CN201436805U - Image processing device - Google Patents
Image processing device Download PDFInfo
- Publication number
- CN201436805U CN201436805U CN2009200064309U CN200920006430U CN201436805U CN 201436805 U CN201436805 U CN 201436805U CN 2009200064309 U CN2009200064309 U CN 2009200064309U CN 200920006430 U CN200920006430 U CN 200920006430U CN 201436805 U CN201436805 U CN 201436805U
- Authority
- CN
- China
- Prior art keywords
- image
- vector
- motion
- block
- image block
- 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.)
- Expired - Lifetime
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The utility model discloses an image processing device, which solves the problem of interpolation error in prior art. The image processing method includes: detecting a first motion vector in a first image frame to determine a flag value of the first image block. The flag value is used for indicating whether the image content of the first image block relatively has a small change. A target motion vector for the image block to be interpolated is calculated according to the flag value.
Description
Technical field
The utility model relates to a kind of image processing mechanism, refers to a kind of by being decided an image processor for the treatment of the employed target motion-vector of interpolation image area piece by the imaged object followed the trail of of photography camera lens in the detecting image image content especially.
Background technology
In general, produce in the interpolation image one and treat that the needed motion-vector of interpolation image area piece system decides motion-vector and produces the image block that this treats the interpolation of answering on the interpolation image area piece position with reference to the image block of two images on the contiguous time point.Yet, directly utilize general image interpolating method to come the image of interpolation any kind of will produce serious interpolation mistake sometimes, for example, in some scene picture, can find the object that the phtographic lens tracking is being moved, because the relation that phtographic lens is followed the trail of, be shown in object tracked in the scene picture and have slower move (that is corresponding to less motion-vector), and other background video can see that in image frame it has mobile (that is corresponding to bigger motion-vector) faster.As a rule, if be positioned near the central authorities of picture by the object followed the trail of of photography camera lens always, this moment is if needed motion-vector can't have problems when using above-mentioned conventional method to decide the image interpolation.Yet this object may be covered by a background suddenly, so promptly is difficult to detect the true moving direction (true motion) of this imaged object; For instance, please refer to Fig. 1, Fig. 1 is to use conventional practice to come drawing frame data is carried out the schematic diagram of image interpolation.Be that phtographic lens is followed the trail of the people that a certain name is being run in Fig. 1, this personage of expression of oblique line circle system, and empty circles and solid circles system represent greenery patches and trees in the background video respectively, this personage of explanation of the example of the left side of Fig. 1 system is covered by trees suddenly fully in the time point of picture frame f3, then is that this personage of expression is covered by trees at some position suddenly in the time point of picture frame f3 ' at the example of the right-hand part of Fig. 1.For convenience of description, only show phtographic lens among Fig. 1 and flatly move and follow the trail of this personage, so be the relative position relation of only expressing this personage, greenery patches and trees simply in Fig. 1.Shown in the left side of Fig. 1, when picture frame f2, still can see this personage's existence, and this personage is covered by trees fully when picture frame f3, right-hand part for Fig. 1, when picture frame f2 ', still can see this personage's existence, and some position of this personage is covered by trees when picture frame f3 '.No matter be under above-mentioned which kind of situation, use conventional practice to decide and treat the needed motion-vector of interpolation image area piece, can be because at next picture frame (that is f3, f3 ') can not find in corresponding to this personage's image block selects the image block interpolation that other is more alike to treating interpolation image area piece position P1, P1 ', but other more alike image block is not the image block that i.e. representative corresponds to this personage, in other words, conventional practice may be with the interpolation of incoherent image block to the former image block position P1 that should interpolation goes out this personage's image, P1 ' and cause the interpolation mistake.
In addition, if a certain imaged object is sheltered from by background video in last picture frame and just appears in the picture at next picture frame, use conventional practice to decide generation one to treat that the needed motion-vector of interpolation image area piece also as above-mentioned said, the situation of interpolation mistake takes place very likely.For example, please refer to Fig. 2, Fig. 2 is to use conventional practice another drawing frame data to be carried out the schematic diagram of image interpolation.Shown in the left side of Fig. 2, certain part of this personage when picture frame f4 (oblique line circle) is still hidden by trees, and this personage has not been covered by trees (solid circles) and can see whole images of this personage when picture frame f5, the right-hand part of Fig. 2 then be explanation when use conventional practice decide interpolation image finter treat the motion-vector of interpolation image area piece position P2 the time, because the image on the position P2 ' of a last picture frame f4 ' and next are opened the position P2 of picture frame f5 ' " on image difference very big; conventional practice can select with other more alike but may the interpolation of incoherent image block to the P2 of position; but since other more alike image block is not i.e. representative correspond to this personage the position that covered, so conventional practice easily causes image interpolation mistake; For example, in the example of the right-hand part of Fig. 2, promptly be with position P2 " ' image (representing the image in greenery patches) directly interpolation go up and cause manifest error to position P2.
