CN103729636B

CN103729636B - Character segmentation method, device and electronic equipment

Info

Publication number: CN103729636B
Application number: CN201310699001.5A
Authority: CN
Inventors: 陈志军; 张涛; 王琳; 秦秋平
Original assignee: Xiaomi Inc
Current assignee: Xiaomi Inc
Priority date: 2013-12-18
Filing date: 2013-12-18
Publication date: 2017-07-25
Anticipated expiration: 2033-12-18
Also published as: CN103729636A

Abstract

The embodiment of the present application discloses a kind of Character segmentation method, device and electronic equipment, and methods described includes：Obtain the current recognition result and at least one history recognition result of target area, the current recognition result is the recognition result of the current frame image of target area, the history recognition result is the recognition result of the history image of target area, and the current recognition result and the history recognition result include at least one character block；Obtain the confidence level of each character block and the position in image in the current recognition result and history recognition result；According to the confidence level of each character block and position, the wrong cutting zone in the current recognition result is obtained.The recognition result auxiliary that the application integrates history image is determined to the wrong cutting zone of current frame image, and then when being cut again to wrong cutting zone, the blindness cut to current frame image can be reduced, cutting accuracy and cutting efficiency is improved.

Description

Character segmentation method, device and electronic equipment

Technical field

The application is related to image identification technical field, more particularly to a kind of Character segmentation method, device and electronic equipment.

Background technology

OCR（Optical Character Recognition, optical character identification）Refer to electronic equipment（For example scan Instrument or camera etc.）The character on paper or on other objects is checked, by detecting that dark, bright pattern determines its shape, word is then used The process of computword is translated into shape in symbol recognition methods；Text information is scanned, then image file entered Row analyzing and processing, obtains the process of word and layout information.

Current OCR Character segmentation technology mainly has the Character segmentation technology based on Recognition feedback, and the technology is first to list Two field picture is pre-processed, including binaryzation, smooth, denoising etc., Character segmentation is then carried out, identifier pair is utilized after dicing Cutting result is identified, and uses recognition result as the foundation for evaluating current cutting result correctness, if cutting knot Fruit None- identified, then it is assumed that current cutting result mistake, then attempts another cutting method to the character in single-frame images Again cut, until finding the cutting result that can be recognized, or, number of attempt exceedes after preset times, reacquires another Two field picture, and repeat said process.

Because cutting has larger blindness every time, progress is generally required to a single-frame images character recognition of completion many It is secondary to attempt cutting and recognize, in some instances it may even be possible to be that cutting is correct, but identifier identification mistake, it is also desirable to cut again, and it is each Operand is very big in identification process, takes longer.So the Character segmentation technical efficiency based on Recognition feedback is low, it is impossible to meet The requirement of real-time of identification.

The content of the invention

This application provides a kind of Character segmentation method, device and electronic equipment, deposited with solving correlation technique Character segmentation The problem of efficiency is low.

In order to solve the above-mentioned technical problem, this application discloses following technical scheme：

In a first aspect, this application provides a kind of Character segmentation method, methods described includes：Obtain the current of target area Recognition result and at least one history recognition result, the current recognition result is the identification knot of the current frame image of target area Really, the history recognition result is the recognition result of the history image of target area, the current recognition result and the history Recognition result includes at least one character block；Obtain each character block in the current recognition result and history recognition result Confidence level and the position in image；According to the confidence level of each character block and position, obtain in the current recognition result Wrong cutting zone.

With reference in a first aspect, in first aspect in the first possible implementation, methods described also includes：Obtain described Cutting character block and adhesion character block are crossed in mistake cutting zone；Cutting character block and the adhesion character block excessively are cut Cut, obtain target cutting result.

The first possible implementation with reference to first aspect, in second of possible implementation of first aspect, institute Cutting character block and the adhesion character block excessively obtained in the wrong cutting zone is stated, including：Obtain target binaryzation parameter, institute Stating target binaryzation parameter includes：Character average height in the recognition result of history image；Joined using the target binaryzation It is several that binary conversion treatment is carried out to the current recognition result；Obtain the character in the current recognition result and history recognition result Confidence width；Character confidence width in the current recognition result and history recognition result, obtains target character confidence Width；Wrong cutting area described in the current recognition result after binary conversion treatment is obtained according to the target character confidence width Cutting character block and adhesion character block are crossed in domain.

With reference to second of possible implementation of first aspect, in first aspect in the third possible implementation, institute The character confidence width in the current recognition result and history recognition result is stated, target character confidence width, bag is obtained Include：Count relation and higher limit between character confidence width lower limit in the current recognition result and history recognition result Between relation；When the first quantity is less than the second quantity, word in the current recognition result is reduced according to default first correction value Confidence width higher limit and lower limit are accorded with, target character confidence width is obtained；When the first quantity is more than the second quantity, according to pre- If the second correction value increases the higher limit and lower limit of the character confidence width in the current recognition result, target character is obtained Confidence width；Wherein, first quantity refers to that character confidence width lower limit is less than character in the current recognition result and put Believe the quantity of the history recognition result of width lower limit, the second quantity refers to that character confidence width higher limit is more than the current knowledge The quantity of the history recognition result of character confidence width higher limit in other result.

The first possible implementation with reference to first aspect, in the 4th kind of possible implementation of first aspect, institute State and cutting character block and the adhesion character block excessively are cut, obtain target cutting result, including：Obtain to described current Recognition result carries out corresponding history Wave crest and wave trough ratio value during history cutting；According to the target character confidence width adjustment institute History Wave crest and wave trough ratio value is stated, target cut point is obtained；In the target cut point cutting character block and adhesion are crossed to described Character block is cut, and obtains target cutting result.

With reference to the 4th kind of possible implementation of first aspect, in the 5th kind of possible implementation of first aspect, institute The history Wave crest and wave trough ratio value according to the target character confidence width adjustment is stated, target cut point is obtained, including：Obtain The first object cutting zone of the adhesion character block, the center of the first object cutting zone and the adhesion character Distance is equal to the target character confidence width between block starting point；According to the history Wave crest and wave trough ratio value in first mesh Search whether there is trough in mark cutting zone；When in the absence of trough, the history Wave crest and wave trough ratio value is turned down, is obtained One refers to Wave crest and wave trough ratio value；Wave crest and wave trough ratio value is referred to using described first, the first object cutting zone is searched With the presence or absence of trough, until finding trough；When there is trough, the trough position found is defined as the adhesion word Accord with the target cut point of block.

With reference to the 5th kind of possible implementation of first aspect, in the 6th kind of possible implementation of first aspect, institute The history Wave crest and wave trough ratio value according to the target character confidence width adjustment is stated, target cut point is obtained, in addition to：Sentence Whether the disconnected adjustment number of times to Wave crest and wave trough ratio value is more than preset times threshold value；When to the adjustment number of times of Wave crest and wave trough ratio value During more than preset times threshold value, by the adhesion character block or mistake cutting character block projected pixel on the axle in character arrangements direction The minimum position of point quantity is defined as target cut point.

With reference to the 4th kind of possible implementation of first aspect, in the 7th kind of possible implementation of first aspect, institute State according to history Wave crest and wave trough ratio value described in the target character confidence width adjustment, obtain target cut point, including：Obtain institute The second target cutting zone of cutting character block was stated, cutting word is crossed in the center of the second target cutting zone with described Distance is equal to the target character confidence width between according with block starting point；According to the history Wave crest and wave trough ratio value described second Search whether there is trough in target cutting zone；When there is trough, the history Wave crest and wave trough ratio value is heightened, is obtained Two refer to Wave crest and wave trough ratio value；Wave crest and wave trough ratio value is referred to using described second, the second target cutting zone is searched With the presence or absence of trough, until can not find trough；When in the absence of trough, it will be found before last time adjustment Wave crest and wave trough ratio value Trough position be defined as it is described cross cutting character block target cut point.

With reference to the 7th kind of possible implementation of first aspect, in the 8th kind of possible implementation of first aspect, institute The history Wave crest and wave trough ratio value according to the target character confidence width adjustment is stated, target cut point is obtained, in addition to：Sentence Whether the disconnected adjustment number of times to Wave crest and wave trough ratio value is more than preset times threshold value；When to the adjustment number of times of Wave crest and wave trough ratio value During more than preset times threshold value, by the adhesion character block or mistake cutting character block projected pixel on the axle in character arrangements direction The minimum position of point quantity is defined as target cut point.

The first possible implementation with reference to first aspect, in the 9th kind of possible implementation of first aspect, institute Stating method also includes：Judge whether to receive the operation requests that user inputs, the operation requests include：Rescan request, Retry request or cancel identification request；When receiving the operation requests, the confidence level of the target cutting result is obtained, and The confidence level is reduced as the objective degrees of confidence of the target cutting result；When not receiving the operation requests, institute is exported State target cutting result.

