CN101339617A - Mobile phones photographing and translation device - Google Patents

Abstract

The invention discloses a cellphone photographing and translating device which comprises a photographing unit, a user interface unit, a character feature image identifying engine (OCR engine) unit, a translation engine unit and an image preprocessing unit. The cellphone photographing and translating device of the invention ensures that the loading of the character feature image identification on a handset device, the obtaining and the process of printed characters which are photographed by a photographing device, the identification of the character correction and the displayed human-computer interaction guideline, the identification to the digital image identification characters, the identification character translation and the identification result storage are all very convenient, in summary, the device can translate the required information in time, has high-efficient input function and provide a new developing opportunity for the handset functions.

Description

Mobile phones photographing and translation device

Technical field

The present invention relates to digital image processing field, area of pattern recognition, and embedded device field, particularly a kind of handheld device interpreting equipment of taking pictures.

Background technology

Along with the development of handheld device with popularize, mobile phone has more and more become the electronic equipment that people's daily life is carried, and the data that how will need to translate is efficiently imported, and data is in time translated, for cell-phone function provides new opportunity to develop.

Summary of the invention

The purpose of this invention is to provide a kind of photographing and translation device.

The objective of the invention is to be achieved through the following technical solutions:

A kind of photographing and translation device comprises:

Character feature image recognition (OCR) engine unit is used for the character conversion of written historical materials digital picture is become character standard ISN.

A translation engine unit is used to translate the written historical materials that OCR identifies;

At least a shooting unit that has 1,300,000 pixels and have macro function is used to take the digital picture of obtaining business card;

An image pretreatment unit, the picture format that the image transitions that is used for taking becomes the translation engine unit to discern carries out the binaryzation compression with image, is used to promote recognition speed;

A user interface elements is used for and user interactions, and the guide user uses this function interface.

Wherein, this OCR engine unit comprises:

The engine library unit is used for the store character proper vector;

Engine is provided with the unit, is used to be provided with operational mode or digital picture parameter;

The engine start unit is used to distribute running space, the engine storehouse is loaded into internal memory, and the startup engine is an executable state;

Engine printed page analysis unit is used to divide the space of a whole page, cuts apart the translation of words zone, and recognizable character is confined with connected region;

The engine recognition unit is used to discern the digital picture that is communicated with in the district, extracts feature according to the digital picture visual pattern and discerns the output character ISN; And

The engine-off unit, engine is closed in the releasing memory space.

This engine library unit comprises:

The translation library unit is used to deposit the speech and the table of comparisons of translating content; And

The translation interface unit is used to provide input translation word, obtains the interface of translation result.

Image pretreatment unit: obtain digital picture from the camera installation unit, this image resolution ratio is more than 1280 * 960, decoding transfers 16 rgb images to the Jpg image through hardware, is converted into 8 gray scale bmp format-patterns from 16 rgb images, then image is carried out binary conversion treatment;

Bianry image is meant the image of only deceiving (gray-scale value is 0) white (gray-scale value is 1) two-value in the entire image picture, does not present the variation of gray scale on them.In Digital Image Processing, bianry image occupies important status.This is because in the image processing system of practicality, requires speed height, the cost of processing low, and it is too big that the shading image that contains much information is handled cost, is not very wise move.And the notion of the image after the binaryzation in can enough geometry analyze and feature description, makes things convenient for manyly compared with gray level image.Thereby binary Images Processing become at present in the Flame Image Process one independently, important branch and obtain to use widely.

If remarked pixel is in that (binary conversion treatment is shown in the following formula for i, the j) gray-scale value of position.

$f(i,j)=\left\{\begin{array}{cc}1& f(i,j)\≥t\\ 0& f(i,j)\≤t\end{array}\right.$

Here t is binary-state threshold (Threshold).The 8-neighborhood (8-Neighbor) of pixel is removed outside d-neighbour's the pixel, and 4 pixels on the remaining diagonal line are called that (symbol is: the i-neighbour for i, non-direct neighborhood j).The linking number of certain pixel can be with the 8-neighborhood value f (x of this pixel ₀) ... f (x ₇) calculate.

$N_c=\underset{k=\mathrm{0,2,4,6}}{\mathrm{\Σ}}[(1-f\left(x_k\right))-(1-f\left(x_k\right))(1-f\left(x_{k+1}\right))(1-f\left(x_{k+2}\right))]$

Work as x _k=x ₈The time, make x ₈=x ₀

For the 8-neighborhood of a pixel the value that might exist, calculate according to following formula, its linking number is always got the value between the 0-4.In the automatic identifying of literal, need carry out refinement to bianry image, can also significantly reduce redundant information.

The binary image refined image

Advantage of the present invention is: the data to required translation has efficient input, to data timely translation can be arranged, for cell-phone function provides new opportunity to develop.

User interface elements can comprise:

Preview interface, the speech interface is selected at the printed page analysis interface, the translation interface.

Description of drawings

Fig. 1 is the structured flowchart of the embodiment of the invention;

Fig. 2 is the schematic flow sheet of the embodiment of the invention;

Fig. 3 is the engine schematic flow sheet of the embodiment of the invention.

Fig. 4-1～Fig. 4-4 shows the operating process synoptic diagram of the embodiment of the invention.

