JPH0983789A - Video input device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To convert character information included in an inputted image into text data. SOLUTION: Still image information inputted from a camera part 1 is stored in a video memory part 4 as image data processed by a video input processing part 3 and stored in a storage part 6. Further, an image decision part 30 decides whether or not the image data have an image including characters by pixels or by image blocks obtained by dividing the area. When the characters are included, a character recognition part 13 converts the data into text data in sequence. The converted text data are put together by a graphic composition part 13 while related with the image data and then stored in the storage part 6. Further, the data are displayed on a monitor 16.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video input device capable of converting character information contained in a captured image into text data.

[0002]

2. Description of the Related Art Conventionally, the effective input pixels of a camera unit in a video input device are about 400,000 pixels, and the resolution is not high enough to handle a still image of a high definition image as it is. An image captured by this camera is generally a standard video signal, for example, output by NTSC, PAL, SECAM or the like. Therefore, in order to obtain a high-quality still image with a conventional camera, it is common to divide the entire captured screen into multiple screens, and after capturing, combine each divided screen to obtain a high-quality still image. Target.
The configuration and operation processing of the above-mentioned conventional video input device will be described below with reference to FIGS. 16 and 17.

FIG. 16 is a block diagram of a conventional video input device.

In the figure, 101 is a camera unit used for a person or a document, 102 is a drive unit for moving the imaging range of the camera unit 101, and 103 is a video signal input from the camera 101 into video data. A video input processing unit for performing conversion processing, 104 is a video memory unit for storing the converted video data, 105 is an overall control unit for controlling processing from video input to output, and 106 is a video A storage unit 107 for storing data, an operation unit 107 for adjusting the device and operating the drive unit 102 of the camera unit 101, and 1
A monitor 08 displays video data.

In this configuration, first, the overall control unit 10
The drive unit 102 is driven according to the instruction of 5, and the image pickup area of the camera 101 is adjusted to a predetermined position. Next, the camera 101
The video signal input from the device becomes video data via the video input processing unit 103. If the input video signal is a composite signal such as NTSC or PAL, the video input processing unit 103 performs YC separation into a luminance signal (hereinafter, Y signal) and a color difference signal (hereinafter, C signal), and further a C signal. Cr,
Color difference separation is performed on the Cb signal to obtain a Y signal, a Cr signal, and a Cb signal, and then A / D conversion is performed. Further, if color space conversion is necessary, color space conversion processing is performed on the R, G, and B signals. Also, format conversion, resolution conversion,
If enlargement / reduction or the like is required, conversion processing of pixel density and interpolation processing by a filter or the like accompanying it are performed. The video data obtained by such video processing is repeated for the divided screens, sequentially stored in a predetermined area of the video memory unit 104, and the video data of each of these screens is pasted and combined. In this way, one high-quality still image data can be obtained. The combined high-quality still image data is stored in the storage unit 106 by the overall control unit 105, and is transferred and displayed on the monitor 108 when display is required. The above operation processing is shown in the flowchart of FIG.

FIG. 17 is a flowchart showing a conventional still image input process.

In the figure, it is judged whether or not the input of a still image is requested by the operation unit 107 (step S101),
If the still image is not input, the moving image is input from the camera unit 101, and the input moving image is displayed on the monitor 108.
To display (step S112 to step S11)
3). On the other hand, when a still image is input, the resolution of the still image to be captured is set (step S102). The imaging capability in the area that a normal camera can image is, for example,
In the case of NTSC, horizontal 768 pixels × vertical 494 lines, and in the case of PAL, horizontal 752 pixels × vertical 5
There are 82 lines, and the above-mentioned imaging capability divided by the size of the imaging target is the resolution. Here, in order to improve the resolution, the number of divided screens of the imaging target is set to n (step S103). Step S104 to Step S
Reference numeral 109 denotes a split screen input routine, which drives the driving unit 102 of the camera unit 101 in accordance with the number of screen divisions n and adjusts the position of the first imaging area. After that, the image of the imaging region is input from the camera unit 101, processed by the image input processing unit 103, and then stored in a predetermined region of the memory input control unit 104. By repeating this process n times while sequentially changing the image pickup area and the memory space of the video memory unit 104, it is possible to input, store, and display a high-quality still image.