With reference to first aspect or first aspect the first to the 9th kind of any one possible implementation, in first aspect In tenth kind of possible implementation, the confidence level of each character block of basis and position obtain the current recognition result In wrong cutting zone, including：According to the confidence information of each character block and position, obtain in the current recognition result Character block different zones, wherein, the current recognition result and history are known in same position in the character block different zones Character block in other result is all differed or part is differed；The character block different zones are defined as wrong cutting area Domain.

With reference to first aspect or first aspect the first to the 9th kind of any one possible implementation, in first aspect In a kind of tenth possible implementation, the confidence information of each character block of basis and position obtain the current knowledge Wrong cutting zone in other result, including：According to the confidence information of character block in the current recognition result, obtain described Confidence level is less than the character block of predetermined threshold value in current recognition result；Confidence level is less than to the character block location of predetermined threshold value Domain, is defined as wrong cutting zone；According to the confidence information of each character block and position, obtain in the current recognition result Character block different zones, wherein, the current recognition result and history are known in same position in the character block different zones Character block in other result is all differed or part is differed；The character block different zones are defined as wrong cutting area Domain.

With reference to a kind of possible implementation of first aspect the tenth, in the 12nd kind of possible implementation of first aspect In, it is described that the character block region that confidence level is less than predetermined threshold value is defined as wrong cutting zone, including：Judge whether to deposit The character block region for being less than predetermined threshold value at least two confidence levels is adjacent；When at least two confidence levels are less than predetermined threshold value Character block region it is adjacent when, merge that position is adjacent and confidence level be less than the character block of predetermined threshold value；By the word after merging The confidence level for according with the region of block and not merging is defined as wrong cutting zone less than the character block region of predetermined threshold value.

Second aspect, this application provides a kind of Character segmentation device, including：Recognition result acquiring unit, for obtaining The current recognition result of target area and at least one history recognition result, the current recognition result is the current of target area The recognition result of two field picture, the history recognition result is the recognition result of the history image of target area, the current identification As a result at least one character block is included with the history recognition result；Character block information acquisition unit, for obtaining described work as The confidence level of each character block and the position in image in preceding recognition result and history recognition result；Mistake cutting zone is obtained Unit is taken, for the confidence level according to each character block and position, the wrong cutting zone in the current recognition result is obtained.

With reference to second aspect, in second aspect in the first possible implementation, described device also includes：First character Block acquiring unit, cutting character block and adhesion character block are crossed for obtaining in the wrong cutting zone；Cutter unit, is used for Cutting character block and the adhesion character block excessively are cut, target cutting result is obtained.

The first possible implementation with reference to second aspect, in second of possible implementation of second aspect, institute Stating the first character block acquiring unit includes：Target binaryzation parameter acquiring subelement, it is described for obtaining target binaryzation parameter Target binaryzation parameter includes：Character average height in the recognition result of history image；Binary conversion treatment subelement, for profit Binary conversion treatment is carried out to the current recognition result with the target binaryzation parameter；Character confidence width obtains subelement, For obtaining the character confidence width in the current recognition result and history recognition result；Target character confidence width obtains son Unit, for the character confidence width in the current recognition result and history recognition result, obtains target character confidence Width；First character block obtains subelement, for working as according to after target character confidence width acquisition binary conversion treatment Described in preceding recognition result cutting character block and adhesion character block are crossed in wrong cutting zone.

With reference to second of possible implementation of second aspect, in second aspect in the third possible implementation, institute Stating target character confidence width acquisition subelement includes：Statistical module, for counting the current recognition result and history identification As a result relation between the relation and higher limit between middle character confidence width lower limit；Quantity comparison module, for comparing The size of one quantity and the second quantity；Character confidence width adjustment module, for when the first quantity be less than the second quantity when, according to Default first correction value reduces character confidence width higher limit and lower limit in the current recognition result, obtains target character and puts Believe width；When the first quantity is more than the second quantity, increase the word in the current recognition result according to default second correction value The higher limit and lower limit of confidence width are accorded with, target character confidence width is obtained；Wherein, first quantity refers to character confidence Width lower limit is less than the quantity of the history recognition result of character confidence width lower limit in the current recognition result, the second number Amount refers to that character confidence width higher limit is more than the history identification knot of character confidence width higher limit in the current recognition result The quantity of fruit.

The first possible implementation with reference to second aspect, in the 4th kind of possible implementation of second aspect, institute Stating cutter unit includes：Ratio value obtains subelement, correspondence when history cutting is carried out to the current recognition result for obtaining History Wave crest and wave trough ratio value；Target cut point obtains subelement, for according to the target character confidence width adjustment institute History Wave crest and wave trough ratio value is stated, target cut point is obtained；Subelement is cut, for being cut in the target cut point to the mistake Divide character block and adhesion character block to be cut, obtain target cutting result.

With reference to the 4th kind of possible implementation of second aspect, in the 5th kind of possible implementation of second aspect, institute Stating target cut point acquisition subelement includes：Target cutting zone acquisition module, first for obtaining the adhesion character block In target cutting zone and the acquisition second target cutting zone for crossing cutting character block, the first object cutting zone Distance is equal to the target character confidence width, the second target cutting area between heart position and the adhesion character block starting point Distance is equal to the target character confidence width between the center in domain and the cutting character block starting point excessively；Trough searches mould Block, for according to the history Wave crest and wave trough ratio value in the first object cutting zone or in the second target cutting zone Search whether there is trough；Ratio value adjusting module, for when trough is not present in the first object cutting zone, turning down The history Wave crest and wave trough ratio value, obtains first and refers to Wave crest and wave trough ratio value；When the second target cutting zone internal memory In trough, the history Wave crest and wave trough ratio value is heightened, second is obtained and refers to Wave crest and wave trough ratio value；The trough searches mould Block is additionally operable to refer to Wave crest and wave trough ratio value using described first, searches the first object cutting zone and whether there is trough, Until trough is found, or, Wave crest and wave trough ratio value is referred to using described second, whether the second target cutting zone is searched There is trough, until can not find trough；First object cut point determining module, Wave crest and wave trough ratio is referred to for working as using first When value finds trough, the trough position found is defined as to the target cut point of the adhesion character block, or, profit When being searched with described second with reference to Wave crest and wave trough ratio value less than trough, it will be found before last time adjustment Wave crest and wave trough ratio value Trough position be defined as it is described cross cutting character block target cut point.

With reference to the 5th kind of possible implementation of second aspect, in the 6th kind of possible implementation of second aspect, institute Stating target cut point acquisition subelement also includes：Number of times judge module, for judging the adjustment number of times to Wave crest and wave trough ratio value Whether preset times threshold value is more than；Second target cut point determining module, for when to the adjustment number of times of Wave crest and wave trough ratio value During more than preset times threshold value, by the adhesion character block or mistake cutting character block projection image's vegetarian refreshments number on character arrangements direction The minimum position of amount is defined as target cut point.

The first possible implementation with reference to second aspect, in the 7th kind of possible implementation of second aspect, institute Stating device also includes：Operation requests receiving unit, the operation requests for receiving user's input, the operation requests include：Two Secondary scan request, retry request or cancel identification request；Judging unit is asked, for judging whether to receive the behaviour that user inputs Ask；Objective degrees of confidence determining unit, for when receiving the operation requests, obtaining putting for the target cutting result Reliability, and the confidence level is reduced as the objective degrees of confidence of the target cutting result；Output unit, for ought not receive institute When stating operation requests, the target cutting result is exported.

With reference to second aspect or second aspect the first to the 7th kind of any one possible implementation, in second aspect In 8th kind of possible implementation, the wrong cutting zone acquiring unit includes：Character block different zones obtain subelement, For the confidence information according to each character block and position, the character block different zones in the current recognition result are obtained, Wherein, the character block in the character block different zones in same position in the current recognition result and history recognition result is complete Portion is differed or part is differed；First wrong cutting zone determination subelement, for the character block different zones to be determined For wrong cutting zone.

With reference to second aspect or second aspect the first to the 7th kind of any one possible implementation, in second aspect In 9th kind of possible implementation, the wrong cutting zone acquiring unit includes：Second character block acquiring unit, for root According to the confidence information of character block in the current recognition result, obtain confidence level in the current recognition result and be less than default threshold The character block of value；Second mistake cutting determination subelement, the character block region for confidence level to be less than to predetermined threshold value, really It is set to mistake cutting zone；Character block different zones obtain subelement, for the confidence information according to each character block and position Put, obtain the character block different zones in the current recognition result, wherein, in the character block different zones in same position Character block in the current recognition result and history recognition result is all differed or part is differed；First wrong cutting area Domain determination subelement, for the character block different zones to be defined as into wrong cutting zone.

With reference to the 9th kind of possible implementation of second aspect, in the tenth kind of possible implementation of second aspect, the Two mistake cutting determination subelements include：Adjacent area judge module, for judging whether that at least two confidence levels are less than The character block region of predetermined threshold value is adjacent；Region merging technique module, for being less than predetermined threshold value when at least two confidence levels When character block region is adjacent, merges that position is adjacent and confidence level is less than the character block of predetermined threshold value；Second mistake cutting is true Cover half block, the character block region that the region of the character block after merging and the confidence level not merged are less than predetermined threshold value is determined For wrong cutting zone.

