JP2007011738A - Character processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a character processor capable of preparing the same character as a character appearing in an objective field. <P>SOLUTION: A γ correction circuit 52 stores six γ correction characteristics for correcting the reproducibility of a field image captured by an image sensor 48 by mutually different forms. A camera CPU 60 selects three γ correction characteristics out of six γ correction characteristics on the basis of the contrast of the field image captured by the image sensor 48. The characters are recognized by the camera CPU 60 from the field image having reproducibility corrected by three γ correction characteristics respectively selected by the camera CPU 60. Three characters recognized by the camera CPU 60 are displayed on an LCD 36. Character selecting operation is received by the camera CPU 60. One character following the character selecting operation out of three characters displayed on the LCD 36 is stored in a flash memory 42. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a character processing device, and more particularly to, for example, a character processing device that recognizes characters appearing in a photographed scene.

An example of a conventional device of this type is disclosed in Patent Document 1. This prior art is intended to perform density correction with optimized contrast on the read image data. Thereby, the recognition rate of OCR can be improved.
Patent Publication 2003-51944 [H04N 1 / 407,1 / 46,1 / 60, G06T 1 / 00,5 / 00]

  However, in the prior art, since it is assumed that the device creates the same character as the character of the subject by itself, strict density correction is required.

  Therefore, a main object of the present invention is to provide a character processing apparatus capable of creating the same character as the character appearing in the scene by simple processing.

  The character processing device (28) according to the first aspect of the present invention includes an image pickup means (48) for capturing a scene, and M pieces (M: M) for correcting the reproducibility of the scene image captured by the image pickup means in different modes. A holding means (52b to 52g) for holding correction characteristics of an integer of 3 or more, and N out of M correction characteristics (N: 2 or more M) based on the contrast of the object scene image captured by the imaging means A selection means (S5) for selecting a correction characteristic of less than (integer) pieces, and a recognition means for recognizing a character from the object scene image having reproducibility corrected by each of the N correction characteristics selected by the selection means ( S9), display means (S15) for displaying N characters recognized by the recognition means, accepting means (S17) for accepting a character selection operation in relation to display processing of the display means, and N displayed by the display means One character that follows the character selection operation is preserved. Comprising storage means (S19) for.

  The object scene is captured by the imaging means. The holding unit holds M correction characteristics for correcting the reproducibility of the object scene image captured by the imaging unit in different modes. The selection unit selects N correction characteristics from the M correction characteristics based on the contrast of the object scene image captured by the imaging unit.

  The character is recognized by the recognition unit from the object scene image having reproducibility corrected by each of the N correction characteristics selected by the selection unit. N characters recognized by the recognition means are displayed by the display means. The character selection operation is displayed by the receiving unit regarding the display process of the display unit. The storage unit stores one character according to the character selection operation among the N characters displayed by the display unit.

  As described above, the correction characteristic for correcting the reproducibility of the object scene image subjected to character recognition is selected based on the contrast of the object scene image. As a result, recognition accuracy is improved. Furthermore, the number of correction characteristics to be selected is smaller than the number of correction characteristics held by the holding means. As a result, inappropriate recognition processing is eliminated. Furthermore, the number of correction characteristics selected for character recognition is two or more, and any one of the two or more recognized characters is selected by a character selection operation. As a result, it is left to the operator's operation which character is finally selected.

  As a result, the same character as that appearing in the object scene can be created by a simple process.

  The character processing device according to the invention of claim 2 is dependent on claim 1, and the holding means divides the M correction characteristics into a plurality of groups so that each of the holding means has N correction characteristics, and the selection means has a plurality of selection means. N correction features are selected from any one of the groups.

  A character processing device according to a third aspect of the invention is dependent on the first or second aspect, and the selection means executes the selection process in response to a character recognition operation.

  A portable terminal (10) according to the invention of claim 4 comprises the character processing device according to any one of claims 1 to 3.

