CN105868683A - Channel logo identification method and apparatus - Google Patents

Abstract

The present invention discloses a channel logo identification method and apparatus, and relates to the technical field of information identification. The method comprises: determining whether a to-be-identified channel logo in a channel logo region is a CCTV (China Central Television) channel logo; when the to-be-identified channel logo is a CCTV channel logo, identifying the to-be-identified channel logo according to a first preset policy; and when the to-be-identified channel logo is a non-CCTV channel logo, identifying the to-be-identified channel logo according to a second preset policy. According to the channel identification method and apparatus provided by the present invention, channel logos are not identified in the same manner any longer, but firstly, it is determined whether the to-be-identified channel logo in the channel logo region is the CCTV channel logo, and then different policies are adopted to identify the CCTVE channel logo and the non-CCTV channel logo, so that the to-be-identified channel logo can be effectively identified targeted at the features of the CCTV channel logo and the non-CCTV channel logo, which improves identification efficiency and identification accuracy.

Description

TV station symbol recognition method and device

Technical field

The present invention relates to information discriminating technology field, particularly to a kind of TV station symbol recognition method and dress Put.

Background technology

Intelligent television just complies with " high Qinghua ", " networking ", the trend of " intelligent " develops rapidly, Possess from the Internet, the multiple channel such as video equipment, computer obtains programme content, by letter The function that the content that consumer is needed most by the mode that breath merges clearly represents on giant-screen.With Traditional tv is compared, intelligent television provided the user more quick, intelligence, hommization should With service.

Intelligent television comprises a large amount of order video, programme televised live, and most of TV programme are protected Stay the station symbol of program product side.Station symbol is to discriminate between the important symbol of television station, contains TV The important semantic informations such as platform platform name, programming source, program orientation, be realize video analysis, Understand and the important semantic source of retrieval.The realization of TV station symbol recognition technology, will effectively realize programme Function, and to understanding that user preferences, input value-added service tool are of great significance.

In prior art when station symbol is identified, generally employing following two scheme:

The first scheme: TV station symbol recognition scheme based on single-frame images.This kind of method with edge away from From transformation matrix as feature, all use the mode of template matching, (sliding including overall situation edge matching Dynamic window travels through whole station symbol region), divided-fit surface (using manual type to filter non-edge), Color form fit (CF feature is mated simultaneously, carries out classification and ordination for program searching) Deng.

First scheme: TV station symbol recognition method based on multiple image.This kind of scheme generally can be adopted By following three kinds of methods: one, utilize sensitizing range (the i.e. station symbol of continuous multiple frames sequence of frames of video Region) the eigenvalue change of pixel splits image, and carries out rim detection, uses and slide Window and method of partition carry out match cognization.Two, comprehensive utilization CF feature, according to Spatio-temporal invariant splits station symbol in frame sequence, utilizes spatial distribution rectangular histogram to combine HSV color Feature is effectively described by Color Histogram, finally utilizes SUV to complete TV station symbol recognition.Method Three: calculate the change of consecutive frame image, extract station symbol and Hu not bending moment thereof, and according to candidate Set and pre-set criteria are identified result etc..

But in prior art, CCTV's station symbol, satellite TV's station symbol and local station symbol etc. are used phase Tongfang Formula is identified, and causes recognition efficiency and recognition accuracy too low.

Summary of the invention

The embodiment of the present invention provides a kind of TV station symbol recognition method and device, in order to solve prior art Middle recognition efficiency and the too low defect of recognition accuracy.

The embodiment of the present invention provides a kind of TV station symbol recognition method, and described method includes:

Judge whether the station symbol to be identified in station symbol region is CCTV's station symbol；

When described station symbol to be identified is CCTV's station symbol, according to the first preset strategy, described waiting is known Other station symbol is identified；

When described station symbol to be identified is non-CCTV's station symbol, treat described according to the second preset strategy Identification station symbol is identified.

The embodiment of the present invention provides a kind of TV station symbol recognition device, and described device includes:

Station symbol judging unit, for judging whether the station symbol to be identified in station symbol region is CCTV's platform Mark；

First recognition unit, for when described station symbol to be identified is CCTV's station symbol, according to first Described station symbol to be identified is identified by preset strategy；

Second recognition unit, for when described station symbol to be identified is non-CCTV's station symbol, according to the Described station symbol to be identified is identified by two preset strategy.

The TV station symbol recognition method and device that the embodiment of the present invention provides, no longer all uses each station symbol Same way is identified, but first judges whether the station symbol to be identified in station symbol region is CCTV Station symbol, then use Different Strategies to be identified CCTV's station symbol and non-CCTV station symbol, it is possible to effectively Station symbol to be identified is identified by ground for the feature of CCTV's station symbol and non-CCTV station symbol, improves Recognition efficiency and recognition accuracy.

Accompanying drawing explanation

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below By the accompanying drawing used required in embodiment or description of the prior art is briefly described, aobvious And easy insight, the accompanying drawing in describing below is some embodiments of the present invention, general for this area From the point of view of logical technical staff, on the premise of not paying creative work, it is also possible to attached according to these Figure obtains other accompanying drawing.

Fig. 1 is the flow chart of the TV station symbol recognition method of one embodiment of the present invention；

Fig. 2 is the signal of the video frame images in one embodiment of the present invention before region segmentation Figure；

Fig. 3 is the signal of the video frame images in one embodiment of the present invention after region segmentation Figure；

Fig. 4 is as a example by CCTV5, and the edge of each video frame images carries out the overall effect synthesized Fruit figure；

Fig. 5 is the flow chart of the TV station symbol recognition method of one embodiment of the present invention；

Fig. 6 is the structured flowchart of the TV station symbol recognition device of one embodiment of the present invention.

Detailed description of the invention

For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below will knot Close the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, Be fully described by, it is clear that described embodiment be a part of embodiment of the present invention rather than Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not having Make the every other embodiment obtained under creative work premise, broadly fall into present invention protection Scope.

It it should be understood that, although mainly for the platform of TV programme in intelligent television in background technology Mark is other, but embodiment of the present disclosure is not limited to this, and it is (all that it could be applicable to other objects Such as, the equipment such as panel computer, mobile phone, PC) in the TV station symbol recognition of TV programme, the most just Being to say, the scene of every TV station symbol recognition relating to TV programme is included in present embodiment In range of application.

Fig. 1 is the flow chart of the TV station symbol recognition method of one embodiment of the present invention；With reference to Fig. 1, Described method includes:

S101: judge whether the station symbol to be identified in station symbol region is CCTV's station symbol；

It should be noted that described station symbol region is only includes to be identified target area.