The third aspect, present invention also provides a kind of electronic equipment, including：Including memory, and one or one More than instruction, one of them or more than one instruction be stored in memory, and be configured to by one or one with Upper computing device is one or more than one instruction includes the instruction for being used to be operated below：Obtain target area Current recognition result and at least one history recognition result, the current recognition result is the knowledge of the current frame image of target area Other result, the history recognition result is the recognition result of the history image of target area, the current recognition result and described History recognition result includes at least one character block；Obtain each character in the current recognition result and history recognition result The confidence level of block and the position in image；According to the confidence level of each character block and position, the current identification knot is obtained Wrong cutting zone in fruit.

Because during OCR identification scannings are carried out, target area typically can continuously come across multiple continuous frames Within, i.e., there is certain relevance between the recognition result obtained after the multiple image identification of target area.

The beneficial effect of the application at least includes：When the recognition result to current frame image is cut, by obtaining The confidence level of each character block and the position in image in current frame image and at least recognition result of a frame history image, The relevance between multiple image recognition result can be utilized, mistake cutting present in the recognition result of current frame image is determined Region.Therefore, the recognition result auxiliary of the application synthesis history image is determined to the wrong cutting zone of current frame image, And then when being cut again to wrong cutting zone, it is possible to reduce the blindness cut to current frame image, is improved Cut accuracy and cutting efficiency.

Brief description of the drawings

, below will be to embodiment or existing in order to illustrate more clearly of the embodiment of the present application or technical scheme of the prior art There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, for those of ordinary skill in the art Speech, without having to pay creative labor, can also obtain other accompanying drawings according to these accompanying drawings.

The schematic flow sheet for the Character segmentation method that Fig. 1 provides for the application one embodiment；

The application scenario diagram for the Character segmentation method that Fig. 2 provides for the embodiment of the present application；

A kind of schematic flow sheet that Fig. 3 is step S103 in Fig. 1；

Another schematic flow sheet that Fig. 4 is step S103 in Fig. 1；

Fig. 5 is the gray level image of the upper target area of card in Fig. 2；

The binary image that Fig. 6 for the gray level image of target area obtain after binaryzation；

Fig. 7 is the connected domain schematic diagram in the binary image shown in Fig. 6；

Fig. 8 is the schematic diagram that the binary image shown in Fig. 6 is tentatively cut；

Fig. 9 is to carry out the recognition result schematic diagram after tentatively cutting to the gray level image of target area；

The schematic flow sheet for the Character segmentation method that Figure 10 provides for another embodiment of the application；

A kind of schematic flow sheet that Figure 11 is step S104 in Fig. 4；

A kind of schematic flow sheet that Figure 12 is step S105 in Fig. 4；

The schematic diagram that image pixel is projected after the binaryzation that Figure 13 provides for the embodiment of the present application；

Figure 14 is the schematic diagram of the corresponding recognition result of previous frame image of current frame image；

Figure 15 is the schematic diagram of the corresponding recognition result of front cross frame image of current frame image；

Figure 16 is the schematic diagram of the corresponding recognition result of current frame image first three two field picture；

Figure 17 is the schematic diagram of the target identification result obtained after being cut again to the current recognition result of current frame image；

The schematic flow sheet for the Character segmentation method that Figure 18 provides for the application another embodiment；

The structural representation for the Character segmentation device that Figure 19 provides for the embodiment of the application one；

The structural representation for the Character segmentation device that Figure 20 provides for another embodiment of the application；

The structural representation for the first character block acquiring unit that Figure 21 provides for another embodiment of the application；

The structural representation for the cutter unit that Figure 22 provides for another embodiment of the application；

The structural representation for the Character segmentation device that Figure 23 provides for the another embodiment of the application；

The structural representation for the electronic equipment that Figure 24 provides for the application another embodiment.

Embodiment

In order that those skilled in the art more fully understand the technical scheme in the embodiment of the present application, and make the application real Applying the above-mentioned purpose of example, feature and advantage can be more obvious understandable, below in conjunction with the accompanying drawings to technical side in the embodiment of the present application Case is described in further detail.

The schematic flow sheet for the Character segmentation method that Fig. 1 provides for the application one embodiment.The embodiment of the present application is provided This method can apply to take pictures or shoot function mobile terminal, for example：Mobile phone, camera, video camera or flat board electricity Brain, as shown in Fig. 2 in the embodiment of the present application, mobile terminal 1 00 is mobile phone, 200 be the card for carrying number to be identified, example 300 be target area to be identified in such as bank card, credit card, rechargeable card or member card, figure.

As shown in figure 1, this method can include：

S101：Obtain the current recognition result and at least one history recognition result of target area.

Current recognition result is the recognition result of the current frame image of target area, and history recognition result is target area The recognition result of history image.During OCR identification scannings are carried out, target area typically can continuously come across multiple companies Within continuous frame, this causes there is certain relevance between the recognition result that the multiple image of target area is obtained after recognizing, So including the character in target area in current recognition result and the history recognition result.

In the embodiment of the present application, the character in target area includes at least one character block, alternatively, and character block can Think one or more combinations in numeral, letter, Chinese character and symbol.Alternatively, in target area content is different, knows Object in other result can also be other patterns, for example：Quick Response Code etc..

Alternatively, the corresponding current frame image of current recognition result history image corresponding with history recognition result can be swept Retouch equipment continuously to take pictures to target area acquisition, or discontinuous take pictures acquisition, and current frame image and history image Photographical condition can be with identical, can also be different.In the embodiment of the present application, photographical condition includes but is not limited to angle of taking pictures Illumination when spending, taking pictures.

S102：Obtain in the current recognition result and history recognition result the confidence level of each character block and positioned at image In position.

The confidence level of character block characterizes the accuracy height of the character block of this in recognition result, and confidence level is higher, recognizes The character block accuracy arrived is higher.Character block is located at the relative position that the position in image refers to be located in image, for example：Work as mesh When marking in region as string number, the positional representation of the character block character block is located at the relative position in numeric string.

S103：According to the confidence level of each character block and position, the wrong cutting area in the current recognition result is obtained Domain.

Mistake cutting zone refers to that the character block in the region is undesirable, such as the character block in wrong cutting zone May be cutting character block, for example：Chinese character " bright " is cut into " day " and " moon ", or, character block may be adhesion character Block, for example：Character string " Guo Mingmin " is cut into " enjoying ", " the positive moon " and " quick ", wherein " enjoying " was cutting character block, and " sun Month " it is adhesion character block, in other embodiments, adhesion character block is possible adjacent multiple Chinese characters.

Alternatively, in the embodiment of the present application, as shown in figure 3, step S103 can include：

S1031：According to the confidence information of each character block and position, the character block in the current recognition result is obtained Different zones.

Character block differs the character block in same position in the current recognition result and history recognition result in region All differ or part is differed.For character block, judge that two character blocks are identical or differ, character can be passed through The confidence information of block compares, for differing the character block in recognition result on same position, if all character blocks Confidence level is all identical, or, in the case where allowing to have error, the confidence level difference between any two character block is both less than pre- If difference, it is possible to think that character block is identical, otherwise, it is believed that character block is different, including：Character block is all differed or portion Split-phase is identical.

S1032：The character block different zones are defined as wrong cutting zone.

Alternatively, in the embodiment of the present application, as shown in figure 4, step S103 can include：

S1033：According to the confidence information of character block in the current recognition result, obtain in the current recognition result Confidence level is less than the character block of predetermined threshold value.

Confidence level is less than predetermined threshold value, represents that the accuracy for the character block that identification is obtained is relatively low.In the embodiment of the present application, The empirical value that confidence level predetermined threshold value can be obtained according to history experimental data.

S1034：Confidence level is less than to the character block region of predetermined threshold value, is defined as wrong cutting zone.

Alternatively, in the embodiment of the present application, step S103 both can individually include step S1031~S1032, can also Individually include step S1033~S1034, furthermore it is also possible to include simultaneously：All steps in S1031~S1034.

From above technical scheme, because during OCR identification scannings are carried out, target area typically can be continuous Come across within multiple continuous frames, i.e., have necessarily between the recognition result obtained after the multiple image identification of target area Relevance.Obtain the current frame image and at least recognition result of a frame history image of target area first in this way, so The confidence level of all character blocks and the position in image in each recognition result are obtained afterwards, finally can be according to each word Confidence level and the position of block are accorded with, the wrong cutting zone in the recognition result of the current frame image is obtained.

And then when subsequently being cut to the recognition result of current frame image, it is possible to use the identification knot of current frame image The wrong cutting zone determined in fruit, auxiliary is cut again, improves the accuracy cut again, and due to cutting again When can be just for wrong cutting zone, it is possible to avoid the blindness cut again to current frame image, raising is cut The efficiency cut.