  A character processing program according to a fifth aspect of the present invention is an image pickup means for capturing a scene and M pieces (M: an integer equal to or greater than 3) for correcting the reproducibility of the scene image captured by the image pickup means in different manners. The processor (60) of the character processing device (10) having the holding means (52b to 52g) for holding the correction characteristics of the M is based on the contrast of the object scene image captured by the imaging means. A selection step (S5) for selecting N (N: integer greater than or equal to 2 and less than M) correction characteristics, and an object field having reproducibility corrected by each of the N correction characteristics selected by the selection means A recognition step for recognizing characters from the image (S9), a display step for displaying N characters recognized by the recognition means (S15), and a reception step for accepting a character selection operation related to the display processing of the display means (S17) And by display means Are N number of storage step of storing a single character following the character selection operation of the character (S19) is executed.

  According to the present invention, the correction characteristic for correcting the reproducibility of the object scene image subjected to character recognition is selected based on the contrast of the object scene image. As a result, recognition accuracy is improved. Furthermore, the number of correction characteristics to be selected is smaller than the number of correction characteristics held by the holding means. As a result, inappropriate recognition processing is eliminated. Furthermore, the number of correction characteristics selected for character recognition is two or more, and any one of the two or more recognized characters is selected by a character selection operation. As a result, it is left to the operator's operation which character is finally selected. As a result, the same character as that appearing in the object scene can be created by a simple process.

  The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

  With reference to FIG. 1, the portable terminal 10 of this embodiment includes an operation key 26. When a speech operation for a call is performed by the operation key 26, the main CPU 30 transmits a speech signal to a partner mobile terminal (not shown) through the signal processing circuit 16, the wireless circuit 14, and the antenna 12. On the other hand, when the other party performs an incoming call operation, it becomes possible to talk.

  When voice is input to the microphone 24 after the call is ready, this voice is converted by the microphone 20 into a voice signal that is an analog signal. The converted audio signal is amplified by the amplifier 22 and converted by the AD / DA conversion circuit 18 into audio data that is a digital signal. The converted audio data is subjected to coding processing by the signal processing circuit 16 and further subjected to modulation processing by the radio circuit 14. The modulated audio data generated by the radio circuit 14 is transmitted from the antenna 12.

  On the other hand, the modulated audio data sent from the other party is received by the antenna 12, subjected to demodulation processing by the radio circuit 14, and subjected to composite processing by the signal processing circuit 16. The audio data combined by the signal processing circuit 16 is converted into an analog audio signal by the AD / DA conversion circuit 18, and the converted audio signal is output from the speaker 26 via the amplifier 24.

  When a call end operation is performed by the operation key 26 while a call is being performed with the other party in this way, the main CPU 30 controls the signal processing circuit 16 and the radio circuit 14 to call the other party. Send a call end signal. After transmitting the call end signal, the main CPU 30 ends the call process. Also when the call end signal is received from the other party first, the main CPU 30 ends the call process.

  When the character recognition function is validated by the operation key 26 in a state where no telephone call is made, a through image output process is executed by the camera unit 28 and the main CPU 30. The camera unit 28 first captures a scene image, generates raw image data corresponding to the captured scene image at a frame rate of 30 fps, and performs a γ correction process on the generated raw image data of each frame. Including signal processing. The image data of each frame obtained in this way is output to the main CPU 30. The main CPU 30 writes the image data output from the camera unit 28 into the VRAM 32 at a frame rate of 30 fps. The image data written in the VRAM 32 is read by the LCD driver 34, and as a result, a real-time moving image (through image) of the object scene image is displayed on the LCD 36.

  When a character recognition start operation is performed by a button (not shown) provided on the operation key 26, a character recognition process is executed by the camera unit 28 and the main CPU 30. The camera unit 28 performs binarization processing on the image data that has been subjected to the γ correction processing, and refers to the text code table to detect a character that approximates the character expressed by the binarized data. The text code table is stored in the flash memory 62. The text code table is a comparison table between character codes and bitmapped characters.