The utility model content
Technical problem to be solved in the utility model provides a kind of by whether the image processor that is decided the value of the target motion-vector for the treatment of interpolation image area piece by the imaged object followed the trail of of photography camera lens is arranged in the detecting image picture, to solve the above problems.
In order to solve above technical problem, the utility model provides following technical scheme:
The utility model provides a kind of image processor.Image processor includes a treatment circuit and a counting circuit at least.Treatment circuit system is used for detecting the motion-vector of the first image block in first image frame to determine the flag value of the first image block, whether this flag value system relatively has less variation in order to the presentation content that indicates the first image block, and counting circuit is the target motion-vector that is coupled to treatment circuit and is used to treat interpolation image area piece according to flag value calculating.
Whether the utility model adopts a kind of by there being the imaged object of being followed the trail of by the photography camera lens to decide the image processor of the value of the target motion-vector for the treatment of interpolation image area piece in the detecting image picture, has overcome the interpolation mistake that prior art exists.
Description of drawings
Fig. 1 comes drawing frame data is carried out the schematic diagram of image interpolation for using conventional practice.
Fig. 2 carries out the schematic diagram of image interpolation for using conventional practice to another drawing frame data.
Fig. 3 is the schematic diagram of the image processor of the utility model one embodiment.
Fig. 4 carries out the image interpolation for image processor shown in Figure 3 at drawing frame data example schematic.
Fig. 5-1 is the operational flowchart of image processor shown in Figure 3.
Fig. 5-2 is the continuity flow chart after Fig. 5-1.
[primary clustering symbol description]
100 image processors
105 statistical circuits
110 treatment circuits
115 counting circuits
Embodiment
Please refer to Fig. 3, Fig. 3 is the schematic diagram of the image processor 100 of the utility model one embodiment.Image processor 100 includes statistical circuit 105, treatment circuit 110 and counting circuit 115.Flag value for a certain image block of decision, treatment circuit 110 decides the flag value Vflag of the first image block by the motion-vector MV ' that detects one first image block in one first image frame at least, wherein whether flag value Vflag system relatively has less variation in order to the presentation content that indicates the first image block, for example in order to indicate the first image block whether to correspond to the imaged object that phtographic lens is followed the trail of; It is noted that herein, the purpose of image processor 100 be to determine interpolation one treat in the interpolation image treat interpolation image area piece MB the time needed target motion-vector MV, first image frame means arbitrary with reference to image frame before this treats the interpolation image, and the usefulness of the content of flag value Vflag system reference when being used for for counting circuit 115 decision target motion-vector MV.
On real the work, the number of the middle motion-vector of first image frame (picture frame or figure) greater than a plurality of image blocks of predetermined vector size Vpre added up in statistical circuit 105 (it is coupled to treatment circuit 110) meeting earlier, and the purpose of adding up be after treatment circuit 110 can inspect these a plurality of image blocks whether number more than or equal to a special value N1 to judge the situation of the image fast moving of whether having powerful connections in this first image frame, that is it is whether the motion-vector number of inspecting greater than predetermined vector size Vpre accounts on a certain ratio of the total number of motion-vector in first image frame; Numerical value of N 1 can preestablish or dynamically adjust.So the reason of doing is when the situation of phtographic lens tracking imaged object occurring, and remaining background video part generally can be occupied ratio bigger in first image frame.Therefore, if greater than the motion-vector number of predetermined vector size Vpre more than or equal to special value N1, then treatment circuit 110 can determine in first image frame to be the operation that situation that the background video fast moving occurs is also then set flag value earlier; Otherwise, then represent not take place in first image frame situation of background video fast moving.No matter whether the arbitrary image block in first image frame has the motion-vector greater than predetermined vector size Vpre, and treatment circuit 110 can be set the flag value of all image blocks in first image frame in advance and indicate the imaged object that its image block is not separately followed the trail of corresponding to phtographic lens.Certainly, in order to determine whether to have the situation of background video fast moving more, whether those motion-vectors that treatment circuit 110 can further be inspected greater than predetermined vector size Vpre have consistent directivity (for example whether those motion-vectors difference to each other all falls within a preset range); If consistent directivity is arranged, the situation of the image fast moving of then can determining further to have powerful connections in first image frame.The operation of treatment circuit 110 decision flag value is then described in detail in next paragraph
As mentioned above, when the situation of background video fast moving in determining first image frame, occurring (that is greater than the motion-vector number of predetermined vector size Vpre more than or equal to special value N1 and have consistent directivity), treatment circuit 110 is chosen the image block of motion-vector less than predetermined vector size Vpre, the existence of those image blocks is arranged, represent the imaged object that very likely has phtographic lens to follow the trail of in first image frame, therefore treatment circuit 110 then is set at the pairing flag value of those image blocks ' 1 ' and corresponds to the imaged object that phtographic lens is followed the trail of to indicate those image blocks systems, and the motion-vector that also writes down those image blocks simultaneously uses for subsequent calculations circuit 115.For instance, for the first image block, treatment circuit 110 relatively motion-vector MV ' and the predetermined vector size Vpre of the first image block with decision flag value Vflag, if motion-vector MV ' is less than predetermined vector size Vpre, then the flag value Vflag of the first image block can be set to ' 1 '.