In the embodiment of the present application, the current recognition result for target area being obtained in above-mentioned steps S101 can use following Mode：

01）：The gray level image of target area is obtained, as shown in figure 5, including a string of Arabic numerals in figure.

02）：The gray level image of acquisition is carried out after self-adaption binaryzation processing, the foreground pixel in gray level image is obtained （Constitute the pixel of character block）, binary image is obtained, as shown in Figure 6.

03）：The connected domain of binary image is searched, and passes through RANSAC（RANdom Sample Consensus, at random Unanimity of samples）Algorithm determines the up-and-down boundary of target area, as shown in fig. 7, character the ratio of width to height of target area can be obtained Character block is connected with the up-and-down boundary of target area.

04）：Most width of similar width of connection character block are chosen as character mean breadth, for example：Determined in Fig. 7 Similar width most width be digital " 1 " and two digital " 0 "s width.

05）：By the foreground pixel in the corresponding binary image in target area along character block bearing of trend（The application is real Apply the bearing of trend for numeric string in example）Axle on projected, and projected pixel values are less than default projected pixel quantity Point merged as cut point, and by the cut point of continuous white space, target area is tentatively cut, such as Fig. 8 Shown, wherein vertical line is exactly preliminary cutting zone.

06）：Utilize step 04）The character mean breadth of middle determination judges character block that above-mentioned cutting obtains for adhesion character Block, excessively cutting character block or correct characters block.In the embodiment of the present application, cutting character block was not present, as shown in figure 8, rising The 4-digit number " 5998 " of beginning, the numeral " 93662 " and the 11st numeral " 0362 " started of the 6th beginning are all judged to Break as adhesion character block, five digit number " 1 " and the 15th, hexadecimal numbers " 0 " are all judged as correct characters block.

07）：Wave crest and wave trough is used for adhesion character block, the ratio of trough and crest is calculated, Wave crest and wave trough ratio is set Value is cut, and for the larger character block of two Characters Stuck spot noises, can also be carried out using character mean breadth Cutting is forced, i.e., is using the original position of the larger character block of two Characters Stuck spot noises of distance as the point of character mean breadth Cut point is cut.

08）：Again the average character duration of current frame image is calculated, and the character block after cutting is sent in grader Line character is recognized, obtains current recognition result, as shown in figure 9, the character block in figure in square frame to be identified after cutting.

The schematic flow sheet for the Character segmentation method that Figure 10 provides for another embodiment of the application.In embodiment illustrated in fig. 1 On the basis of, as shown in Figure 10, this method can also include：

S104：Obtain in the wrong cutting zone and cross cutting character block and adhesion character block.

Referring to the description as described in wrong cutting zone in above-described embodiment, cutting character block and adhesion were apparent that The implication of character block, will not be repeated here.

In the embodiment of the present application, as shown in figure 11, step S104 can include：

S1041：Obtain target binaryzation parameter.

In the embodiment of the present application, using the self-adaption binaryzation method after a kind of improvement, concretely comprise the following steps：Carrying out two During value, centered on current pixel, with the average gray value of all pixels in the rectangle of S*S sizes, for current picture The gray value of element is compared, if current pixel gray value is less than average gray value, then it is assumed that be prospect（As the one of character Part）, it is on the contrary then for background colour.So in the embodiment of the present application, target binaryzation parameter is S.

When carrying out binaryzation, it is contemplated that the problems such as binaryzation efficiency（S value it is bigger, it is necessary to the pixel compared more It is many）, generally select as S can divide rectangle using the height of character in target area.In the embodiment of the present application, in order to carry The accuracy of the height of character in high target area, can obtain character height in history recognition result, calculate all history and know The average value of character height in other result, and it regard the average value as target binaryzation parameter.

S1042:Binary conversion treatment is carried out to the current recognition result using the target binaryzation parameter.

After target binaryzation parameter is got, a pixel is as current pixel point in selection target area, to work as Centered on preceding pixel point, choose and a square area is divided by side of the average value of character height in history recognition result, meter The average gray value of all pixels point in the square area is calculated as gray reference value, finally by the gray value of current pixel point It is compared with the gray reference value, if the gray value of current pixel point is more than gray reference value, it is believed that current pixel point is A part for character, i.e. foreground, otherwise it is assumed that current pixel point is background colour.

S1043：Obtain the character confidence width in the current recognition result and history recognition result.

S1044：Character confidence width in the current recognition result and history recognition result, obtains target character Confidence width.

In the embodiment of the present application, target character confidence width can be in current recognition result and history recognition result The average value of character confidence width.

Pass through target character confidence width, it will be appreciated that the desired width that character takes in target area.Alternatively, at this Apply in other embodiments, target character confidence width can also be obtained using priori.

S1045：According to the target character confidence width, obtain after binary conversion treatment described in the current recognition result Cutting character block and adhesion character block are crossed in mistake cutting zone.

After target character confidence width is got, width and target word by character block in comparison error cutting zone Accord with confidence width, it is possible to the type for the character in cutting zone of judging to make mistake, be specially：When character block in wrong cutting zone Width be more than target character confidence width when, the character block is judged as adhesion character block, when character block in wrong cutting zone Width be less than target character confidence width when, judge the character block as cross cutting character block.

S105：Cutting character block and the adhesion character block excessively are cut, target cutting result is obtained.

In the embodiment of the present application, as shown in figure 12, step S105 can include：

S1051：Obtain corresponding history Wave crest and wave trough ratio value when history cutting is carried out to the current recognition result.

When obtaining current recognition result, it will usually preliminary cutting is carried out once to the character block in current frame image, in detail Carefully reference can be made to shown in above-mentioned Fig. 5-Fig. 9, when tentatively being cut, while character block corresponding crest ripple during record preliminary cutting Paddy ratio value, is used as history Wave crest and wave trough ratio value, it usually needs the character block of log history Wave crest and wave trough ratio value is adhesion Character block and excessively cutting character block.

S1052：The history Wave crest and wave trough ratio value according to the target character confidence width adjustment, obtains target cutting Point.

All pixels point on image after binaryzation is projected on the axle in the direction extended along character string, after projection Obtain image as shown in figure 13, it is different by the shape of each character block, obtain that pixel quantity can be represented after all projections Curve.

In the embodiment of the present application, when obtaining the target cut point of adhesion character block, the step 1052 can include：

11）, obtain the first object cutting zone of the adhesion character block, the centre bit of the first object cutting zone Put the distance between the adhesion character block starting point and be equal to the target character confidence width.

First object cutting zone is illustrated by taking Figure 13 as an example below, it is as shown in figure 13, white in pixel in figure Color vertical line is the cut point just cut, by taking adhesion character block as an example, and the character block in figure between A points and C points is adhesion character block, The starting point A of adhesion character block coordinate a is chosen, the distance searched apart from starting point A is w cutting reference point B, cuts reference point Coordinate be a+w, w be target character confidence width.The region of [a+w-c, a+w+c], figure are chosen centered on cutting reference point B Left side width is 2c oblique line filling region in 13, and wherein c is region spacing value, in practical application, region spacing value c can be with Freely set as needed.[a+w-c, a+w+c] region is first object cutting zone.

12）, according to the ratio value of history Wave crest and wave trough search whether there is trough in first object region.

According to formula（1）The ratio f of medium wave peak and trough is calculated,

f=h2/h1 （1）

Wherein, h1 is the height of crest, and h2 is the height of trough.

Judge the magnitude relationship of the ratio value of f and history Wave crest and wave trough, when f is less than the ratio value of history Wave crest and wave trough, really Determine there is trough in first object cutting zone.

13）, when in the absence of trough, turn down the history Wave crest and wave trough ratio value, obtain first refer to Wave crest and wave trough ratio Example value.

14）, using described first refer to Wave crest and wave trough ratio value, search the first object cutting zone and whether there is ripple Paddy, until finding trough.

15）When there is trough, the target that the trough position found is defined as the adhesion character block is cut Point.

21）Obtain the second target cutting zone for crossing cutting character block, the center of the second target cutting zone Distance is equal to the target character confidence width between position and the cutting character block starting point excessively.

Referring to Figure 13, choose with the first of adjacent multiple cutting character blocks excessively the original position M, M point for crossing cutting character block Conduct be a ＇, the reference cut point of the second target cutting zone is N, N points as a ＇+w, choose the second obtained target and cut It is the oblique line filling region that right side width is 2c in Figure 13 to cut region.Alternatively, for crossing cutting character block and adhesion character block, When choosing target cutting zone, region spacing value c can be with identical, can also be different.

22）Search whether there is ripple in the second target cutting zone according to the history Wave crest and wave trough ratio value Paddy；

23) when there is trough, the history Wave crest and wave trough ratio value is heightened, second is obtained and referring to Wave crest and wave trough ratio Value；

24) Wave crest and wave trough ratio value is referred to using described second, searches the second target cutting zone and whether there is ripple Paddy, until can not find trough；

25）It is when in the absence of trough, the trough position found before last time adjustment Wave crest and wave trough ratio value is true It is set to the target cut point for crossing cutting character block.