  The main CPU 30 draws bitmap data representing the detected character in the VRAM 32. The LCD driver 34 reads the bitmap data from the VRAM 32 and displays the corresponding characters on the LCD 36. In this way, the same (or similar) character as the character appearing in the scene is displayed on the LCD 36.

  Such character recognition processing and character display processing are executed a plurality of times using different γ correction curves until a character recognition end operation is performed. A plurality of characters are displayed on the LCD 36.

  When the operator performs an operation of selecting one desired character from the plurality of characters displayed on the LCD 36, that is, a character selection operation, the main CPU 30 sets a character code corresponding to the character selected by the operator as a text code. The character code read from the table is stored in the flash memory 42.

  As shown in FIG. 2, the LCD 36 includes a pictogram area 36a, a preview area 36b, a recognition result area 36c, and a key operation function area 36d. The LCD driver 34 outputs the bitmap data drawn in the VRAM 32 to the LCD 36 based on a command issued from the main CPU 30. During the through image output process, the through image is displayed in the preview area 36b. During the character recognition process, a through image is displayed in the preview area 36b, and a character is displayed in the recognition result area 36c. During display processing of the editing screen, a plurality of characters are displayed in the recognition result area 36c.

  The camera unit 28 is configured in detail as shown in FIG. The optical image of the object scene is irradiated onto the imaging surface of the image sensor 48 via the optical lens 46. On the imaging surface, a charge corresponding to the optical image of the object scene, that is, a raw image signal is generated by photoelectric conversion. The camera CPU 60 commands the timing generator (TG) 58 to repeat pre-exposure and thinning-out reading.

  The TG 58 repeatedly executes pre-exposure of the image sensor 48 and thinning-out reading of the image signal generated thereby. The TG 58 also generates a plurality of timing signals including a pixel clock, a horizontal synchronization signal Hsync, and a vertical synchronization signal Vsync. The driver 56 drives the image sensor 48 in response to the timing signal. As a result, the vertical synchronization signal Vsync is generated every 1/30 seconds, and the image signal is output from the image sensor 48 at a rate of 1 frame per 1/30 seconds.

  The output raw image signal is subjected to a series of processes of noise removal, level adjustment and A / D conversion by the CDS / AGC / AD circuit 50, thereby obtaining raw image data as a digital signal. The γ correction circuit 52 performs γ correction processing on the raw image data output from the CDS / AGC / AD circuit 50, and the camera signal processing circuit 54 adjusts the white balance to the γ correction image data output from the γ correction circuit 52. , Color separation, YUV conversion, and the like are performed to generate YUV format image data. The generated image data is written into the SDRAM 68 by the memory controller 66. The camera CPU 60 reads the image data written in the SDRAM 68 and performs the character recognition process described above on the image data. The character code obtained in this way is written again into the SDRAM 68 by the memory controller 66. The memory controller 66 also reads the character code written in the SDRAM 68 and outputs it to the I / F 70. Thus, the character code generated by the camera unit 28 is output to the main CPU 30.

  The γ correction circuit 52 is configured in detail as shown in FIG. The clamp circuit 52a performs clamp processing on the raw image data output from the CDS / AGC / AD circuit 50, and outputs the raw image data subjected to the clamp processing to the switch SW1.

  The camera CPU 60 determines whether or not the contrast of the image data stored in the SDRAM 68 is larger than a threshold value, and issues a switching command to the γ correction circuit 52 according to the determination result. The switch SW1 is connected to the terminal T1 if the contrast is greater than the threshold value, and is connected to the terminal T2 if the contrast is equal to or less than the threshold value. That is, the connection destination of the switch SW1 is changed based on the contrast level of the image data. Thereby, the time required for the character recognition process is shortened.

  The six LUTs 52b to 52g formed in the γ correction circuit 52 are divided into two groups, a binarization processing group (GRP1) and a linear processing group (GRP2). GRP1 is composed of LUT 52b, LUT 52c and LUT 52d, and GRP2 is composed of LUT 52e, LUT 52f and LUT 52g.