It will be appreciated that described station symbol region can be extracted in several ways, in order to prevent The impact on TV station symbol recognition of the noise such as random noise, picture noise, in present embodiment, passes through Following steps obtain and include the station symbol region of station symbol to be identified:

(1) in the predeterminable area of the video including station symbol to be identified, video frame images sequence is obtained Row；

According to priori, the substantially all upper left corner being positioned at video frame images of television station's station symbol (certainly, if being in other positions, it is possible to carry out accommodation as required), because of During the detection of this station symbol, only need to extract fixing upper left corner area (i.e. predeterminable area) as station symbol Detection region.Existing TV station symbol recognition method is generally according to optimal region rule (GSR) Obtaining station symbol region, present embodiment is with existing TV station symbol recognition method difference: (1) Calculate the proportional positions that all station symbols effectively identify in each video frame images；(2) calculating is all The maximum magnitude of proportional positions is as the region of station symbol region segmentation.Video with 1920*1080 As a example by, station symbol cut zone is row starting position 80 (1/24), arranges starting position 40 (1/27), line width 450 (15/64), col width 180 (1/6), region segmentation effect such as Fig. 2 Shown in Fig. 3, certainly, described proportional positions can the most suitably adjust, this enforcement This is not any limitation as by mode.

For information unrelated in elimination image, recover or strengthen useful relevant information, improve spy The detectability levied, simplifies data to greatest extent, to guarantee the reliability identified, this enforcement In mode, each video frame images can carry out pretreatment, described pretreatment includes: region segmentation, At least one in gray processing and image enhaucament, certainly, may also include other processing procedures, this This is not any limitation as by embodiment.

Described pretreatment can use formula Gray=0.33R+0.59G+0.11B to carry out gray processing, Certainly, it is possible to substituted by modes such as triple channel mean value method or triple channel maximum value process, wherein, Gray is the gray value of pixel, and R is the red component of pixel, and G is the green component of pixel, B is the blue component of pixel.

The purpose of described image enhaucament is prominent station symbol region effective information, as icon, word, Numerals etc., image enhaucament uses the gray scale stretching of 0～255 gray levels, it is also possible to rectangular histogram converts Method substitutes.

(2) each video frame images is carried out edge extracting；

It will be appreciated that the edge violent part that is variation of image grayscale, edge extracting is station identification Other key, the integrated degree at edge directly affects TV station symbol recognition result, certainly, edge extracting Method have a lot, such as Canny, LOG, Sobel, Laplace operator method etc..Comprehensively Consider the requirements such as denoising, edge integrity, edge precision, present embodiment uses Canny edge detection method.

In implementing, the parameter of Canny edge detection method is set to: weak edge threshold 50, strong edge threshold 200, certainly, it is possible to the most suitably float, such as, threshold Value is floated in the range of ± 10.

(3) edge of each video frame images is synthesized；

In implementing, presetting of correspondence can be determined according to the quantity of described video frame images Whether image threshold, be less than described pre-further according to described each marginal point in the quantity of video frame images If image threshold judges whether to retain this marginal point.

It is to say, pre-build between the quantity of video frame images and pre-set image threshold value is right Should be related to, search corresponding relation according to the quantity of described video frame images, to determine the pre-of correspondence If image threshold, there is each marginal point quantity at video frame images less than described pre-set image During threshold value, do not retain this marginal point, be optionally greater than in the quantity of video frame images at each marginal point During described pre-set image threshold value, retain this marginal point.

Illustrate to close the edge of each video frame images with a specific embodiment below Become, but do not limit protection scope of the present invention: setting the N quantity as video frame images, X is Pre-set image threshold value.

As N=6, correspondingly, X=4, say, that only marginal point (bag more than 4 Include 4) video frame images in the presence of just retain, if marginal point (includes 3) below 3 Video frame images in the presence of then give up；

When 6 > N > 3 time, correspondingly, X=3, say, that only marginal point is more than 3 Just retain in the presence of in the video frame images of (including 3), if marginal point (bag below 2 Include 2) video frame images in the presence of then give up；

When N≤3, correspondingly, X=N, say, that only marginal point is at all videos Just retaining in the presence of in two field picture, other situations are all given up.

Certainly, the parameter in described corresponding relation can be adjusted according to the resolution of image, this This is not any limitation as by embodiment.

As a example by CCTV5, Fig. 4 illustrates the whole structure of synthesis.

All recognition accuracy can be caused shadow due to edge noise, black surround and inessential word etc. Ringing, for improving recognition accuracy further, the edge that can synthesize is optimized process, this enforcement In mode, described optimization processes and includes: edge noise is deleted, black surround is removed and inessential word At least one in deletion.

(4) the minimum external matrix at the edge of synthesis is obtained；

(5) respectively each video frame images is split according to described minimum external matrix, and The image being partitioned into is synthesized by average weighted mode, includes to be identified to obtain Target station symbol region.

S102: when described station symbol to be identified is CCTV's station symbol, according to the first preset strategy to institute State station symbol to be identified to be identified；

In implementing, owing to the difference of CCTV's station symbol is only that numeral is different with word, therefore And, described station symbol to be identified can be entered according to the first preset strategy for the feature of CCTV's station symbol Row identifies.

S103: when described station symbol to be identified is non-CCTV's station symbol, according to the second preset strategy pair Described station symbol to be identified is identified.

It will be appreciated that can be for the feature of non-CCTV station symbol, according to the second preset strategy to institute State station symbol to be identified to be identified.

Each station symbol is no longer all used same way to be identified by present embodiment, but first judges Whether the station symbol to be identified in station symbol region is CCTV's station symbol, then to CCTV's station symbol and non-CCTV platform Mark uses Different Strategies to be identified, it is possible to effectively for CCTV's station symbol and non-CCTV station symbol Station symbol to be identified is identified by feature, improves recognition efficiency and recognition accuracy.

Fig. 5 is the flow chart of the TV station symbol recognition method of one embodiment of the present invention；With reference to Fig. 2, Described method includes:

S501: judge whether the station symbol to be identified in station symbol region is CCTV's station symbol；

Will be understood that, it is judged that when whether the station symbol to be identified in station symbol region is CCTV's station symbol, Various ways can be used, for ensureing the accuracy rate judged, in present embodiment, according to station symbol district The length-width ratio in territory, gray scale and color judge whether the station symbol to be identified in station symbol region is CCTV's platform Mark.

Research finds, CCTV's station symbol finds relative to satellite TV's station symbol, local broadcasting stations' target feature and difference: (1) length-width ratio difference, the length-width ratio of CCTV's station symbol is (long: vertical direction length；Wide: water Square to width) generally it is significantly less than other station symbols；(2) cromogram of CCTV's station symbol has extensively White pixel feature, especially show left side 2/3 at；(3) gray-scale map of CCTV's station symbol divides After block, meet the related constraint that gray scale is close between sub-block, such as average, variance etc..

So, can be come institute by the difference of length-width ratio, gray scale and the color in described station symbol region State station symbol region to classify.