More than the mode of decision flag value is the mode of intuition comparatively, yet, because the content of flag value can have influence on the operation of subsequent calculations circuit 115, thus in fact can be in addition when the decision flag value with reference to other rule accurately to determine the content of flag value.For instance, as the motion-vector MV ' of the first image block during less than predetermined vector size Vpre, treatment circuit 110 can judge in addition whether the motion-vector of at least one contiguous image block of the first image block also decides the flag value Vflag of the first image block less than predetermined vector size Vpre.This is because in general an object of being followed the trail of by the photography camera lens can only not have single image area block size, so, if the motion-vector MV ' of the first image block is less than predetermined vector size Vpre, then the motion-vector of its at least one contiguous image block also should just can reasonably indicate first image block system less than predetermined vector size Vpre and correspond to the imaged object that phtographic lens is followed the trail of.Therefore, if arbitrary motion-vector is not arranged abnormally less than predetermined vector size Vpre in the motion-vector of the motion-vector MV ' of the first image block its contiguous image block less than predetermined vector size Vpre, then might be cause the first image block because of the error of calculation motion-vector MV ' less than predetermined vector size Vpre, treatment circuit 110 can be set at the flag value Vflag of the first image block ' 0 ' and expressed the first image block and do not correspond to the imaged object that phtographic lens is followed the trail of this moment.
In addition, from the above, for confirming flag value Vflag more, treatment circuit 110 also can calculate the first image block and motion-vector in addition less than the block comparison difference D1 between another image block MB1 (being that MB1 corresponds to the imaged object that phtographic lens is followed the trail of equally) of predetermined vector size Vpre, and calculate the first image block and motion-vector, and judge whether block comparison difference D1 compares the flag value Vflag that difference D2 decides the first image block much smaller than block greater than the block comparison difference D2 between the image block MB2 (being that MB2 corresponds to background video) of predetermined vector size Vpre.For example, example with above-mentioned people and trees, the block comparison difference (that is D1) that corresponds to equally between people's the two image blocks certainly will be compared difference (that is D2) much smaller than the block between the two image blocks that correspond to people and trees respectively, therefore, if block comparison difference D1 is much smaller than block comparison difference D2, then treatment circuit 110 can further be determined the flag value Vflag of the first image block is set at ' 1 '; Otherwise treatment circuit 110 can be set at ' 0 ' with its flag value Vflag.
As mentioned above, after the content that determines flag value Vflag, the counting circuit 115 that is coupled to treatment circuit 110 can calculate the target motion-vector MV that is used to treat interpolation image area piece MB according to the content of flag value Vflag.In the present embodiment, treat that the position system that interpolation image area piece MB is arranged in an interpolation image is same as the first image block in the relative position of first image frame; Yet, this non-qualification of the present utility model.In particular, counting circuit 115 be the motion-vector MV ' of the first image block that whether treatment circuit 110 write down according to the decision of the content of flag value Vflag as candidate's motion-vector for the treatment of interpolation image area piece MB, then from all candidate's motion-vectors for the treatment of interpolation image area piece MB, select target motion-vector MV again.Be noted that first image block system corresponds to the imaged object that phtographic lens is followed the trail of, its motion-vector ideal value is zero, therefore after in the step of decision target motion-vector MV, motion-vector zero also can be included into consideration.Please refer to Fig. 4, Fig. 4 is image processor 100 shown in Figure 3 carries out the image interpolation at drawing frame data a example schematic.Object (people of for example running) of being followed the trail of by the photography camera lens of oblique line circle system expression shown in Figure 4, empty circles, solid circles, intersection mark are then represented different background video (for example greenery patches, trees or the like) respectively with the triangle mark.Suppose that the first above-mentioned image frame means f ', the position for the treatment of the interpolation image area piece MB and the first image block lays respectively on P4 and the P4 ', and treatment circuit 110 can be set at the flag value Vflag of the first image block ' 1 ' and the motion-vector MV ' (theoretical value is zero) of the first image block noted; For example, if motion-vector MV ' is zero, then representative can produce a certain needed image of interpolation image area piece for the treatment of with reference to the image block on the same spatial location in the adjacent image frame.