It is optional when not finding target cut point in crossing cutting character block and adhesion character block through the above way Ground, in the embodiment of the present application, the step 1052 can also include：

31）Judge whether preset times threshold value is more than to the adjustment number of times of Wave crest and wave trough ratio value；

32）When being more than preset times threshold value to the adjustment number of times of Wave crest and wave trough ratio value, by the adhesion character block or Cross cutting character block and be defined as target cut point in the minimum position of the Zhou Shang projection images vegetarian refreshments quantity in character arrangements direction.

In addition, in the embodiment of the present application, when run into have multiple characters in adhesion character block when, with adhesion character block Starting point is original position, respectively the position choosing of selected distance original position one, two and multiple target character confidence width Select and refer to cut point.Simultaneously for multiple continuous cutting character blocks excessively, the starting point of cutting character block is crossed as start bit using first Put, the position of selected distance original position one, two and multiple target character block confidence width is chosen with reference to cutting respectively Point.

S1053：Cutting character block and the adhesion character block excessively are cut in the target cut point, target is obtained Cutting result.

After the target cut point of cutting character block and adhesion character block was got, respectively in respective target cut point Cut position.

In the embodiment of the present application, by taking three history recognition results as an example, referring to the former frame that Figure 14 is current frame image The schematic diagram of the corresponding recognition result of image, Figure 15 is the signal of the corresponding recognition result of front cross frame image of current frame image Figure, Figure 16 is the schematic diagram of the corresponding recognition result of current frame image first three two field picture.Figure 17 is to the current of current frame image The schematic diagram for the target identification result that recognition result is obtained after cutting again.

From Figure 14-Figure 17, this method is when the recognition result to current frame image is cut, it is possible to use current The wrong cutting zone determined in the recognition result of two field picture, auxiliary is cut again, improves the accuracy cut again, and And can be cut during due to cutting again just for wrong cutting zone, it is possible to avoid carrying out again current frame image The blindness of secondary cutting, improves the efficiency of cutting.

The schematic flow sheet for the Character segmentation method that Figure 18 provides for the application another embodiment.As shown in figure 18, exist On the basis of embodiment illustrated in fig. 10, this method can also include：

S106：Judge whether to receive the operation requests that user inputs.

Operation requests include but is not limited to：Identification request is asked, retries request or cancelled to rescan.When user's input Operation requests after, represent user think that this recognition result is undesirable, it is understood that there may be identification mistake.

If it is, carrying out S107, otherwise, S108 is carried out.

S107：The confidence level of the target cutting result is obtained, and reduces the confidence level as the target and cuts knot The objective degrees of confidence of fruit.

When being identified in correspondence target area, it is possible that repeatedly there is same mistake in identification, such as：To numeral When string " 123456789 " is identified, digital " 7 " can be identified as digital " 1 ", so, and for there is same mistake Multiple recognition results confidence level it is all identical, so repeatedly identification after, it is necessary to user confidence level identical it is multiple identification knot One is selected in fruit again as target identification result, it is cumbersome.

By the step, after result is identified every time, if all receiving the operation requests of user's input, it incite somebody to action this Secondary obtained target identification result confidence level reduction, until be finally not connected to user input operation requests, before repeatedly The confidence level of obtained target identification result is by less than the confidence level of the target identification result finally given.Avoid for repeatedly knowledge Not occurring same wrong situation subsequently needs the problem of user is selected in multiple recognition results.

Alternatively, in the embodiment of the present application, the confidence level of target identification result can be reduced according to default decrement, The confidence level of target identification result can also be reduced otherwise.

S108：Export the target cutting result.

The way of output of target cutting result, including but not limited to：On a display screen display, voice broadcast, with short message Form transmission etc..

The description of embodiment of the method more than, it is apparent to those skilled in the art that the application can Realized by the mode of software plus required general hardware platform, naturally it is also possible to by hardware, but in many cases the former It is more preferably embodiment.Understood based on such, the technical scheme of the application substantially makes tribute to prior art in other words The part offered can be embodied in the form of software product, and the computer software product is stored in a storage medium, bag Some instructions are included to cause a computer equipment（Can be personal computer, server, or network equipment etc.）Perform The all or part of step of each embodiment methods described of the application.And foregoing storage medium includes：Read-only storage （ROM）, random access memory（RAM）, magnetic disc or CD etc. are various can be with the medium of store program codes.

The embodiment of the present application also provides the device and equipment for performing above method embodiment, and citing is done below and is introduced.

The structural representation for the Character segmentation device that Figure 19 provides for the embodiment of the application one.As shown in figure 19, the device Including：

Recognition result acquiring unit 1, current recognition result and at least one history for obtaining target area recognize knot Really, the current recognition result is the recognition result of the current frame image of target area, and the history recognition result is target area The recognition result of the history image in domain, the current recognition result and the history recognition result include at least one character Block；

Character block information acquisition unit 2, for obtaining each character in the current recognition result and history recognition result The confidence level of block and the position in image；

Mistake cutting zone acquiring unit 3, for the confidence level according to each character block and position, obtains the current knowledge Wrong cutting zone in other result.

Alternatively, in the embodiment of the present application, mistake cutting zone acquiring unit 3 can include：

Character block different zones obtain subelement, for the confidence information according to each character block and position, obtain institute The character block different zones in current recognition result are stated, wherein, it is described current in same position in the character block different zones Character block in recognition result and history recognition result is all differed or part is differed；

First wrong cutting zone determination subelement, for the character block different zones to be defined as into wrong cutting area Domain.

Second character block acquiring unit, for the confidence information according to character block in the current recognition result, is obtained Confidence level is less than the character block of predetermined threshold value in the current recognition result；

Second mistake cutting determination subelement, the character block region for confidence level to be less than to predetermined threshold value, it is determined that For wrong cutting zone；

Wherein, the second mistake cutting determination subelement can include：

Adjacent area judge module, for judging whether that at least two confidence levels are less than the character block institute of predetermined threshold value It is adjacent in region；

Region merging technique module, when the character block region for being less than predetermined threshold value when at least two confidence levels is adjacent, Merge that position is adjacent and confidence level is less than the character block of predetermined threshold value；

Second mistake cutting determining module, is less than default threshold by the region of the character block after merging and the confidence level not merged The character block region of value is defined as wrong cutting zone.

The current frame image of device acquisition target area and at least recognition result of a frame history image, are then obtained every The confidence level of all character blocks and the position in image in individual recognition result, finally can putting according to each character block Reliability and position, obtain the wrong cutting zone in the recognition result of the current frame image.And then subsequently to present frame figure When the recognition result of picture is cut, it is possible to use the wrong cutting zone determined in the recognition result of current frame image, auxiliary Cut again, improve the accuracy cut again, and can be just for wrong cutting zone, institute during due to cutting again So that the blindness cut again to current frame image can be avoided, the efficiency of cutting is improved.

The structural representation for the Character segmentation device that Figure 20 provides for another embodiment of the application.As shown in figure 20, in figure On the basis of 19 illustrated embodiments, the device can also include：

First character block acquiring unit 4, cutting character block and adhesion word are crossed for obtaining in the wrong cutting zone Accord with block；

Wherein, in the embodiment of the present application, as shown in figure 21, the first character block acquiring unit 4 can include：

Target binaryzation parameter acquiring subelement 41, for obtaining target binaryzation parameter, the target binaryzation parameter Including：Character average height in the recognition result of history image；

Binary conversion treatment subelement 42, for carrying out two to the current recognition result using the target binaryzation parameter Value is handled；

Character confidence width obtains subelement 43, for obtaining the word in the current recognition result and history recognition result Accord with confidence width；

Target character confidence width obtains subelement 44, for according in the current recognition result and history recognition result Character confidence width, obtain target character confidence width；

First character block obtains subelement 45, for obtaining institute after binary conversion treatment according to the target character confidence width State and cross cutting character block and adhesion character block in wrong cutting zone described in current recognition result.

Wherein, target character confidence width obtains subelement and can included：Statistical module, for counting the current identification As a result the relation between relation and higher limit in history recognition result between character confidence width lower limit；Quantity compares mould Block, the size for comparing the first quantity and the second quantity；Character confidence width adjustment module, for being less than the when the first quantity During two quantity, character confidence width higher limit and lower limit in the current recognition result are reduced according to default first correction value, Obtain target character confidence width；It is described current according to the increase of default second correction value when the first quantity is more than the second quantity The higher limit and lower limit of character confidence width in recognition result, obtain target character confidence width.