  When the raw image data output from the clamp circuit 52a is applied to the LUT 52b, the LUT 52b outputs the raw image data corrected according to the γ curve shown in FIG. 5A toward the terminal T3. In this way, the γ-corrected image data that has been subjected to intensity binarization processing is output from the LUT 52b.

  When the raw image data output from the clamp circuit 52a is applied to the LUT 52c, the LUT 52c outputs the raw image data corrected according to the γ curve shown in FIG. 5B toward the terminal T4. In this way, the γ-corrected image data subjected to the medium binarization process is output from the LUT 52c.

  When the raw image data output from the clamp circuit 52a is given to the LUT 52d, the LUT 52d outputs the raw image data corrected according to the γ curve shown in FIG. 5C toward the terminal T5. In this way, the γ-corrected image data that has been subjected to the binarization processing of the weakness is output from the LUT 52d.

  When the raw image data output from the clamp circuit 52a is applied to the LUT 52e, the LUT 52e outputs the raw image data corrected according to the γ curve shown in FIG. 6A toward the terminal T6. In this way, the γ corrected image data subjected to the intensity linear process is output from the LUT 52e.

  When the raw image data output from the clamp circuit 52a is given to the LUT 52f, the LUT 52f outputs the raw image data corrected according to the γ curve shown in FIG. 6B toward the terminal T7. In this way, the γ corrected image data subjected to the medium linear processing is output from the LUT 52f.

  When the raw image data output from the clamp circuit 52a is given to the LUT 52f, the LUT 52f outputs the raw image data corrected according to the γ curve shown in FIG. 6C toward the terminal T8. In this way, the γ-corrected image data subjected to the weak linear processing is output from the LUT 52g.

  Returning to FIG. 4, the camera CPU 60 issues a command for simultaneously switching the switches SW2 and SW3 to the γ correction circuit 52, and the first area 68a, the second area 68b and the third area 68c formed in the SDRAM 68 according to the switching command. A write destination switching command to (see FIG. 7) is issued to the memory controller 66.

  Specifically, the switches SW2 and SW3 are connected to terminals T3 and T6, respectively. The controller 66 writes the image data given from the camera signal processing circuit 54 in the first area 68a, and reads out the image data written in the first area 68a. The camera CPU 60 converts the image data obtained thereby into a character code. The converted character code is written into the first area 68a by the controller 66.

  When a switching command is issued from the camera CPU 60, the switches SW2 and SW3 are connected to the terminals T4 and T7, respectively, and the controller 66 writes the image data given from the camera signal processing circuit 54 into the second area 68b. When a switching command is issued from the camera CPU 60, the switches SW2 and SW3 are connected to the terminals T5 and T8, respectively, and the controller 66 writes the image data given from the camera signal processing circuit 54 in the third area 68c. Further, when a switching command is issued from the camera CPU 60, the switches SW2 and SW3 are connected to the terminals T3 and T6, respectively, and the controller 66 writes the image data given from the camera signal processing circuit 54 in the first area 68a.

  That is, the connection destination of the switches SW2 and SW3 is changed every time the character code is stored in the SDRAM 68, and the controller 66 stores the three raw image data having different γ correction intensities in the SDRAM 68.

  According to this embodiment, when a subject having a large contrast between the character and the background as shown in FIG. 8A is photographed, the character being subjected to the character recognition process is displayed in the recognition result area 42c (see FIG. 9). ). When the character recognition process is completed, the characters corresponding to the character codes stored in the first area 68a, the second area 68b, and the third area 68c are displayed as shown in FIG.

  On the other hand, when a subject having a small contrast between the character and the background is photographed as shown in FIG. 8B, when the character recognition process is completed, the characters are stored in the first area 68a, the second area 68b, and the third area 68c. Characters corresponding to the character codes are displayed as shown in FIG.

  When the character recognition function is validated, the camera CPU 60 executes processing according to the flowcharts shown in FIGS. A control program corresponding to the flowcharts shown in FIGS. 11 to 13 is stored in the flash memory 62.