Because the content of the flag value Vflag of the first image block means that the first image block is shown to correspond to by the imaged object of photography camera tracking, so candidate's motion-vector that counting circuit 115 may use in the time of the motion-vector MV ' of the first image block can being treated interpolation image area piece MB as generation.Specifically, when counting circuit 115 is treated the target motion-vector MV of interpolation image area piece MB in decision, can calculate its each self-corresponding candidate block comparison difference D1 ' ~ Dn ' (supposing to have n candidate's motion-vector) respectively according to each candidate's motion-vector.Yet, counting circuit 115 can suitably be adjusted candidate block comparison difference D1 ' ~ Dn ', make that (for example the value as if motion-vector MV ' is zero to the candidate block comparison difference D1 ' that calculates according to motion-vector MV ', then candidate block comparison difference D1 ' promptly is the position P4 of picture frame f6 " on image and the position P4 of picture frame f7 " ' on image between difference) than other candidate block compare difference D2 ' ~ Dn ' little, for example, counting circuit 115 can reduce candidate block comparison difference D1 ' (for example with D1 ' divided by 2) or improve other candidate block comparison difference D2 ' ~ Dn '.So the purpose of doing is to wish can have bigger probability to choose D1 ' when counting circuit 115 is selected a minimum block comparison difference from candidate block comparison difference the D1 ' ~ Dn ' that treats interpolation image area piece MB, so that counting circuit 115 can use motion-vector (that is MV ') corresponding to D1 ' to be used as treating the target motion-vector MV of interpolation image area piece MB.
With the quality of image, when taking place in next picture frame f7, to be occurred suddenly by the imaged object that the photography camera lens is followed the trail of, in order to make the image that interpolation image area piece MB is presented for the treatment of on the P4 of position not produce artificial false shadow, use the position P4 of picture frame f6 " on image with or the position P4 of picture frame f7 " ' on image to produce the image for the treatment of interpolation image area piece MB be the practice preferably, so the time preference counting circuit 115 should select the target motion-vector MV that motion-vector MV ' (its value is zero) is used as treating interpolation image area piece MB.On the contrary, do not correspond to arbitrary if the content of the flag value Vflag of the first image block indicates the first image block by the imaged object of photography camera tracking (that is flag value Vflag system is set at ' 0 '), therefore candidate's motion-vector when then counting circuit 115 can be not painstakingly be treated interpolation image area piece MB with the motion-vector MV ' of the first image block as generation can not produce any interpolation mistake; For instance, suppose among the first image frame f ' of Fig. 4 that the system of the first image block on the P4 ' of position corresponds to trees (being solid circles) but not corresponds to personage's (being the oblique line circle), then the flag value of the first image block can be ' 0 ', therefore counting circuit 115 promptly carries out known block comparison and searches and produce the motion-vector for the treatment of interpolation image area piece MB among the interpolation image finter on the P4 of position, thus can be not painstakingly in addition the motion-vector that other is relevant with flag value elect candidate's motion-vector as.
In addition, treatment circuit 110 may need the flag value Vflag content that is set at ' 1 ' is fallen to make it to become empty (null) clearly.With Fig. 4 is example, if be arranged in the situation that arbitrary image frame between the first image frame f ' and the image frame f6 detects occurrence scene conversion (scene change) (that is detect the first image frame f ' and image frame f6/f7 system adheres to different scene pictures separately), then painstakingly the motion-vector MV ' of the first image block is set at when the target motion-vector MV of interpolation image area piece MB is treated in decision for fear of counting circuit 115 and has higher possibility, the flag value Vflag that treatment circuit 110 can will originally be set to ' 1 ' the first image block does not fall clearly so that counting circuit 115 can not be made as motion-vector MV ' to have higher possibility; At this moment, flag value Vflag means that the presentation content that the first image block is shown may not relatively have less variation.Therefore, even scene change has taken place, counting circuit 115 will can not produce mistake because of motion-vector MV ' has higher possibility when the target motion-vector MV of interpolation image area piece MB is treated in decision.