Embodiment

Provide better embodiment of the present invention according to Fig. 1～Fig. 4-4 below, and described in detail,, rather than be used for limiting scope of the present invention so that those skilled in the art is easier to understand architectural feature of the present invention and function characteristics.

See also shown in Figure 1, mobile phone with the interpretative function of taking pictures, comprise image pretreatment component 1, user interface 2, image recognition engine 3, dictionary engine 4 and camera installation 5, wherein: user interface 2 comprises preview interface 21, participle interface 22, correction interface 23 and the translation interface 24 that is set up in parallel; The image recognition engine 3 comprises that the engine that is set up in parallel is provided with 31, data printed page analysis engine 32, character recognition engine 33 and engine-off 34; Dictionary engine 4 comprises dictionary library 41 and translation interface 42; Camera installation 5 comprises that camera withdraws from unit 51, camera take pictures unit 52, camera adjustments unit 53 and camera preview unit 54.

See also photograph flow process 100 shown in Figure 2,

S ₁₀₀₁, initialization one OCR engine initialization, camera initialization and the initialization of dictionary engine;

S ₁₀₀₂, data preview one camera preview, preview regulate;

S ₁₀₀₃, data takes that a camera is taken and image transitions;

S ₁₀₀₄, printed page analysis one word piecemeal;

S ₁₀₀₅, select speech identification one to select word, identified word and word correction;

S ₁₀₀₆, translation result one calls dictionary engine, display result;

In each step of above-mentioned flow process, at S ₁₀₀₆In the translation result, if judge that also need proceed just rebound carries out S ₁₀₀₅, select speech identification; If deterministic process finishes, then skip to S ₁₀₀₇, withdraw from mobile phone shooting translation and be in exit status.

At S ₁₀₀₅Select in the speech identification step, judge and select speech to carry out, just carry out S ₁₀₀₇, withdrawing from, mobile phone is taken translation and is in holding state.

At S ₁₀₀₂, in the data preview step, the judgement data need not taken, and just skips to S ₁₀₀₇, withdrawing from, mobile phone is taken translation and is in holding state.

At S ₁₀₀₁In the initialization step, differentiate to take to translate and need not carry out, just carry out S ₁₀₀₇, withdrawing from, mobile phone is taken translation and is in holding state.

At S ₁₀₀₄In the printed page analysis step, or S ₁₀₀₃The undesirable need of discovery image are taken again in the data shooting step, then rebound S ₁₀₀₂, carry out the data preview step again.

See also Fig. 3, it shows the synoptic diagram of engine flow process 200 of the present invention, as shown in the figure,

S ₂₀₀₁, start;

S ₂₀₀₂, start the business card image recognition engine;

S ₂₀₀₃, the business card image attribute is set.

S ₂₀₀₄, business card image is handled

S ₂₀₀₅, the output of business card word

S ₂₀₀₆, also have word output? if word output is arranged, then S is carried out in rebound ₂₀₀₄Otherwise, carry out S ₂₀₀₇,

S ₂₀₀₇, close business card image

S ₂₀₀₈, EOP (end of program).

See also Fig. 4-1～Fig. 4-4

It has provided user's operating process of the present invention: promptly 1 ', shooting data, the preview data image;

2 ', click " identification ", after several seconds, the column picture frame appears on the business card; For example, selected " precdent " column by keyboard or stylus; 3 ', eject English word by identification, precedent if discern wrong can the modification again, clicks translation.4 ', show the translator of Chinese of this word, repeat " continuation "+" translation " operation after, can finish translation to whole section each word of data.

Claims (4)

1, a kind of mobile phones photographing and translation device comprises:

An OCR engine unit is used for the character conversion of written historical materials digital picture is become character standard ISN;

A translation engine unit is used to translate OCR and identifies written historical materials;

At least a shooting unit that has 1,300,000 pixels and macro function is arranged is used to take the digital picture of obtaining business card;

An image pretreatment unit, the picture format that the image transitions that is used for taking becomes the translation engine unit to discern carries out the binaryzation compression with image, to promote recognition speed; And a user interface elements, being used for and user interactions, the guide user uses this interface.

2, mobile phones photographing and translation device according to claim 1 is characterized in that, described ORC engine unit comprises: the engine library unit is used for the store character proper vector; Engine is provided with the unit, is used to establish operational mode or digital picture parameter; The engine start unit is used to distribute running space, and the engine storehouse is loaded into internal memory, and starting this engine start unit is executable state; Engine face version analytic unit is used to divide the space of a whole page, cuts apart the translation of words zone, and recognizable character is confined with connected region; The engine recognition unit is used to discern the digital picture in the connected region, discerns the output character ISN according to the visual pattern extraction feature of digital picture; The engine-off unit, above-mentioned each engine unit is closed in the releasing memory space.

3, mobile phones photographing and translation device according to claim 2 is characterized in that, described engine library unit comprises:

The translation library unit is used to deposit the speech and the table of comparisons of translating content;

The translation interface unit is used to provide input translation word, is the interface that obtains translation result.

4, mobile phones photographing and translation device according to claim 1 is characterized in that, described user interface elements comprises: preview interface, printed page analysis interface, select the speech interface and the translation interface.

Images