Owing to length-width ratio is one of the most direct feature of station symbol.So, can be first to described station symbol district Territory is tentatively judged by length-width ratio, say, that first calculate the length-width ratio in each station symbol region, The method calculating length-width ratio is: calculate length H and width W, the length-width ratio in station symbol region Ratio=W/H.

The ratio of CCTV's station symbol generally below 0.3, so, the preliminary Rule of judgment that can build For: ratio < 0.3.But, satellite TV's platform such as Inner Mongol satellite TV, Chongqing satellite TV, BTV (bag Include the local broadcasting stations of these satellite TVs) ratio of station symbol all below 0.3, so, length and width can be passed through After screening, more again screened by gray scale and color.

When again being screened by gray scale and color, can carry out point according to following Rule of judgment Class, say, that the following condition for determining whether is set:

(1) red component in first preset range in the upper left corner, described station symbol region and described The average of the red component in second preset range in the lower right corner, station symbol region is less than presetting redness Component.

It is to say, can be divided into 5*3 sub-block (certainly, it is possible to logical by row * row in station symbol region Cross the modes such as 6*3 or 4*3 and carry out piecemeal), extract first sub-block area1 in the upper left corner (i.e. First preset range) and first sub-block area2 (the i.e. second preset range) in the lower right corner.Centre The red distribution in the two region such as television stations mark and Chongqing satellite TV station symbol, BTV's station symbol cuts The most different.

Integrated interference station symbol and the color characteristic of CCTV's station symbol, can build condition 1 (i.e. Condition1) the red average of area1 and area2 is less than 150.

(2) on the left of described station symbol region, the gray average in the 3rd preset range is grey less than presetting Angle value.

Consider fault-tolerance and the translucent feature of CCTV's station symbol of Condition1, take sub-block area3.Station symbol region can be taken the most left with 50 pixels of width in station symbol region as standard 8 row pixels (the widest 4/25) constitute area3 (the i.e. the 3rd preset range).

Analysis interference station symbol and CCTV station symbol, at the gray difference of area3, build condition 2 (Condition2) gray average of area3 is less than 100.

It will be appreciated that triple channel classics synthetic method Gray=0.33R+0.59G+0.11B can be passed through Obtain gray level image, it is possible to by triple channel maximum value process, triple channel mean value method etc., this This is not any limitation as by embodiment.

(3) it is at least 4 parts by described station symbol region segmentation, the predetermined fraction after segmentation Between pixel average absolute difference less than preset absolute difference；

Owing to Condition1 and Condition2 contains only the self information of each sub-block, need by Constraint expansion is to the relation between sub-block.Analyze and find, the word in station symbol region, digital pixel It is predominantly located at rear 1/3 row in station symbol region.To this end, station symbol region is divided into 2*3 according to row * row Sub-block (certainly, it is possible to carry out piecemeal by modes such as 3*3), takes front 2*2 sub-block and (i.e. divides Predetermined fraction after cutting) it is expressed as area4, area5, area6 and area7.

Multisample strictly calculates discovery, and CCTV station symbol is substantially following condition at these 4 sub-blocks:

The average of condition 3 (Condition3) area4, area5, area6 and area7 it Absolute difference is less than 100.

(4) variance of the pixel average between the predetermined fraction after segmentation is less than presetting variance.

It is to say, the variance of the pixel average of above 4 sub-blocks all has difference, so, CCTV station symbol meets following condition at these 4 sub-blocks:

The average sequence of condition 4 (Condition4) area4, area5, area6 and area7 The variance of row is less than 1600.

When certain station symbol sample meets above Condition1～Condition4 simultaneously, can be judged For CCTV's station symbol, otherwise it is judged as non-CCTV station symbol.When above-mentioned condition judges simultaneously, accuracy rate The highest, it is demonstrated experimentally that lack any one condition all can improve the error rate of multisample classification.

S502: when described station symbol to be identified is CCTV's station symbol, extract in described station symbol region Numeric area, carries out figure place differentiation to the numeric area extracted, and differentiates result identification according to figure place Numeral in described numeric area, by the combination between mark and the numeral of identification of CCTV's station symbol Recognition result as described station symbol to be identified；

It should be noted that described CCTV station symbol comprises mark (i.e. CCTV), word sum Word, the difference of common CCTV station symbol is word and numeral.In area of pattern recognition, numeral is known Easier than Text region, stable, quick.Meanwhile, the numeral of CCTV's station symbol can individually describe Concrete channel, thus present embodiment by the word removed in station symbol region (such as " comprehensively ", " wealth Warp " etc.), extract numeral in station symbol region (i.e. 1,2 etc.).Wherein, word is in station symbol Lower section and be flagged with obvious pixel separation, can arrange presetted pixel interval, carried by described Whether the pixel separation in the edge taken and between mark exceedes presetted pixel interval, pre-exceeding If during pixel separation, then confirm as word, it is deleted, so, following station symbol region It is and only includes numeral and the region of mark.

It will be appreciated that described numeric area is positioned at described station symbol region, and exist certain Position relationship, therefore, the position between described station symbol region and numeric area can be pre-build Relation, the positional information further according to described numeric area carries out dividing processing to described image, with Obtain described numeric area.

For station symbol region, numeric area and mark region (i.e. CCTV region) Between there is following corresponding relation:

(1) numeric area is positioned on the right side of mark region, and shared width is approximately equal to mark region 1/4；

(2) letter in numeric area and mark region is contour, accounts for CCTV's station symbol entirety high The 0.8 of degree.

So, can extract described according to the position relationship between described station symbol region and numeric area Numeric area in station symbol region.

It addition, for the ease of the numeral in numeric area is identified, can be to described digital block Numerical portion and background parts in territory carry out binary conversion treatment, if numerical portion is set to white, Background parts is set to black.

Owing to the position at described four angles of numeric area easily produces white pixel block/point, and There is also noise spot in described numeric area, numeral identification can be caused shadow by these interference information Ring, in present embodiment, can carry out the numeric area after binaryzation disturbing information deletion.

In implementing, the white at described four angles of numeric area can be deleted according in the following manner Block of pixels/point: set the horizontal width of described numeric area as W (equal to 0.25W_A, W_AFor institute State the width in station symbol region), vertical a length of H is (equal to H_A, H_AFor described station symbol region Highly), each pixel gray value is that (i, j), i is pixel vertical rows coordinate to gray, and j is pixel Point horizontal row coordinate, the gray value after conversion be Gray (i, j),

$Gray(i,j)=\left\{\begin{array}{c}\begin{array}{ccc}0,& if& (0.2H\&GreaterEqual;i,0.1W\&GreaterEqual;jorj\&GreaterEqual;0.9W)\end{array}\\ \begin{array}{ccc}0,& if& (0.8H\&le;i,0.1W\&GreaterEqual;jorj\&GreaterEqual;0.9W)\end{array}\\ \begin{array}{cc}gray(i,j),& else\end{array}\end{array}\right.$

For the noise spot in described numeric area, then can carry out noise filtering, weaken further Affect with reducing noise spot.