In addition, phtographic lens takes place in determining the first image frame f ' treatment circuit 110 follows the trail of after the situation of imaged object, and it also can check constantly whether the situation that phtographic lens is followed the trail of imaged object also takes place in the follow-up image frame; If check out the situation that does not have phtographic lens to follow the trail of imaged object in the follow-up image frame, then treatment circuit 110 can fall to have among the first image frame f ' content clearly and is ' 1 ' flag value (for example Vflag among Fig. 4), for example, detect/count motion-vector at that time and become when being less than special value N1 when being arranged in a certain image frame between the first image frame f ' and the image frame f6 greater than the image block number of predetermined vector size Vpre, represent to have in this image frame the situation of the specific image of camera tracking, in other words, the presentation content of the first image block may not relatively have less variation.At this moment, treatment circuit 110 content that can fall the flag value Vflag of the first image block does not clearly have higher possibility so that counting circuit 115 can be not painstakingly when decision target motion-vector MV be set at the motion-vector MV ' of the first image block.
Moreover, in the present embodiment, for ' 1 ' afterwards, treatment circuit 110 also checks constantly in the follow-up image frame whether the contiguous image block with the image block of the first image block same spatial location still has the flag value of at least one contiguous image block also to be set at ' 1 ' with the flag value Vflag decision of the first image block of the first image frame f '; If there is not the flag value of arbitrary contiguous image block to be set to ' 1 ', then the treatment circuit 110 flag value Vflag of the first image block that can fall the first image frame f ' clearly painstakingly is set at the motion-vector MV ' of the first image block when avoiding determining treating the target motion-vector MV of interpolation image area piece MB and has higher possibility.
Be noted that, though the mode of operation of image processor 100 when the imaged object of being followed the trail of by the photography camera lens occurs suddenly only is described in the example of Fig. 4, image processor 100 also can be used for determining preferable target motion-vector to avoid artificial false shadow when the imaged object of being followed the trail of by the photography camera lens suddenly disappears; In order to simplify this explanation book, do not give unnecessary details in addition at this.At last, can understand spirit of the present utility model rapidly for making the reader, enclose the operational flowchart (that is Fig. 5-1 and Fig. 5-2) of image processor shown in Figure 3 100 at this, too tediously long for fear of length, its step explanation repeats no more.
The above only is preferred embodiment of the present utility model, and all equalizations of being done according to the utility model claim change and modify, and all should belong to covering scope of the present utility model.
Claims (9)
1. image processor is characterized in that it includes:
One treatment circuit, a motion-vector that is used for detecting one first image block in one first image frame decides a flag value of this first image block, and whether this flag value relatively has less variation in order to the presentation content of indicating this first image block; And
One counting circuit is coupled to this treatment circuit, is used for calculating a target motion-vector that is used for treating interpolation image area piece according to this flag value.
2. image processor as claimed in claim 1 is characterized in that, the wherein said interpolation image area piece for the treatment of is arranged in an interpolation image and this identical locus of first image block.
3. image processor as claimed in claim 1, it is characterized in that, described counting circuit has the judgement structure, optionally this motion-vector of this first image block is treated candidate's motion-vector of interpolation image area piece as this according to this flag value, and select this target motion-vector from these a plurality of candidate's motion-vectors for the treatment of interpolation image area piece.
4. image processor as claimed in claim 3, it is characterized in that, wherein when described flag value indicates the presentation content of this first image block relatively to have less variation, this counting circuit can be with this motion-vector of this first image block as this candidate's motion-vector.
5. image processor as claimed in claim 4, it is characterized in that, wherein when described flag value indicates the presentation content of this first image block relatively to have less variation, this counting circuit is used for this candidate block for the treatment of interpolation image area piece comparison difference for this motion-vector that uses this first image block calculates, and reduces this candidate block comparison difference or improve all the other candidate block comparison differences except this candidate block comparison difference in these a plurality of candidate block comparison differences for the treatment of interpolation image area piece; And this counting circuit can treat to select a minimum block comparison difference a plurality of candidate block comparison of this of interpolation image area piece differences from this, and uses the motion-vector corresponding to this minimum block comparison difference to be used as this target motion-vector.