[0008]

However, in the above conventional example, in order to capture a high-quality still image, it was necessary to capture the entire screen as a high-quality still image. Therefore, depending on the object to be imaged, only one part of the entire screen needs a high-quality image, and as a result, the high-quality image data of the other part and the processing time of the apparatus for obtaining it are wasted. There was a problem of becoming. In addition, when an image capture target including a document or characters is captured as image data, that portion is also treated as image data, so a considerably larger storage capacity than that required for storing as text data is required. There was a problem that you had to type as.

Therefore, an object of the present invention is to provide a video input device capable of converting character information contained in a captured image into text data.

[0010]

As a configuration of the video input apparatus of the present invention for achieving the above object, the following features are provided.

That is, in a video input device having a camera unit capable of moving an image pickup area and capturing an image picked up by the camera unit as a still image, an area having a character characteristic among the image characteristics of the still image is determined. An image discriminating means, a character area dividing means for dividing an area having the character characteristic, a character recognizing means for recognizing a character for each character area divided by the character area dividing means, and the character recognizing means. And a code conversion means for converting the recognized character into a code. As a result, the character area included in the image data is converted into an individual character code such as text data.

Further, there is provided an extraction area scaling means for extracting an area where the character cannot be recognized by the character recognition means and increasing the scaling ratio until the characters can be recognized in the extracted area. It is characterized in that the area after scaling is recognized again by the character recognition means. Alternatively, the character recognition means extracts an area in which characters cannot be recognized, and a resolution increasing means is provided for increasing the resolution until the characters can be recognized in the extracted area. Character recognition is performed again by the character recognition means. Alternatively, the present invention is characterized in that it comprises a scaling means for the extraction area and the resolution increasing means, and the character recognition means newly recognizes the area after the scaling and / or the resolution increasing processing. As a result, the image quality is improved only in the area where characters cannot be recognized with the image quality at the time of image capturing.

Further, it is characterized by further comprising first synthesizing means for synthesizing the character coded by the code converting means by associating with the still image or the character area. Alternatively, a second synthesizing unit for synthesizing the region in which the character recognizing unit cannot recognize the character with the still image or the character region is provided. With these, still image data that can handle data in the character area as a code is generated.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

<First Embodiment> First, a high-resolution video input device by screen synthesis according to a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a high resolution video input device by screen synthesis as the first embodiment of the present invention.

In the figure, 1 is a camera unit used for a person or a document, 2 is a drive unit for moving the image capturing range of the camera unit 1, and 3 is a video signal input from the camera 1 into video data. A video input processing unit 4 for performing conversion processing, a video memory unit 4 for storing the converted video data, and a video memory unit 5 for storing the converted video data.
An overall control unit that controls the processing from the input of the video to the output, 6 is a storage unit that stores the video data, 7 is an operation unit that adjusts the device and operates the drive unit 2 of the camera unit 1,
1 is a function that improves the resolution by increasing the magnification by using the zoom function to reduce the image size, and a function that shifts the optical axis slightly from the imaging range of each element of the CCD by the optical axis variable function using a lens. A lens unit 12 having a function of improving the image quality is output from the image memory unit 4 by a lens control unit 13 that controls a magnification and an optical axis shift amount with respect to the lens unit 12 according to an instruction from the overall control unit 5. And a graphic memory unit for synthesizing the image data and the graphic data from the graphic data memory unit, a graphic memory unit for storing the graphic data from the overall control unit, and a display unit for the video signal from the graphic synthesizing unit. The video output processing unit 17, which converts the signal into a signal for display, outputs a plurality of image data output from the video input processing unit. Video synthesis processing unit for combining capture surface, 18 is an image synthesis memory unit for storing image data of a plurality of screens by the control of the image synthesis processing unit. The apparatus according to the first embodiment may have a configuration in which the lens unit 11 and the lens driving unit 12 are not provided (second unit described later).
Is required in the embodiment of).