For improving the accuracy rate that numeral identifies further, in present embodiment, first to the number extracted Territory, block carries out figure place differentiation, differentiates the number in numeric area described in result identification further according to figure place Word.

Can in several ways it will be appreciated that the numeric area extracted is carried out differentiation, this reality Execute in mode, by row, the gray value of pixel each in described numeric area is projected, to constitute The projection vector of a length of described numeric area horizontal width, has super in string projection vector When crossing the pixel belonging to numerical portion of predetermined number, this row projection vector is identified, if There is the minimum range between adjacent two identified projection vectors and be more than predeterminable range, then will Described numeral differentiates that result is set to two, otherwise described numeral being differentiated, result is set to one.

For improving the efficiency that numeral identifies further, in present embodiment, can be by following 3 Step is according to the numeral in numeric area described in figure place differentiation result identification:

(1) obtain the white pixel region A in numeric area, and calculate the horizontal width of A W and vertical line width h.If h/w > 2, then this station symbol is the station symbol of CCTV-1.Otherwise enter Enter (2).

(2) described numeric area is carried out edge extracting, if being m*n by row * row piecemeal Individual sub-block, the most as shown in table 1.Build marginal point probability space distribution histogram, with standard 0, the digital picture marginal point probability space distribution histogram of 2～9 calculates matching probability.

Table 1 piecemeal parameter

Parameter	0.8<h/w<1.25	Other
			m	4	6
n	4	3

(3) numeral in numeric area described in result identification is differentiated according to numeral.If it is digital Differentiate result be one, then station symbol be the probability of CCTV-1, CCTV-10～CCTV15 be 0, If numeral differentiates that result is two, then station symbol is that the probability of CCTV1～CCTV9 is 0, Thus complete the numeral identification in described numeric area.

Consider to identify whether problem accurately, in present embodiment, differentiate knot according to described figure place Described numeric area is mated, at the highest matching rate and second highest matching rate by fruit with standard digital Time unequal, according to standard digital corresponding to the highest described matching rate as in described numeric area Numeral.

Certainly, when the highest matching rate and second highest matching rate are equal, the most unique identification.

S503: when described station symbol to be identified is non-CCTV's station symbol, calculate described station symbol region Edge and each standard unit target matching rate；

It will be appreciated that owing to non-CCTV station symbol generally includes: satellite TV's station symbol and local station symbol, In present embodiment, it is contemplated that the feature of satellite TV's station symbol and difference, can be divided into without word station symbol (as Dragon TV), (such as southeast satellite TV, the most above-mentioned is " inessential separable word station symbol Word ") and have and can not be kept completely separate word station symbol (such as Hebei satellite TV, the most above-mentioned " must Want word "), so, phase will be carried out according to three kinds of different satellite TVs platform station symbol types in java standard library Answer the structure of type standard station symbol.

For ease of calculating edge and each standard unit target matching rate, this enforcement in described station symbol region In mode, the edge in described station symbol region is entered with standard station symbol by spatial distribution rectangular histogram Row coupling, to calculate the edge in described station symbol region and each standard unit target matching rate.

Owing to the station symbols such as Jiangsu satellite TV, Dragon TV are only left station symbol (such circular, oval Station symbol is defined as short station symbol), length and width are smaller, the television station such as Inner Mongol satellite TV, Xinjiang satellite TV The length-width ratio of station symbol (this kind of station symbol is defined as long station symbol) is significantly greater than the length of these satellite TV's station symbols Wide ratio.So, can be using length-width ratio as class condition.Long table target ratio more than 1.5, Identification layer length-width ratio comparison condition of taking second place is: ratio >=1.5, after sorting, when mating, Then can directly mate according to the standard station symbol of identical aspect ratio type, and without with all marks Quasi-station symbol all mates, and further shorten coupling duration.

In present embodiment, the edge Edge in station symbol region is divided into the sub-block that quantity is identical A1, A2, A3 ... Am, wherein m is the number of sub-block.To each sub-block, statistics is wherein Marginal point probability, i.e. obtain spatial distribution rectangular histogram.Such as: the Edge of short station symbol is divided into 5*5 sub-block, i.e. 25 sub-blocks；Long table target Edge is divided into 5*10 (row * row) height Block, i.e. 50 sub-blocks.Each standard station symbol in standard gallery is also according to short, progress row space Distribution histogram calculates, and the spatial distribution rectangular histogram in described station symbol region and standard station symbol are by sub-block Mate.The sub-block number of meter coupling is n (initial value is 0), and total sub-block number is N, if certain The probability of sub-block is sufficiently close to (such as: probability difference < 0.05), then n=n+1；Coupling terminates, Calculating matching rate is p=n/N.Thus obtain traveling through all standard unit target matching rate arrays P.

It is to say, parameter, Δ p is used for judging whether to need to carry out Pixel Information coupling.Δ p is fixed Justice is: Δ p=p_max-p_second, wherein, p_maxMaximum for matching rate array P is (the highest Matching rate), p_secondFor second largest value (the most second highest in addition to maximum in matching rate array P Join rate).

S504: when the difference of the highest matching rate with second highest matching rate is more than the first preset difference value, will The standard unit that the highest matching rate is corresponding is denoted as the recognition result into described station symbol to be identified；

There is certain fault-tolerance in view of match cognization, set Rule of judgment Δ p >=0.25 (i.e. the One preset difference value, certainly, can be also other values), if it is determined that condition Δ p >=0.25 is set up, Then by p_maxCorresponding standard unit is designated as the recognition result as described station symbol to be identified.

S505: the difference at the highest matching rate with second highest matching rate is less than or equal to the first preset difference value Time, it is judged that whether the highest described matching rate or standard station symbol corresponding to second highest matching rate belong to local Station symbol；

When being false in Rule of judgment Δ p >=0.25, then it needs to be determined that described station symbol to be identified whether For local station symbol, in present embodiment, need the highest matching rate or second highest matching rate described in judgement Whether corresponding standard station symbol belongs to local station symbol.

S506: if the highest described matching rate or standard station symbol corresponding to second highest matching rate are not admitted to Local station symbol, then be denoted as standard unit corresponding with second highest matching rate for the highest described matching rate as treating Coupling standard station symbol, enters the colour information in described station symbol region with described standard station symbol to be matched Row coupling, to update described standard unit target matching rate to be matched, treats higher for matching rate Join the standard recognition result as described station symbol to be identified.

Owing to the highest described matching rate or standard station symbol corresponding to second highest matching rate are not admitted to ground Fang Taibiao, so, described station symbol to be identified also will not belong to local station symbol, belongs to satellite TV's station symbol Probability the highest.