  Referring to FIG. 11, initial setting is performed in step S1. Specifically, an instruction to connect the switch SW1 to the terminal T1, an instruction to connect the switch SW2 to the terminal T3, and an instruction to connect the switch SW3 to the terminal T6 are issued to the γ correction circuit 52, and the count value N is set to 1. To do.

  In step S3, the process waits until a recognition start operation is performed. When a recognition start operation is performed, a selection process is performed in step S5. In step S7, the process waits until the vertical synchronization signal Vsync is generated. When the vertical synchronization signal Vsync is generated, character recognition processing is performed in step S9. In step S11, a γ curve switching process is executed. When this process is completed, it is determined in step S13 whether or not a recognition end operation has been performed. If the determination result is YES, the process proceeds to step S15, and if the determination result is NO, the process returns to step S7. In step S15, a list display process is executed to display the character codes stored in the SDRAM 68 in the recognition result area 42c. When a desired character is selected by the operator in step S17, it is determined that the selection is completed, and the process proceeds to step S19. In step S19, the character code corresponding to the desired character is stored in the flash memory 42, and this process ends.

  The selection process in step S5 will be described with reference to FIG. In step S21, it is determined whether or not the contrast of the image data stored in the SDRAM 68 is greater than a threshold value. If the determination result is YES, the process proceeds to step S23, and if the determination result is NO, the process proceeds to step S25. In step S23, a command for connecting the switch SW1 to the terminal T1 is issued to the γ correction circuit 52, and the process returns to the upper routine. In step S25, a command to connect the switch SW1 to the terminal T2 is issued to the γ correction circuit 52, and the process returns to the upper routine.

  The γ curve switching process in step S11 will be described with reference to FIG. In step S31, the count value N is incremented. In step S33, it is determined whether the count value N is greater than 3. If the determination result is YES, the process proceeds to step S35, and if the determination result is NO, the process proceeds to step S37. In step S35, 1 is substituted into the count value N, and the process proceeds to step S37. In step S37, an instruction to switch the switches SW2 and SW3 based on the count value N is issued to the γ correction circuit 52. That is, when the count value N is 1, the switches SW2 and SW3 are connected to the terminals T3 and T6, respectively, and when the count value N is 2, the switches SW2 and SW3 are connected to the terminals T4 and T7, respectively. Is 3, the switches SW2 and SW3 are connected to the terminals T5 and T8.

  As can be seen from the above description, the object scene is captured by the image sensor 48. The LUTs 52b to 52g hold six γ correction characteristics that correct the reproducibility of the object scene image captured by the image sensor 48 in different modes. The camera CPU 60 selects three correction characteristics from the six correction characteristics based on the contrast of the object scene image captured by the image sensor 48.

  The character is recognized by the camera CPU 60 from the object scene image having reproducibility corrected by each of the three γ correction characteristics selected by the camera CPU 60. Three characters recognized by the camera CPU 60 are displayed on the LCD 36. The character selection operation is accepted by the camera CPU 60. Of the three characters displayed on the LCD 36, one character following the character selection operation is stored in the flash memory 42.

As described above, the γ correction characteristic for correcting the reproducibility of the object scene image subjected to character recognition is selected based on the contrast of the object scene image. As a result, recognition accuracy is improved. Further, the number of γ correction characteristics to be selected is smaller than the number of γ correction characteristics held by the LUTs 52a to 52g. As a result, inappropriate recognition processing is eliminated. Furthermore, the number of γ correction characteristics selected for character recognition is two or more, and any one of the two or more recognized characters is selected by a character selection operation. As a result, it is left to the operator's operation which character is finally selected.
As a result, the same character as that appearing in the object scene can be created by a simple process.

  In this embodiment, the object of the recognition process is a character. However, the present invention is not limited to this, and character data such as numbers, codes, and alphabets can also be recognized.

