-of-bounds information to outside dma module;
Receiving element, for receiving the not view data of out-of-bounds part of described image that described outside DMA reads from described external memory storage according to described out-of-bounds information.
13. according to claim 10 to the described device of any one in 12, it is characterized in that, described packing module comprises:
The first filler cells, for obtaining described image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part;
The first memory cell, be stored to internal storage for the described image that will carry out after the level filling.
14. according to claim 10 to the described device of any one in 12, it is characterized in that, described packing module comprises:
The second memory cell, for the described image by obtaining not the view data of out-of-bounds parts be stored to internal storage;
The second filler cells, carry out the vertical direction filling for the view data of utilizing described image boundary member in the vertical direction to the image of described image out-of-bounds part;
Described the second memory cell, also be stored to described internal storage for the described figure that will carry out after the vertical direction filling.
15. according to claim 10 to the described device of any one in 12, it is characterized in that, described packing module comprises:
Horizontal filler cells, for obtaining described image from external memory storage not during the view data of out-of-bounds parts, the view data of utilizing described image boundary member is in the horizontal direction carried out the horizontal direction filling to the image of described image out-of-bounds part;
The 3rd memory cell, be stored to internal storage for the described image that will carry out after the level filling;
Vertical filler cells, the view data of carrying out the described image boundary member in the vertical direction after the level filling for utilization is carried out the vertical direction filling to the image of described image out-of-bounds part;
Described the 3rd memory cell, also be stored to described internal storage for the described figure that will carry out after the vertical direction filling.
16. device according to claim 15, is characterized in that, described packing module also comprises:
The 4th memory cell, be stored to inner buffer for the described image boundary image data in the vertical direction that will carry out after the level filling;
Reading unit, for reading described described image boundary image data in the vertical direction of carrying out the level filling from described inner buffer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101501643A CN103428482A (en) | 2012-05-15 | 2012-05-15 | Processing method and device of image data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101501643A CN103428482A (en) | 2012-05-15 | 2012-05-15 | Processing method and device of image data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103428482A true CN103428482A (en) | 2013-12-04 |
Family
ID=49652580
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012101501643A Pending CN103428482A (en) | 2012-05-15 | 2012-05-15 | Processing method and device of image data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103428482A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022115999A1 (en) * | 2020-12-01 | 2022-06-09 | 深圳市大疆创新科技有限公司 | Data processing method and data processing device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1578468A (en) * | 2003-07-28 | 2005-02-09 | 松下电器产业株式会社 | Video decoding apparatus |
| CN101729904A (en) * | 2008-10-24 | 2010-06-09 | 安凯(广州)微电子技术有限公司 | Interframe prediction data reading method, multimedia processor and multimedia playing terminal |
| CN101729903A (en) * | 2008-10-24 | 2010-06-09 | 安凯(广州)微电子技术有限公司 | Method, system and multimedia processor for reading reference frame data |
| CN102300086A (en) * | 2010-06-23 | 2011-12-28 | 中国科学院微电子研究所 | Method for expanding reference frame boundary and limiting position of motion compensation reference sample |
-
2012
- 2012-05-15 CN CN2012101501643A patent/CN103428482A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1578468A (en) * | 2003-07-28 | 2005-02-09 | 松下电器产业株式会社 | Video decoding apparatus |
| CN101729904A (en) * | 2008-10-24 | 2010-06-09 | 安凯(广州)微电子技术有限公司 | Interframe prediction data reading method, multimedia processor and multimedia playing terminal |
| CN101729903A (en) * | 2008-10-24 | 2010-06-09 | 安凯(广州)微电子技术有限公司 | Method, system and multimedia processor for reading reference frame data |
| CN102300086A (en) * | 2010-06-23 | 2011-12-28 | 中国科学院微电子研究所 | Method for expanding reference frame boundary and limiting position of motion compensation reference sample |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022115999A1 (en) * | 2020-12-01 | 2022-06-09 | 深圳市大疆创新科技有限公司 | Data processing method and data processing device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3664443B1 (en) | Panoramic image generation method and device | |
| CN110662087B (en) | Point cloud coding and decoding method and coder-decoder | |
| US9105090B2 (en) | Wide-angle lens image correction | |
| CN103984513B (en) | Massive geographic information system (GIS) information ultrahigh resolution displaying method | |
| US10708619B2 (en) | Method and device for generating predicted pictures | |
| CN107967669A (en) | Method, apparatus, computer equipment and the storage medium of picture processing | |
| CN113132558A (en) | Architecture for high performance power efficient programmable image processing | |
| CN103369331B (en) | The complementing method of image cavity and device and the treating method and apparatus of video image | |
| EP2786348B1 (en) | Methods, systems and computer program products for creating three dimensional meshes from two dimensional images | |
| KR20200020704A (en) | Methods, systems, and apparatus for reducing memory latency when fetching pixel kernels | |
| US9201781B2 (en) | Data processing apparatus, data processing method and data sharing system | |
| CN103428482A (en) | Processing method and device of image data | |
| KR20170073937A (en) | Method and apparatus for transmitting image data, and method and apparatus for generating 3dimension image | |
| CN104268098A (en) | On-chip cache system for transformation on ultrahigh-definition video frame rates | |
| US20080049035A1 (en) | Apparatus and method for accessing image data | |
| US20210281882A1 (en) | Compact description of region-wise packing information | |
| CN101729903B (en) | Method, system and multimedia processor for reading reference frame data | |
| CN110800301A (en) | Control method and device of coding equipment and storage medium | |
| CN102300086A (en) | Method for expanding reference frame boundary and limiting position of motion compensation reference sample | |
| CN115205099A (en) | Image data transmission method and device and electronic equipment | |
| CN104469210B (en) | A kind of fringing method for SDI HD video character addings | |
| CN110662040B (en) | Depth map CU (CU) rapid partitioning method based on reconfigurable array processor | |
| CN111435992B (en) | Point cloud decoding method and device | |
| CN106034216A (en) | Camera image positioning system and camera image positioning method based on image identification | |
| CN110941399A (en) | Data processing method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131204 |
|
| RJ01 | Rejection of invention patent application after publication |