In present embodiment, by the colour information in station symbol region and described standard station symbol to be matched Mate, to update described standard unit target matching rate to be matched, following steps can be passed through:

(1) station symbol region RGB color triple channel gray matrix M is extracted_R、M_G、M_B。

(2) triple channel gray areas matrix M in described station symbol region is extracted_0.5R、M_0.5G、 M_0.5B, the width of three, high respectively M_R、M_G、M_BHalf, the pixel initial point of three Shown in coordinate such as formula (1).Wherein, gray represents R, G, B respectively, and x is row-coordinate, y For row coordinate, H is matrix line number (i.e. height), and W is matrix columns (i.e. width).

$\left\{\begin{array}{c}{x}_{0,M_{0.5gray}}=\frac{H_{M_{gray}}}{4}\\ y_{0,M_{0.5gray}}=\frac{W_{M_{gray}}}{4}\end{array}\right.$ Formula (1)

The purpose extracting the high region of half-breadth half is, eliminates the background color at corner, station symbol region and does Disturb, in order to subsequent match.

(3) partitioning of matrix.By M_0.5R、M_0.5G、M_0.5BCarry out piecemeal respectively.Due to district Domain sizes reduces, thus segment partition scheme is modified to: the Edge of short station symbol is divided into 3*3 sub-block, I.e. 9 regions；Long table target Edge is divided into 3*6 (row * row) individual sub-block, Ji18Ge district Territory.

(4) RGB triple channel spatial distribution rectangular histogram builds.Piecemeal side by step (3) Case, calculates the average red gray value m in each sub-block region_Ri, average green gray value m_Gi、 Average blue gray value m_Bi, the standard station symbol in java standard library carries out piecemeal mean value computation in like manner, And carry out triple channel mean match.The number of regions of meter coupling is n_rgb, overall area numerical digit N_rgb, If the coupling of three passages is poor, (in same sub-block region, sample gray average and standard grayscale are equal The difference of value) both less than 50, then n_rgb=n_rgb+1.Coupling terminates, and calculates matching rate and is p_rgb=n_rgb/N_rgb.Thus obtain traveling through all standard unit target RGB matching rate arrays P_rgb, The matching rate of the most close station symbol is p_rgb, the matching rate of non-close station symbol is 0.

(5) matching rate updates.Comprehensive matching rate array P and P_rgbResult, update coupling Rate array P_new=P+P_rgb。

S507: if the highest described matching rate or standard station symbol corresponding to second highest matching rate belong to local Station symbol, then determine described according to the edge in described station symbol region with each standard unit target matching rate Area belonging to station symbol to be identified, is partitioned into the character area in described station symbol region, separates institute State the individual character part in character area, isolated individual character part is identified, by individual character portion The standard station symbol dividing recognition result corresponding with described area carries out characters matching, the standard that will mate Station symbol is as the recognition result of described station symbol to be identified.

It will be appreciated that the standard station symbol of described area correspondence can be regarded as all of described area Standard station symbol.

Owing to the highest described matching rate or standard station symbol corresponding to second highest matching rate belong to local broadcasting stations Mark, now, it is the highest that described station symbol to be identified belongs to local broadcasting stations' target probability, it will usually with The standard station symbol in multiple identical areas is respectively provided with higher matching rate, in the case, and can basis These matching rates determine the area belonging to described station symbol to be identified.

In present embodiment, according to the standard station symbol quantity of each department in the standard station symbol obtained Determine the area belonging to described station symbol to be identified with the standard station symbol sum of each department, can by with Lower standard:

Obtain the difference standard station symbol (bag less than the first preset difference value of matching rate and the highest matching rate Include the standard station symbol that the highest matching rate is corresponding), according to the area belonging to each standard station symbol obtained Determine the objective area belonging to most standard station symbol, it is judged that target area in each standard station symbol of acquisition Ratio between standard unit target sum corresponding to the standard station symbol quantity in territory and described target area Whether example exceedes preset ratio, the most then using described target area as described station symbol to be identified Affiliated area.

Such as: matching rate is designated as less than the standard unit of the first preset difference value with the difference of the highest matching rate In 10, and 10 standard station symbols, 8 belong to BTV, and 1 belongs to Sichuan electricity Television stations, 1 belongs to Xinjiang television station, if BTV one has 12, Sichuan TV Platform one has 6, and Xinjiang television station one has 5, and preset ratio is 50%, it may be determined that mesh Mark region is Beijing, Pekinese's standard station symbol quantity (i.e. 8) in each standard station symbol of acquisition With the ratio between standard unit target sum (i.e. 12) corresponding to Beijing is 72.7%, super Cross preset ratio (50%), now, using Beijing as the ground belonging to described station symbol to be identified District.

It should be noted that the character area extracted in described station symbol region has various ways, For improving the extraction efficiency of character area and accuracy rate, present embodiment, according to station symbol region and Position relationship and the positional information in described station symbol region between character area determine described literary composition The positional information in territory, block, and according to the positional information of described character area from described station symbol region Middle extraction character area, certainly, also by other modes, present embodiment is to this most in addition Limit.

As a example by described station symbol to be identified is for BTV's station symbol, character area and Beijing Television Mark " BTV " in platform is contour, so, described can be doubled in cut label " BTV " right side The region of mark region height is as character area, say, that the height of character area and mark The height of " BTV " is identical, the twice of a length of described station symbol region height of character area；

For improving recognition accuracy further, in present embodiment, described character area is carried out Ostu binary conversion treatment, and use connected domain mode to remove noise jamming.

It will be appreciated that for some local station symbol, such as BTV's station symbol, its difference It is word, such as " BTV ", " physical culture ", " life " etc..This difference can be by character area In first word uniquely describe, in order to improve recognition efficiency, in present embodiment, isolate The individual character part that individual character part is first word place in described character area, can pass through Below scheme realizes separating:

(1) interference straight line (section) horizontal, longitudinal in character area is removed；

(2) pixel separation between individual character part in described character area is obtained；

(3) complete individual character part is extracted according to described pixel separation, such as: literary composition (literature and art), Shadow (video display)；

(4) individual character part is carried out Threshold segmentation, cutting procedure takes upper left corner 5*5 subdomain, Calculate black ratio r, if r is less than 0.5, the monochrome pixels of exchange individual character part.

Certainly, also whole individual character parts all can be separated, then be identified respectively, this reality This is not any limitation as by mode of executing.

There is the multiple method that isolated individual character part is identified in the prior art, but In order to improve recognition accuracy and stability, with reference to Fig. 2, described to isolated individual character part It is identified including:

A1: described individual character part is carried out micronization processes；

In present embodiment, described individual character part is carried out micronization processes, to obtain described individual character The refinement edge graph of part, the process of micronization processes can refer to be executed what roc flew to write by Lv Yuehe " a kind of practical parallel thinning algorithm and realization thereof ", present embodiment does not repeats them here.

For the ease of subsequent treatment, also the standard one-letter that individual character is concentrated can be also carried out at refinement Reason, to obtain the refinement edge graph of each standard one-letter；

For making the refinement edge graph refining edge graph and each standard one-letter of described individual character part Size keeps consistent, such as: can be by the refinement edge graph of described individual character part and each standard one-letter Refinement edge graph be transformed to 50*50 pixel size.