  In this embodiment, since the binarization processing group is composed of LUTs 52b to 52d and the linear processing group is composed of LUTs 52e to 52g, each of the LUTs 52b to 52g is both a binarization processing group and a linear processing group. Does not belong to. However, the present invention is not limited to this, and any one of the LUTs 52b to 52g may have a configuration belonging to both the binarization processing group and the linear processing group.

It is a block diagram which shows the structure of one Example of this invention. 4 is an illustrative view showing one example of a configuration of an LCD 36. FIG. It is a block diagram which shows the structure of the camera unit 28 applied to the FIG. 1 Example. FIG. 4 is a block diagram showing a configuration of a γ correction circuit 52 applied to the embodiment in FIG. 3. (A) is an illustrative view showing an input / output mode of the LUT 52b, (B) is an illustrative view showing an input / output mode of the LUT 52c, and (C) is an illustrative view showing an input / output mode of the LUT 52d. (A) is an illustrative view showing an input / output aspect of the LUT 52e, (B) is an illustrative view showing an input / output aspect of the LUT 52f, and (C) is an illustrative view showing an input / output aspect of the LUT 52g. FIG. 4 is an illustrative view showing one example of mapping of SDRAM 68 applied to the embodiment in FIG. 3; (A) is an illustration figure which shows an example of the display mode of the preview area 42b, (B) is an illustration figure which shows another example of the display mode of the preview area 42b. It is an illustration figure which shows an example of the display mode of the preview area 42c applied to the FIG. 1 Example. (A) is an illustration figure which shows another example of the display mode of the preview area 42c, (B) is an illustration figure which shows another example of the display mode of the preview area 42c. It is a flowchart which shows a part of operation | movement of camera CPU60 applied to the FIG. 1 Example. It is a flowchart which shows a part of other operation | movement of camera CPU60 applied to the FIG. 1 Example. It is a flowchart which shows a part of other operation | movement of camera CPU60 applied to the FIG. 1 Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Mobile telephone apparatus 12 ... Antenna 26 ... Operation key 28 ... Camera unit 30 ... Main CPU
36 ... LCD
48 ... Image sensor 52 ... Gamma correction circuit 52a ... Clamp circuit 52b-52g ... LUT
60 ... Camera CPU
68 ... SDRAM
48 ... USB controller 50 ... USB terminal 52 ... PC

Claims (5)

Imaging means for capturing the object scene,
Holding means for holding M (M: integer greater than or equal to 3) correction characteristics for correcting the reproducibility of the object scene image captured by the imaging means in different manners;
Selection means for selecting N (N: an integer less than or equal to 2 and less than M) correction characteristics from the M correction characteristics based on the contrast of the object scene image captured by the imaging means;
Recognizing means for recognizing characters from an object scene image having reproducibility corrected by each of the N correction characteristics selected by the selecting means;
Display means for displaying N characters recognized by the recognition means;
Character processing comprising: a receiving unit that receives a character selection operation related to the display process of the display unit; and a storage unit that stores one character according to the character selection operation among N characters displayed by the display unit apparatus. The holding means divides the M correction characteristics into a plurality of groups so that each has the N correction characteristics,
The character processing apparatus according to claim 1, wherein the selection unit selects the N correction features from any one of the plurality of groups.   The character processing apparatus according to claim 1, wherein the selection unit executes a selection process in response to a character recognition operation.   A portable terminal comprising the character processing device according to claim 1. Imaging means for capturing an object scene, and holding means for retaining M (M: an integer of 3 or more) correction characteristics for correcting the reproducibility of the object scene image captured by the image capturing means in different modes. In the processor of the character processor
A selection step of selecting N (N: integer from 2 to less than M) correction characteristics from the M correction characteristics based on the contrast of the object scene image captured by the imaging means;
A recognition step for recognizing characters from a scene image having reproducibility corrected by each of the N correction characteristics selected by the selection means;
A display step of displaying N characters recognized by the recognition means;
A character receiving step for receiving a character selection operation in relation to the display process of the display means, and a storage step for storing one character according to the character selection operation among the N characters displayed by the display means. Processing program.

Images