In the embodiment of the present application, first quantity refers to that character confidence width lower limit is less than the current identification knot The quantity of the history recognition result of character confidence width lower limit in fruit, the second quantity refers to that character confidence width higher limit is more than The quantity of the history recognition result of character confidence width higher limit in the current recognition result

Cutter unit 5, for being cut to cutting character block and the adhesion character block excessively, obtains target cutting result

Wherein, in the embodiment of the present application, as shown in figure 22, cutter unit 5 can include：

Ratio value obtains subelement 51, corresponding history when history cutting is carried out to the current recognition result for obtaining Wave crest and wave trough ratio value；

Target cut point obtains subelement 52, for the history crest ripple according to the target character confidence width adjustment Paddy ratio value, obtains target cut point；

Subelement 53 is cut, for being cut in the target cut point to cutting character block and the adhesion character block excessively Cut, obtain target cutting result.

Alternatively, in the embodiment of the present application, target cut point obtains subelement and can included：

Target cutting zone acquisition module, for obtaining the first object cutting zone of the adhesion character block and obtaining institute Stated the second target cutting zone of cutting character block, the center of the first object cutting zone and the adhesion character Distance is equal to the target character confidence width, center and the mistake of the second target cutting zone between block starting point Distance is equal to the target character confidence width between cutting character block starting point；

Trough searching modul, for according to the history Wave crest and wave trough ratio value in the first object cutting zone or Search whether there is trough in second target cutting zone；

Ratio value adjusting module, for when trough is not present in the first object cutting zone, turning down the history Wave crest and wave trough ratio value, obtains first and refers to Wave crest and wave trough ratio value；When there is trough in the second target cutting zone, The history Wave crest and wave trough ratio value is heightened, second is obtained and refers to Wave crest and wave trough ratio value；

The trough searching modul is additionally operable to refer to Wave crest and wave trough ratio value using described first, searches the first object Cutting zone whether there is trough, until trough is found, or, Wave crest and wave trough ratio value is referred to using described second, institute is searched The second target cutting zone is stated with the presence or absence of trough, until can not find trough；

First object cut point determining module, for when using first with reference to Wave crest and wave trough ratio value find trough when, The trough position found is defined as to the target cut point of the adhesion character block, or, utilize the described second reference When Wave crest and wave trough ratio value is searched less than trough, the trough position that will be found before last time adjustment Wave crest and wave trough ratio value It is defined as the target cut point for crossing cutting character block.

Alternatively, in the embodiment of the present application, target cut point obtains subelement and can also included：

Number of times judge module, for judging whether the adjustment number of times to Wave crest and wave trough ratio value is more than preset times threshold value；

Second target cut point determining module, for being more than preset times threshold when the adjustment number of times to Wave crest and wave trough ratio value It is during value, the adhesion character block or the position for crossing cutting character block projection image's vegetarian refreshments quantity minimum on character arrangements direction is true It is set to target cut point.

The device is when the recognition result to current frame image is cut, it is possible to use the recognition result of current frame image The wrong cutting zone of middle determination, auxiliary is cut again, improves the accuracy cut again, and during due to cutting again It can be cut just for wrong cutting zone, it is possible to avoid the blindness cut again to current frame image, Improve the efficiency of cutting.

The structural representation for the Character segmentation device that Figure 23 provides for the another embodiment of the application.As shown in figure 23, in figure On the basis of 20 illustrated embodiments, the device can also include：

Operation requests receiving unit 6, the operation requests for receiving user's input, the operation requests include：It is secondary to sweep Request is retouched, request is retried or cancels identification request；

Judging unit 7 is asked, for judging whether to receive the operation requests that user inputs；

Objective degrees of confidence determining unit 8, for when receiving the operation requests, obtaining the target cutting result Confidence level, and the confidence level is reduced as the objective degrees of confidence of the target cutting result；

Output unit 9, for when not receiving the operation requests, exporting the target cutting result.

The device, after result is identified every time, if all receiving the operation requests of user's input, this is obtained The target identification result arrived confidence level reduction, until be finally not connected to user input operation requests, before repeatedly obtain Target identification result confidence level by less than the confidence level of the target identification result finally given.Avoid equal for repeatedly identification Same wrong situation occur subsequently needs the problem of user is selected in multiple recognition results.

Suggestion device embodiment is scanned corresponding to above- mentioned information card, the disclosure additionally provides the implementation of a kind of electronic equipment Example.

Figure 24 is the structural representation for a kind of electronic equipment that the embodiment of the present application is provided, referring to Figure 24, the electronic equipment The Character segmentation method for implementing to provide in above-described embodiment is provided.Specifically：

Electronic equipment 500 can include radio circuit 110, include one or more computer-readable recording mediums Memory 120, input block 130, display unit 140, sensor 150, voicefrequency circuit 160, WiFi module 170, include The parts such as the processor 180 and power supply 190 of one or more than one processing core.It will be understood by those skilled in the art that The electronic devices structure shown in Figure 24 does not constitute the restriction to electronic equipment, can include than illustrating more or less portions Part, either combines some parts or different parts arrangement.Wherein：

Radio circuit 110 can be used for receive and send messages or communication process in, the reception and transmission of signal.Especially, radio frequency electrical Road 110 will base station downlink information receive after, transfer to one or more than one processor 180 handled；In addition, will be related to up Data be sent to base station.Usually as communication unit radio circuit include but is not limited to antenna, at least one amplifier, Tuner, one or more oscillators, subscriber identity module（SIM）Card, transceiver, coupler, LNA（Low Noise Amplifier, low-noise amplifier）, duplexer etc..In addition, radio circuit 110 can also pass through radio communication and network and its His equipment communication.The radio communication can use any communication standard or agreement, including but not limited to GSM (Global System of Mobile communication, global system for mobile communications), GPRS (General Packet Radio Service, general packet radio service), CDMA (Code Division Multiple Access, CDMA), WCDMA (Wideband Code Division Multiple Access, WCDMA), LTE (Long Term Evolution, Long Term Evolution), Email, SMS (Short Messaging Service, Short Message Service) etc..Memory 120 can be used for storage software program and module, processor 180 by operation be stored in memory 120 software program and Module, so as to perform various function application and data processing.Memory 120 can mainly include storing program area and data storage Area, wherein, the application program that storing program area can be needed for storage program area, at least one function（Such as sound-playing function, Image player function etc.）Deng；Storage data field can be stored uses created data according to electronic equipment 500（Such as audio Data, phone directory etc.）Deng.In addition, memory 120 can include high-speed random access memory, it can also include non-volatile Memory, for example, at least one disk memory, flush memory device or other volatile solid-state parts.Correspondingly, store Device 120 can also include Memory Controller, to provide the access of processor 180 and input block 130 to memory 120.

Input block 130 can be used for the numeral or character information for receiving input, and generation to be set with user and function The relevant keyboard of control, mouse, action bars, optics or the input of trace ball signal.Specifically, input block 130 may include to touch Sensitive surfaces 131 and other input equipments 132.Touch sensitive surface 131, also referred to as touch display screen or Trackpad, collect and use Touch operation of the family on or near it（Such as user is using any suitable objects such as finger, stylus or annex in touch-sensitive table Operation on face 131 or near touch sensitive surface 131）, and corresponding attachment means are driven according to formula set in advance.It is optional , touch sensitive surface 131 may include both touch detecting apparatus and touch controller.Wherein, touch detecting apparatus detection is used The touch orientation at family, and the signal that touch operation is brought is detected, transmit a signal to touch controller；Touch controller is from touch Touch information is received in detection means, and is converted into contact coordinate, then gives processor 180, and can reception processing device 180 The order sent simultaneously is performed.Furthermore, it is possible to using polytypes such as resistance-type, condenser type, infrared ray and surface acoustic waves Realize touch sensitive surface 131.Except touch sensitive surface 131, input block 130 can also include other input equipments 132.Specifically, Other input equipments 132 can include but is not limited to physical keyboard, function key（Such as volume control button, switch key etc.）、 One or more in trace ball, mouse, action bars etc..

Display unit 140 can be used for the information and electronic equipment for showing the information inputted by user or being supplied to user 500 various graphical user interface, these graphical user interface can by figure, text, icon, video and its any combination Lai Constitute.Display unit 140 may include display panel 141, optionally, can using LCD (Liquid Crystal Display, Liquid crystal display), the form such as OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) configure display Panel 141.Further, touch sensitive surface 131 can cover display panel 141, when touch sensitive surface 131 is detected on or near it Touch operation after, processor 180 is sent to determine the type of touch event, with preprocessor 180 according to touch event Type provides corresponding visual output on display panel 141.Although in fig. 24, touch sensitive surface 131 is with display panel 141 Realized as two independent parts input and input function, but in some embodiments it is possible to by touch sensitive surface 131 with Display panel 141 is integrated and realizes input and output function.

Electronic equipment 500 may also include at least one sensor 150, such as optical sensor, motion sensor and other Sensor.Specifically, optical sensor may include ambient light sensor and proximity transducer, wherein, ambient light sensor can basis The light and shade of ambient light adjusts the brightness of display panel 141, proximity transducer can when electronic equipment 500 is moved in one's ear, Close display panel 141 and/or backlight.As one kind of motion sensor, gravity accelerometer can detect all directions On（Generally three axles）The size of acceleration, can detect that size and the direction of gravity when static, available for identification mobile phone posture Application（Such as horizontal/vertical screen switching, dependent game, magnetometer pose calibrating）, Vibration identification correlation function（Such as pedometer, strike Hit）Deng;Gyroscope, barometer, hygrometer, thermometer, infrared ray sensor for can also configure as electronic equipment 500 etc. other Sensor, will not be repeated here.