A2: according to individual character part described in the refinement edge calculations of the individual character part after micronization processes And the edge matching rate between standard one-letter；

The edge feature of refinement word contains the stroke featuring word, can reduce coupling standard The scope of data, and provincial characteristics is the content of refinement word, enriches region description, can repair Just determining matching literal, owing to the data volume of refinement is less than provincial characteristics, in present embodiment, First pass through Edge Feature Matching, then by provincial characteristics correction, thus reduce operand.

In present embodiment, according to the refinement edge calculations of the individual character part after micronization processes During edge matching rate between individual character part and standard one-letter, below scheme can be passed through:

(1) travel through the refinement edge graph si of standard one-letter, si is divided into 100 of 10*10 Subgraph；

(2) calculating each subgraph refinement edge ratio pi_j, composition refines edge scaling matrices Mi, In like manner it is calculated the refinement edge scaling matrices M of individual character part；

(3) compare the degree of approximation of same position element in M and Mi, and then calculate edge square Battle array matching degree Pe_i；

(4) according to Ps_i+=Pe_i the i-th element of Ps (Ps_i be), new probability is formed Array Ps；

A3: whether judge the matching rate difference of high rim matching rate and second highest edge matching rate Higher than the second preset difference value；

That is, it is judged that the peak Ps_m1 of Ps and time high level Ps_m2 meet condition:

Ps_m1-Ps_m2 0.25 (0.25 can be set to other value),

A4: when described matching rate difference is higher than the second preset difference value, by described high rim Standard one-letter corresponding to matching rate is as the individual character recognition result in described individual character region；

It is to say, when meeting Ps_m1-Ps_m2 0.25, be identified as Ps_m1 corresponding Word and update matching rate.

A5: when described matching rate difference is not higher than the second preset difference value, according to micronization processes After the provincial characteristics of individual character part and standard one-letter mate, described edge matching rate is entered Row updates；

It is to say, when being unsatisfactory for Ps_m1-Ps_m2 0.25, can lead to according to provincial characteristics Cross below scheme to be modified:

Individual character part region-by-region pixel after micronization processes is determined the standard list after micronization processes Whether the pixel of word same position is area pixel, calculates whole figure matching rate pi_ds；

Standard one-letter region-by-region pixel after micronization processes is determined the individual character portion after micronization processes Whether the pixel of split-phase co-located is area pixel, calculates whole figure matching rate pi_sd；

According to Ps_i+=(pi_ds+2*pi_sd)/3 the i-th element of Ps (Ps_i be), composition is new Probability array Ps；

A6: using standard one-letter corresponding for high rim matching rate as the list in described individual character region Word recognition result.

It is to say, using word corresponding for the maximum Ps_m of Ps as identifying word, and repair Positive matching rate.

Table 1 gives the method the using present embodiment average identification to several typical case's station symbols Time.Traversal recognition time is shorter and difference little, all at about 2s.

The average recognition time of table 1 typical electrical television stations station symbol

Television station	TV station symbol recognition time (s)
		CCTV1	1.44
CCTV5	1.29
		Phoenix Satellite TV	1.76
Southeast satellite TV	1.93
		Live in Beijing	2.17
Beijing physical culture	2.25

Table 2 gives the method the using present embodiment Mean match to several typical case's station symbols Rate and discrimination.Table 2 demonstrates the effectiveness of the method for present embodiment, stability and reliable Property.

The Mean match rate of table 2 typical electrical television stations and discrimination

Television station	Mean match rate	Discrimination
			CCTV1	100%	100%
CCTV5	92.4%	95.1%
			Phoenix Satellite TV	95.2%	97.8%
Southeast satellite TV	84.0%	96.6%
			Live in Beijing	76.3%	95.4%
Beijing physical culture	82.6%	96.3%

The method of present embodiment is suitable for CCTV's platform, satellite TV's platform and the match cognization of local broadcasting stations, real The high discrimination of existing more than 95% correctly identifies, and traversal recognition time is controlled within 3s.

For embodiment of the method, in order to be briefly described, therefore it is all expressed as a series of action Combination, but those skilled in the art should know, and the embodiment of the present invention is not by described The restriction of sequence of movement, because according to the embodiment of the present invention, some step can use other suitable Sequence or simultaneously carry out.Secondly, those skilled in the art also should know, is retouched in description The embodiment stated belongs to preferred embodiment, and the involved action not necessarily present invention implements Necessary to example.

Fig. 6 is the structured flowchart of the TV station symbol recognition device of one embodiment of the present invention；With reference to figure 6, described device includes:

Station symbol judging unit 601, for judging whether the station symbol to be identified in station symbol region is centre Television stations mark；

First recognition unit 602, is used for when described station symbol to be identified is CCTV's station symbol, according to Described station symbol to be identified is identified by the first preset strategy；

Second recognition unit 603, for when described station symbol to be identified is non-CCTV's station symbol, presses According to the second preset strategy, described station symbol to be identified is identified.

In a kind of alternative embodiment of the present invention, described station symbol judging unit, it is further used for Length-width ratio, gray scale and color according to station symbol region judges that the station symbol to be identified in station symbol region is No for CCTV's station symbol.

In a kind of alternative embodiment of the present invention, described first recognition unit, it is further used for Extract the numeric area in described station symbol region, the numeric area extracted carried out figure place differentiation, Differentiate the numeral in numeric area described in result identification according to figure place, by the mark of CCTV's station symbol and Combination between the numeral identified is as the recognition result of described station symbol to be identified.

In a kind of alternative embodiment of the present invention, described first recognition unit, it is further used for Extract in described station symbol region according to the position relationship between described station symbol region and numeric area Numeric area, and the numerical portion in described numeric area and background parts are carried out binaryzation Process.

In a kind of alternative embodiment of the present invention, described first recognition unit, it is further used for By row, the gray value of pixel each in described numeric area is projected, a length of described to constitute The projection vector of numeric area horizontal width, has in string projection vector and exceedes predetermined number The pixel belonging to numerical portion time, this row projection vector is identified, if having adjacent two Minimum range between individual identified projection vector is more than predeterminable range, then described numeral sentenced Other result is set to two, otherwise described numeral being differentiated, result is set to one.

In a kind of alternative embodiment of the present invention, described second recognition unit, it is further used for Calculate the edge in described station symbol region and each standard unit target matching rate；At the highest matching rate with secondary When the difference of high matching rate is more than the first preset difference value, standard unit corresponding for the highest matching rate is denoted as Recognition result for described station symbol to be identified；Difference at the highest matching rate with second highest matching rate is less than During equal to the first preset difference value, it is judged that the highest described matching rate or standard corresponding to second highest matching rate Whether station symbol belongs to local station symbol；In the standard that the highest described matching rate or second highest matching rate are corresponding Station symbol is not admitted to local broadcasting stations' timestamp, by mark corresponding with second highest matching rate for the highest described matching rate The colour information in described station symbol region as standard station symbol to be matched, is treated by quasi-station symbol with described Join standard station symbol to mate, to update described standard unit target matching rate to be matched, will coupling The higher standard to be matched of rate is as the recognition result of described station symbol to be identified.