6. image processor as claimed in claim 1 is characterized in that it also includes:
One statistics circuit is coupled to this treatment circuit, is used for adding up in this first image frame motion-vector greater than the number of a plurality of image blocks of a predetermined vector size;
Wherein this treatment circuit judges that in addition whether the number of these a plurality of image blocks is greater than a special value; When the number of these a plurality of image blocks all fell within a preset range greater than this special value and these a plurality of motion-vectors difference to each other, this treatment circuit was by relatively this motion-vector of this first image block and this flag value that this predetermined vector size decides this first image block.
7. image processor as claimed in claim 6, it is characterized in that, wherein when this motion-vector of the described first image block during less than this predetermined vector size, this treatment circuit judges that in addition whether a motion-vector of at least one contiguous image block of this first image block decides this flag value of this first image block less than this predetermined vector size.
8. image processor as claimed in claim 6, it is characterized in that, wherein when this motion-vector of the described first image block during less than this predetermined vector size, this treatment circuit calculates this first image block in addition and has less than the comparison of one first block between the image block of a motion-vector of this predetermined vector size difference, and calculate this first image block and have, and judge that whether this first block comparison difference decides this flag value of this first image block less than this second block comparison difference greater than the comparison of one second block between the image block of a motion-vector of this predetermined vector size difference.
9. image processor as claimed in claim 1 is characterized in that it also includes:
One statistics circuit is coupled to this treatment circuit, is used for adding up after this first image frame in one second image frame motion-vector greater than the number of a plurality of image blocks of a predetermined vector size;
Wherein this treatment circuit judges that in addition whether the number of these a plurality of image blocks is less than a special value; When the number of these a plurality of image blocks during less than this special value, this treatment circuit can be set this flag value of this first image block, and on behalf of the presentation content of this first image block, this flag value may not relatively have less variation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009200064309U CN201436805U (en) | 2009-03-03 | 2009-03-03 | Image processing device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009200064309U CN201436805U (en) | 2009-03-03 | 2009-03-03 | Image processing device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN201436805U true CN201436805U (en) | 2010-04-07 |
Family
ID=42373250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009200064309U Expired - Lifetime CN201436805U (en) | 2009-03-03 | 2009-03-03 | Image processing device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN201436805U (en) |
-
2009
- 2009-03-03 CN CN2009200064309U patent/CN201436805U/en not_active Expired - Lifetime
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102673979B (en) | Method and device for judging deviation of conveying belt | |
| CN111709407B (en) | Method and device for improving video target detection performance in monitoring edge calculation | |
| CN102136147B (en) | Target detecting and tracking method, system and video monitoring device | |
| CN103688292A (en) | Image display apparatus and image display method | |
| CN117351008A (en) | Smart phone panel surface defect detection method | |
| JP2016526713A (en) | Tracker auxiliary image capture | |
| CN109711332B (en) | Regression algorithm-based face tracking method and application | |
| CN106659370A (en) | Capsule endoscope, endoscope system, and capsule endoscope operation method | |
| US11763470B2 (en) | Method and apparatus for aligning image frames, electronic device and storage medium | |
| CN110021055A (en) | The method and its image processing apparatus and system of generation disparity map | |
| US20140185866A1 (en) | Optical navigation method and device using same | |
| CN111325798A (en) | Camera model correction method and device, AR implementation equipment and readable storage medium | |
| JP2013101419A (en) | Image signal processor | |
| CN101631241B (en) | Image processing method and device | |
| CN201436805U (en) | Image processing device | |
| CN113838097A (en) | Camera lens angle deviation detection method and device and storage medium | |
| CN112884803A (en) | Real-time intelligent monitoring target detection method and device based on DSP | |
| CN112907433A (en) | Digital watermark embedding method, digital watermark extracting device, digital watermark embedding apparatus, digital watermark extracting apparatus, and digital watermark extracting medium | |
| US20080239097A1 (en) | Image Processing Apparatus and Image Processing Method | |
| CN109685802B (en) | Low-delay video segmentation real-time preview method | |
| CN111640134A (en) | Face tracking method and device, computer equipment and storage device thereof | |
| US20200034960A1 (en) | Correlation value calculation device | |
| CN105184809A (en) | Moving object detection method and moving object detection device | |
| CN104978731A (en) | Information processing method and electronic equipment | |
| CN114170271B (en) | Multi-target tracking method, equipment and storage medium with self-tracking consciousness |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CX01 | Expiry of patent term | ||
| CX01 | Expiry of patent term |
Granted publication date: 20100407 |