Voicefrequency circuit 160, loudspeaker 161, microphone 162 can provide the COBBAIF between user and electronic equipment 500. Electric signal after the voice data received conversion can be transferred to loudspeaker 161, is changed by loudspeaker 161 by voicefrequency circuit 160 Exported for voice signal；On the other hand, the voice signal of collection is converted to electric signal by microphone 162, is connect by voicefrequency circuit 160 Be converted to voice data after receipts, then after voice data output processor 180 is handled, it is such as another to be sent to through RF circuits 110 One electronic equipment, or voice data is exported to memory 120 so as to further processing.Voicefrequency circuit 160 is also possible that Earphone jack, to provide the communication of peripheral hardware earphone and electronic equipment 500.

In order to realize radio communication, WiFi module 170 can be configured with the electronic equipment.WiFi belongs to short-distance wireless Transmission technology, electronic equipment 500 can help user to send and receive e-mail, browse webpage and access stream by WiFi module 170 Formula media etc., it has provided the user wireless broadband internet and accessed.Although Figure 24 shows WiFi module 170, can With understanding, it is simultaneously not belonging to must be configured into for electronic equipment 500, can not change the essence of invention as needed completely In the range of and omit.

Processor 180 is the control centre of electronic equipment 500, utilizes each of various interfaces and connection whole mobile phone Part, by operation or performs and is stored in software program and/or module in memory 120, and calls and be stored in memory Data in 120, perform the various functions and processing data of electronic equipment 500, so as to carry out integral monitoring to mobile phone.It is optional , processor 180 may include one or more processing cores；It is preferred that, processor 180 can integrated application processor and modulatedemodulate Processor is adjusted, wherein, application processor mainly handles operating system, user interface and application program etc., modem processor Main processing radio communication.It is understood that above-mentioned modem processor can not also be integrated into processor 180.

Electronic equipment 500 also includes the power supply 190 powered to all parts（Such as battery）, it is preferred that power supply can lead to Cross power-supply management system and processor 180 be logically contiguous, thus by power-supply management system realize management charging, electric discharge and The functions such as power managed.Power supply 190 can also include one or more direct current or AC power, recharging system, electricity The random component such as source failure detector circuit, power supply changeover device or inverter, power supply status indicator.

Although not shown, electronic equipment 500 can also include camera, bluetooth module etc., will not be repeated here.Specifically exist In the present embodiment, the display unit of electronic equipment is touch-screen display, and electronic equipment also includes memory, the memory Also include one or more than one program, and be configured to by one or more than one processor 180 performs described one Individual or more than one program bag, which contains, to be used to perform the Character segmentation method that any illustrated embodiments of Fig. 1-Figure 18 are provided.

As on the other hand, the application another embodiment additionally provides a kind of computer-readable recording medium, the computer Readable storage medium storing program for executing can be the computer-readable recording medium included in the memory in above-described embodiment；Can also be single Solely exist, without the computer-readable recording medium in supplying terminal.The computer-readable recording medium storage have one or More than one program of person, one or more than one program by one or more than one processor be used for performing Fig. 1- The Character segmentation method that any illustrated embodiments of Figure 18 are provided.

It is understood that the application can be used in numerous general or special purpose computing system environments or configuration.For example：It is individual People's computer, server computer, handheld device or portable set, laptop device, multicomputer system, based on microprocessor The system of device, set top box, programmable consumer-elcetronics devices, network PC, minicom, mainframe computer including to take up an official post DCE of what system or equipment etc..

The application can be described in the general context of computer executable instructions, such as program Module.Usually, program module includes performing particular task or realizes routine, program, object, the group of particular abstract data type Part, data structure etc..The application can also be put into practice in a distributed computing environment, in these DCEs, by Remote processing devices connected by communication network perform task.In a distributed computing environment, program module can be with Positioned at including in the local and remote computer-readable storage medium including storage device.

It should be noted that herein, such as first and second or the like relational terms are used merely to a reality Body or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or deposited between operating In any this actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant are intended to Nonexcludability is included, so that process, method, article or equipment including a series of key elements not only will including those Element, but also other key elements including being not expressly set out, or also include being this process, method, article or equipment Intrinsic key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that Also there is other identical element in process, method, article or equipment including the key element.

Described above is only the embodiment of the application, it is noted that for the ordinary skill people of the art For member, on the premise of the application principle is not departed from, some improvements and modifications can also be made, these improvements and modifications also should It is considered as the protection domain of the application.

Claims

1. a kind of Character segmentation method, it is characterised in that methods described includes：

The current recognition result and at least one history recognition result of target area are obtained, the current recognition result is target area The recognition result of the current frame image in domain, the history recognition result is the recognition result of the history image of target area, described Current recognition result and the history recognition result include at least one character block；

Obtain the confidence level of each character block and the position in image in the current recognition result and history recognition result；

According to the confidence level of each character block and position, the wrong cutting zone in the current recognition result is obtained.

2. according to the method described in claim 1, it is characterised in that methods described also includes：

Obtain in the wrong cutting zone and cross cutting character block and adhesion character block；

Cutting character block and the adhesion character block excessively are cut, target cutting result is obtained.

3. method according to claim 2, it is characterised in that cross cutting word in the acquisition wrong cutting zone Block and adhesion character block are accorded with, including：

Target binaryzation parameter is obtained, the target binaryzation parameter includes：Character in the recognition result of history image is averaged Highly；

Binary conversion treatment is carried out to the current recognition result using the target binaryzation parameter；

Obtain the character confidence width in the current recognition result and history recognition result；

Character confidence width in the current recognition result and history recognition result, obtains target character confidence width；

Wrong cutting area described in the current recognition result after binary conversion treatment is obtained according to the target character confidence width Cutting character block and adhesion character block are crossed in domain.

4. method according to claim 3, it is characterised in that described to be tied according to the current recognition result and history identification Character confidence width in fruit, obtains target character confidence width, including：

Count relation and the upper limit between character confidence width lower limit in the current recognition result and history recognition result Relation between value；

When the first quantity is less than the second quantity, character confidence in the current recognition result is reduced according to default first correction value Width higher limit and lower limit, obtain target character confidence width；

When the first quantity is more than the second quantity, increases the character in the current recognition result according to default second correction value and put Believe the higher limit and lower limit of width, obtain target character confidence width；

Wherein, first quantity refers to that character confidence width lower limit is less than character confidence width in the current recognition result The quantity of the history recognition result of lower limit, the second quantity refers to that character confidence width higher limit is more than the current recognition result The quantity of the history recognition result of middle character confidence width higher limit.

5. method according to claim 3, it is characterised in that described to enter to cutting character block and the adhesion character block excessively Row cutting, obtains target cutting result, including：

Obtain corresponding history Wave crest and wave trough ratio value when history cutting is carried out to the current recognition result；

History Wave crest and wave trough ratio value, obtains target cut point according to the target character confidence width adjustment；

Cutting character block and the adhesion character block excessively are cut in the target cut point, target cutting result is obtained.

6. method according to claim 5, it is characterised in that described according to the target character confidence width adjustment History Wave crest and wave trough ratio value, obtains target cut point, including：

Obtain the first object cutting zone of the adhesion character block, the center of the first object cutting zone with it is described Distance is equal to the target character confidence width between adhesion character block starting point；

Search whether there is trough in the first object cutting zone according to the history Wave crest and wave trough ratio value；

When in the absence of trough, the history Wave crest and wave trough ratio value is turned down, first is obtained and refers to Wave crest and wave trough ratio value；

Wave crest and wave trough ratio value is referred to using described first, the first object cutting zone is searched and whether there is trough, until Find trough；

When there is trough, the trough position found is defined as to the target cut point of the adhesion character block.

7. method according to claim 6, it is characterised in that described according to the target character confidence width adjustment History Wave crest and wave trough ratio value, obtains target cut point, in addition to：

Judge whether preset times threshold value is more than to the adjustment number of times of Wave crest and wave trough ratio value；

When the adjustment number of times to Wave crest and wave trough ratio value is more than preset times threshold value, by the adhesion character block or mistake cutting word Symbol block is defined as target cut point in the minimum position of the Zhou Shang projection images vegetarian refreshments quantity in character arrangements direction.

8. method according to claim 5, it is characterised in that described according to the target character confidence width adjustment History Wave crest and wave trough ratio value, obtains target cut point, including：

Obtain the second target cutting zone for crossing cutting character block, the center of the second target cutting zone and institute Distance was equal to the target character confidence width between stating cutting character block starting point；

Search whether there is trough in the second target cutting zone according to the history Wave crest and wave trough ratio value；

When there is trough, the history Wave crest and wave trough ratio value is heightened, second is obtained and referring to Wave crest and wave trough ratio value；

Wave crest and wave trough ratio value is referred to using described second, the second target cutting zone is searched and whether there is trough, until It can not find trough；

When in the absence of trough, the trough position found before last time adjustment Wave crest and wave trough ratio value is defined as described Cross the target cut point of cutting character block.