In a kind of alternative embodiment of the present invention, described second recognition unit, it is further used for Edge in described station symbol region is mated with standard station symbol by spatial distribution rectangular histogram, To calculate the edge in described station symbol region and each standard unit target matching rate.

In a kind of alternative embodiment of the present invention, described second recognition unit, it is further used for The standard station symbol corresponding at the highest described matching rate or second highest matching rate belongs to local broadcasting stations' timestamp, root Edge and each standard unit target matching rate according to described station symbol region determine described to be identified Area belonging to mark, is partitioned into the character area in described station symbol region, separates described literal field Individual character part in territory, is identified isolated individual character part, individual character part identification is tied The fruit standard station symbol the most corresponding with described area carries out characters matching, the standard unit of coupling is denoted as into The recognition result of described station symbol to be identified.

In a kind of alternative embodiment of the present invention, described second recognition unit, it is further used for Described individual character part is carried out micronization processes；Refinement limit according to the individual character part after micronization processes Edge calculates the edge matching rate between described individual character part and standard one-letter；Judge high rim Whether join the matching rate difference of rate and second highest edge matching rate higher than the second preset difference value；Described When matching rate difference is higher than the second preset difference value, by the described standard that high rim matching rate is corresponding Individual character is as the individual character recognition result in described individual character region；It is not higher than in described matching rate difference During two preset difference values, provincial characteristics and standard one-letter according to the individual character part after micronization processes are entered Row coupling, is updated described edge matching rate；By standard corresponding for high rim matching rate Individual character is as the individual character recognition result in described individual character region.

In a kind of alternative embodiment of the present invention, described device also includes:

Area acquisition unit, for obtaining in the predeterminable area of the video including station symbol to be identified Video frame images sequence, carries out edge extracting to each video frame images, by each video frame images Edge synthesizes, and obtains the minimum external matrix at the edge of synthesis, external according to described minimum Each video frame images is split by matrix respectively, and the image being partitioned into is passed through weighted average Mode synthesize, include the station symbol region of station symbol to be identified to obtain.

For device embodiment, due to itself and embodiment of the method basic simlarity, so describing Fairly simple, relevant part sees the part of embodiment of the method and illustrates.

Device embodiment described above is only schematically, wherein said as separated part The unit of part explanation can be or may not be physically separate, shows as unit Parts can be or may not be physical location, i.e. may be located at a place, or also Can be distributed on multiple NE.Can select according to the actual needs part therein or The whole module of person realizes the purpose of the present embodiment scheme.Those of ordinary skill in the art are not paying In the case of going out performing creative labour, i.e. it is appreciated that and implements.

Through the above description of the embodiments, those skilled in the art it can be understood that The mode of required general hardware platform can be added by software to each embodiment to realize, certainly Hardware can also be passed through.Based on such understanding, technique scheme is the most in other words to existing The part having technology to contribute can embody with the form of software product, and this computer is soft Part product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, light Dish etc., including some instructions with so that computer equipment (can be personal computer, Server, or the network equipment etc.) perform some part institute of each embodiment or embodiment The method stated.

Last it is noted that above example is only in order to illustrate technical scheme, and Non-to its restriction；Although the present invention being described in detail with reference to previous embodiment, ability The those of ordinary skill in territory is it is understood that it still can be to the skill described in foregoing embodiments Art scheme is modified, or wherein portion of techniques feature is carried out equivalent；And these are repaiied Change or replace, not making the essence of appropriate technical solution depart from various embodiments of the present invention technical side The spirit and scope of case.

Claims (20)

1. a TV station symbol recognition method, it is characterised in that described method includes:

Judge whether the station symbol to be identified in station symbol region is CCTV's station symbol；

When described station symbol to be identified is CCTV's station symbol, according to the first preset strategy, described waiting is known Other station symbol is identified；

When described station symbol to be identified is non-CCTV's station symbol, treat described according to the second preset strategy Identification station symbol is identified.

Method the most according to claim 1, it is characterised in that described judgement station symbol district Whether the station symbol to be identified in territory is CCTV's station symbol, farther includes:

Length-width ratio, gray scale and color according to station symbol region judges to be identified in station symbol region Whether mark is CCTV's station symbol.

Method the most according to claim 1, it is characterised in that described pre-according to first If described station symbol to be identified is identified by strategy, farther include:

Extract the numeric area in described station symbol region, the numeric area extracted is carried out figure place and sentences Not, the numeral in numeric area described in result identification is differentiated according to figure place, by the mark of CCTV's station symbol Combination between will and the numeral of identification is as the recognition result of described station symbol to be identified.

Method the most according to claim 3, it is characterised in that described of described extraction Numeric area in mark region, farther includes:

Described station symbol district is extracted according to the position relationship between described station symbol region and numeric area Numeric area in territory, and the numerical portion in described numeric area and background parts are carried out two Value processes.

Method the most according to claim 4, it is characterised in that the described number to extracting Territory, block carries out figure place differentiation, farther includes:

By row, the gray value of pixel each in described numeric area is projected, a length of to constitute The projection vector of described numeric area horizontal width, have in string projection vector exceed default During the pixel belonging to numerical portion of quantity, this row projection vector is identified, if there is phase Minimum range between adjacent two identified projection vectors is more than predeterminable range, then by described number Word differentiates that result is set to two, otherwise described numeral being differentiated, result is set to one.

Method the most according to claim 1, it is characterised in that described pre-according to second If described station symbol to be identified is identified by strategy, farther include:

Calculate the edge in described station symbol region and each standard unit target matching rate；

When the difference of the highest matching rate with second highest matching rate is more than the first preset difference value, by the highest Join standard unit corresponding to rate and be denoted as the recognition result into described station symbol to be identified；

When the difference of the highest matching rate with second highest matching rate is less than or equal to the first preset difference value, it is judged that Whether the highest described matching rate or standard station symbol corresponding to second highest matching rate belong to local station symbol；

If the highest described matching rate or standard station symbol corresponding to second highest matching rate are not admitted to place Station symbol, then be denoted as standard unit corresponding with second highest matching rate for the highest described matching rate as to be matched Standard station symbol, is carried out the colour information in described station symbol region and described standard station symbol to be matched Join, to update described standard unit target matching rate to be matched, by mark to be matched higher for matching rate The accurate recognition result as described station symbol to be identified.

Method the most according to claim 6, it is characterised in that described of described calculating The edge in mark region and each standard unit target matching rate, farther include:

Edge in described station symbol region is carried out with standard station symbol by spatial distribution rectangular histogram Coupling, to calculate the edge in described station symbol region and each standard unit target matching rate.

Method the most according to claim 6, it is characterised in that described in described judgement After whether high matching rate or standard station symbol corresponding to second highest matching rate belong to local station symbol, described Method also includes:

If the highest described matching rate or standard station symbol corresponding to second highest matching rate belong to local station symbol, Then according to the edge in described station symbol region and each standard unit target matching rate determine described in wait to know Area belonging to other station symbol, is partitioned into the character area in described station symbol region, separates described literary composition Individual character part in territory, block, is identified isolated individual character part, individual character part is known The other result standard station symbol corresponding with described area carries out characters matching, will the standard station symbol of coupling Recognition result as described station symbol to be identified.

Method the most according to claim 8, it is characterised in that described to isolated Individual character part is identified, and farther includes:

Described individual character part is carried out micronization processes；

According to individual character part and mark described in the refinement edge calculations of the individual character part after micronization processes Edge matching rate between quasi-individual character；

Whether judge the edge matching rate difference of high rim matching rate and second highest edge matching rate Higher than the second preset difference value；

When described edge matching rate difference is higher than the second preset difference value, by described high rim Join standard one-letter corresponding to the rate individual character recognition result as described individual character region；

When described edge matching rate difference is not higher than the second preset difference value, after micronization processes The provincial characteristics of individual character part and standard one-letter mate, described edge matching rate is carried out Update；

Standard one-letter corresponding for high rim matching rate is known as the individual character in described individual character region Other result.

10. according to the method according to any one of claim 1～9, it is characterised in that described Before judging whether the station symbol to be identified in station symbol region is CCTV's station symbol, described method is also wrapped Include:

Video frame images sequence is obtained in the predeterminable area of the video including station symbol to be identified, right Each video frame images carries out edge extracting, is synthesized at the edge of each video frame images, obtains The minimum external matrix at the edge of synthesis, according to described minimum external matrix respectively to each frame of video Image is split, and is synthesized by average weighted mode by the image being partitioned into, with Obtain the station symbol region including station symbol to be identified.

11. 1 kinds of TV station symbol recognition devices, it is characterised in that described device includes:

Station symbol judging unit, for judging whether the station symbol to be identified in station symbol region is CCTV's platform Mark；

First recognition unit, for when described station symbol to be identified is CCTV's station symbol, according to first Described station symbol to be identified is identified by preset strategy；

Second recognition unit, for when described station symbol to be identified is non-CCTV's station symbol, according to the Described station symbol to be identified is identified by two preset strategy.

12. devices according to claim 11, it is characterised in that described station symbol judges Unit, is further used for the length-width ratio according to station symbol region, gray scale and color and judges station symbol region In station symbol to be identified whether be CCTV's station symbol.

13. devices according to claim 11, it is characterised in that described first identifies Unit, is further used for extracting the numeric area in described station symbol region, to the digital block extracted Territory carries out figure place differentiation, differentiates the numeral in numeric area described in result identification according to figure place, will Combination between mark and the numeral of identification of CCTV's station symbol is as the knowledge of described station symbol to be identified Other result.

14. devices according to claim 13, it is characterised in that described first identifies Unit, is further used for extracting according to the position relationship between described station symbol region and numeric area Numeric area in described station symbol region, and to the numerical portion in described numeric area and background Part carries out binary conversion treatment.

15. devices according to claim 14, it is characterised in that described first identifies Unit, is further used for projecting the gray value of pixel each in described numeric area by row, To constitute the projection vector of a length of described numeric area horizontal width, in string projection vector When there is the pixel belonging to numerical portion exceeding predetermined number, this row projection vector is marked Know, if the minimum range existed between adjacent two identified projection vectors more than preset away from From, then described numeral is differentiated that result is set to two, otherwise described numeral is differentiated that result is set to One.

16. devices according to claim 11, it is characterised in that described second identifies Unit, is further used for the edge calculating described station symbol region and each standard unit target matching rate； When the difference of the highest matching rate with second highest matching rate is more than the first preset difference value, by the highest matching rate Corresponding standard unit is denoted as the recognition result into described station symbol to be identified；At the highest matching rate with secondary When the difference of high matching rate is less than or equal to the first preset difference value, it is judged that described the highest matching rate or second highest Whether the standard station symbol that matching rate is corresponding belongs to local station symbol；At described the highest matching rate or second highest The standard station symbol that matching rate is corresponding is not admitted to local broadcasting stations' timestamp, by the highest described matching rate with secondary The standard unit that high matching rate is corresponding is denoted as standard station symbol to be matched, by the coloured silk in described station symbol region Color information is mated with described standard station symbol to be matched, to update described standard station symbol to be matched Matching rate, using standard to be matched higher for matching rate as described station symbol to be identified identification tie Really.

17. devices according to claim 16, it is characterised in that described second identifies Unit, is further used for the edge in described station symbol region by spatial distribution rectangular histogram and mark Quasi-station symbol mates, and mates with each standard unit target calculating the edge in described station symbol region Rate.

18. devices according to claim 16, it is characterised in that described second identifies Unit, is further used for the standard station symbol the highest described matching rate or second highest matching rate are corresponding and belongs to In local broadcasting stations' timestamp, come really with each standard unit target matching rate according to the edge in described station symbol region Fixed area belonging to described station symbol to be identified, is partitioned into the character area in described station symbol region, Separate the individual character part in described character area, isolated individual character part is identified, will The standard station symbol that individual character partial recognition result is corresponding with described area carries out characters matching, will coupling Standard unit be denoted as the recognition result into described station symbol to be identified.

19. devices according to claim 18, it is characterised in that described second identifies Unit, is further used for described individual character part is carried out micronization processes；After micronization processes Edge matching between individual character part and standard one-letter described in the refinement edge calculations of individual character part Rate；Judge that the matching rate difference of high rim matching rate and second highest edge matching rate is whether higher than the Two preset difference values；When described matching rate difference is higher than the second preset difference value, by described flash Standard one-letter corresponding to edge matching rate is as the individual character recognition result in described individual character region；Described When matching rate difference is not higher than the second preset difference value, according to the district of the individual character part after micronization processes Characteristic of field and standard one-letter mate, and are updated described edge matching rate；By flash Standard one-letter corresponding to edge matching rate is as the individual character recognition result in described individual character region.

20. according to the device according to any one of claim 11～19, it is characterised in that institute State device also to include:

Area acquisition unit, for obtaining in the predeterminable area of the video including station symbol to be identified Video frame images sequence, carries out edge extracting to each video frame images, by each video frame images Edge synthesizes, and obtains the minimum external matrix at the edge of synthesis, external according to described minimum Each video frame images is split by matrix respectively, and the image being partitioned into is passed through weighted average Mode synthesize, include the station symbol region of station symbol to be identified to obtain.