9. method according to claim 8, it is characterised in that described according to the target character confidence width adjustment History Wave crest and wave trough ratio value, obtains target cut point, in addition to：

10. method according to claim 2, it is characterised in that methods described also includes：

Judge whether to receive the operation requests that user inputs, the operation requests include：Rescan request, retry request or Cancel identification request；

When receiving the operation requests, the confidence level of the target cutting result is obtained, and reduces the confidence level conduct The objective degrees of confidence of the target cutting result；

When not receiving the operation requests, the target cutting result is exported.

11. the method according to claim any one of 1-10, it is characterised in that the confidence level of each character block of basis And position, the wrong cutting zone in the current recognition result is obtained, including：

According to the confidence information of each character block and position, the character block different zones in the current recognition result are obtained, Wherein, the character block in the character block different zones in same position in the current recognition result and history recognition result is complete Portion is differed or part is differed；

The character block different zones are defined as wrong cutting zone.

12. the method according to claim any one of 1-10, it is characterised in that the confidence level of each character block of basis Information and position, obtain the wrong cutting zone in the current recognition result, including：

According to the confidence information of character block in the current recognition result, obtain confidence level in the current recognition result and be less than The character block of predetermined threshold value；

Confidence level is less than to the character block region of predetermined threshold value, is defined as wrong cutting zone；

The character block different zones are defined as wrong cutting zone.

13. method according to claim 12, it is characterised in that the character block institute that confidence level is less than to predetermined threshold value It is defined as wrong cutting zone in region, including：

Judge whether that at least two confidence levels are adjacent less than the character block region of predetermined threshold value；

When the character block region that at least two confidence levels are less than predetermined threshold value is adjacent, merging position is adjacent and confidence level is small In the character block of predetermined threshold value；

The character block region that the region of character block after merging and the confidence level not merged are less than predetermined threshold value is defined as Mistake cutting zone.

14. a kind of Character segmentation device, it is characterised in that including：

Recognition result acquiring unit, current recognition result and at least one history recognition result, institute for obtaining target area The recognition result that current recognition result is the current frame image of target area is stated, the history recognition result is going through for target area The recognition result of history image, the current recognition result and the history recognition result include at least one character block；

Character block information acquisition unit, for obtaining putting for each character block in the current recognition result and history recognition result Reliability and the position in image；

Mistake cutting zone acquiring unit, for the confidence level according to each character block and position, obtains the current identification knot Wrong cutting zone in fruit.

15. device according to claim 14, it is characterised in that described device also includes：

First character block acquiring unit, cutting character block and adhesion character block are crossed for obtaining in the wrong cutting zone；

Cutter unit, for being cut to cutting character block and the adhesion character block excessively, obtains target cutting result.

16. device according to claim 15, it is characterised in that the first character block acquiring unit includes：

Target binaryzation parameter acquiring subelement, for obtaining target binaryzation parameter, the target binaryzation parameter includes：Go through Character average height in the recognition result of history image；

Binary conversion treatment subelement, for being carried out using the target binaryzation parameter to the current recognition result at binaryzation Reason；

Character confidence width obtains subelement, for obtaining the character confidence in the current recognition result and history recognition result Width；

Target character confidence width obtains subelement, for the character in the current recognition result and history recognition result Confidence width, obtains target character confidence width；

First character block obtains subelement, for obtaining described current after binary conversion treatment according to the target character confidence width Described in recognition result cutting character block and adhesion character block are crossed in wrong cutting zone.

17. device according to claim 16, it is characterised in that the target character confidence width obtains subelement bag Include：

Statistical module, for counting in the current recognition result and history recognition result between character confidence width lower limit Relation between relation and higher limit；

Quantity comparison module, the size for comparing the first quantity and the second quantity；

Character confidence width adjustment module, for when the first quantity is less than the second quantity, reducing according to default first correction value Character confidence width higher limit and lower limit in the current recognition result, obtain target character confidence width；When the first quantity During more than the second quantity, increase the higher limit of the character confidence width in the current recognition result according to default second correction value And lower limit, obtain target character confidence width；Wherein, first quantity refers to character confidence width lower limit less than described The quantity of the history recognition result of character confidence width lower limit in current recognition result, the second quantity refers to character confidence width Higher limit is more than the quantity of the history recognition result of character confidence width higher limit in the current recognition result.

18. device according to claim 16, it is characterised in that the cutter unit includes：

Ratio value obtains subelement, corresponding history crest ripple when history cutting is carried out to the current recognition result for obtaining Paddy ratio value；

Target cut point obtains subelement, for the history Wave crest and wave trough ratio according to the target character confidence width adjustment Value, obtains target cut point；

Subelement is cut, for being cut in the target cut point to cutting character block and the adhesion character block excessively, is obtained To target cutting result.

19. device according to claim 18, it is characterised in that the target cut point, which obtains subelement, to be included：

Target cutting zone acquisition module, for obtaining the first object cutting zone of the adhesion character block and obtaining the mistake Risen with the adhesion character block second target cutting zone of cutting character block, the center of the first object cutting zone Distance is equal to the target character confidence width between point, and cutting is crossed in the center of the second target cutting zone with described Distance is equal to the target character confidence width between character block starting point；

Trough searching modul, for according to the history Wave crest and wave trough ratio value in the first object cutting zone or second Search whether there is trough in target cutting zone；

Ratio value adjusting module, for when trough is not present in the first object cutting zone, turning down the history crest Trough ratio value, obtains first and refers to Wave crest and wave trough ratio value；When there is trough in the second target cutting zone, heighten The history Wave crest and wave trough ratio value, obtains second and refers to Wave crest and wave trough ratio value；

The trough searching modul is additionally operable to refer to Wave crest and wave trough ratio value using described first, searches the first object cutting Region whether there is trough, until find trough, or, Wave crest and wave trough ratio value is referred to using described second, described the is searched Two target cutting zones whether there is trough, until can not find trough；

First object cut point determining module, for when finding trough with reference to Wave crest and wave trough ratio value using first, will look into The trough position found is defined as the target cut point of the adhesion character block, or, refer to crest using described second When trough ratio value is searched less than trough, the trough position found before last time adjustment Wave crest and wave trough ratio value is determined For the target cut point for crossing cutting character block.

20. device according to claim 19, it is characterised in that the target cut point, which obtains subelement, also to be included：

Second target cut point determining module, for being more than preset times threshold value when the adjustment number of times to Wave crest and wave trough ratio value When, the adhesion character block or cutting character block minimum position of projection image's vegetarian refreshments quantity on character arrangements direction excessively are determined For target cut point.

21. device according to claim 15, it is characterised in that described device also includes：

Operation requests receiving unit, the operation requests for receiving user's input, the operation requests include：Rescan please Ask, retry request or cancel identification request；

Judging unit is asked, for judging whether to receive the operation requests that user inputs；

Objective degrees of confidence determining unit, for when receiving the operation requests, obtaining the confidence of the target cutting result Degree, and the confidence level is reduced as the objective degrees of confidence of the target cutting result；

Output unit, for when not receiving the operation requests, exporting the target cutting result.

22. the device according to claim any one of 14-21, it is characterised in that the wrong cutting zone acquiring unit Including：

Character block different zones obtain subelement, for the confidence information according to each character block and position, obtain described work as Character block different zones in preceding recognition result, wherein, the current identification in same position in the character block different zones As a result all differed with the character block in history recognition result or part is differed；

First wrong cutting zone determination subelement, for the character block different zones to be defined as into wrong cutting zone.

23. the device according to claim any one of 14-21, it is characterised in that the wrong cutting zone acquiring unit Including：

Second character block acquiring unit, for the confidence information according to character block in the current recognition result, obtains described Confidence level is less than the character block of predetermined threshold value in current recognition result；

Second mistake cutting determination subelement, the character block region for confidence level to be less than to predetermined threshold value is defined as mistake Miss cutting zone；

24. device according to claim 23, it is characterised in that the second mistake cutting determination subelement includes：

Adjacent area judge module, for judging whether that at least two confidence levels are less than the character block location of predetermined threshold value Domain is adjacent；

Region merging technique module, when the character block region for being less than predetermined threshold value when at least two confidence levels is adjacent, merges Position is adjacent and confidence level is less than the character block of predetermined threshold value；

Second mistake cutting determining module, is less than predetermined threshold value by the region of the character block after merging and the confidence level not merged Character block region is defined as wrong cutting zone.

25. a kind of electronic equipment, it is characterised in that including：Including memory, and one or more than one instruction, its In one or more than one instruction be stored in memory, and be configured to by one or more than one computing device institute State one or more than one instruction includes the instruction for being used to be operated below：