Next, the camera section 1 will be described with reference to FIG.

FIG. 3 is an internal block diagram of the camera unit 1 according to the first embodiment of the present invention.

In the figure, 201 is an aperture shutter, 202
Is an optical low-pass filter, and 203 is the photoelectric conversion element 20 in synchronization with a sync signal from a sync signal generator (not shown).
4, a CCD driver for driving 4, a photoelectric conversion element 204 for receiving an optical signal of an imaged object, converting it into an electric signal and outputting the electric signal, 205 an automatic gain control for amplifying the signal output from the photoelectric conversion element Department (AGC), 20
Reference numeral 6 is an aperture photometering circuit for measuring the aperture amount by a signal from the photoelectric conversion element, 207 is an iris drive unit for driving the aperture shutter by the aperture photometering circuit to adjust the aperture, and 208 is a minute movement of the optical axis. A parallel plate that shifts the pixels to perform pixel shifting, 209 is a field read control unit that controls an output method of a signal output from the photoelectric conversion element, 210
Is a low-pass filter drive unit for inserting / removing the low-pass filter 202, and 211 is a parallel plate drive unit for controlling the optical axis shift of the parallel plate 208.

The apparatus of the first embodiment may have a configuration in which the parallel plate 208 and the parallel plate drive unit 211 are not provided (essential in the second embodiment described later). Further, the low-pass filter driving section 210 is unnecessary when the parallel plate 208 and the parallel-plate driving section 211 are not provided, and in that case, the low-pass filter 202 is left attached.

As described above, the optical information picked up by the camera unit 1 is converted into an electric signal and outputted, and the aperture level and the gain adjustment automatically adjust the output level so that it falls within an appropriate range. Has been done. Here, the signal output from the camera unit 1 has a configuration capable of outputting a signal suitable for each of a moving image and a still image, and a moving image mode / still image is output in response to a request from an operator in the operation unit 7. The image mode can be switched.

Next, the mode of the output signal will be described with reference to FIGS.

FIG. 5 is a diagram showing a filter array on the elements of the photoelectric conversion element 204 according to the first embodiment of the present invention.

In the figure, Cy (cyan), Ye (yellow), G (green), and Mg (magenta) filters are arranged in an arrangement as shown in FIG. 5, which is called a complementary color checkered arrangement. (However, G is a primary color, not a complementary color). In this arrangement, it is better to receive light through the complementary colors Cy, Ye, and Mg than the primary color R (red), G (green), and B (blue) filters, for each complementary color. Since the two primary colors are mixed and received, more information is obtained and the sensitivity is improved. The G and Mg arrays are alternately arranged for each line, and the signals of the filters above and below the filter array are added and output. Therefore, the signal output from the camera unit 1 is the field of FIG. As shown in the read mode.

FIG. 6 is a diagram showing a filter array in the field reading according to the first embodiment of the present invention.

In the figure, the odd lines are Cy + G and Ye + M.
g is repeated, and Cy + Mg and Ye + G are repeated for even lines. Here, the filter characteristics of the Y signal and the C signal are set so as to be obtained by the following equation. Y = {(Cy + G) + (Ye + Mg)} × 1/2 R−Y = {(Ye + Mg) − (Cy + G)} − (B−Y) = {(Ye + G) − (Cy + Mg)} Therefore, the video input processing unit In 3, the Y signal and the C signal are generated by adding and subtracting as described above. That is, in the case of moving image processing or still image processing by spatially dividing the screen, the field reading mode as described above is used.

Next, the video input processing section 3 will be described with reference to FIG.

FIG. 4 is an internal block diagram of the video input processing unit 3 according to the first embodiment of the present invention.

In the figure, reference numeral 301 denotes a delay circuit for two horizontal lines, which outputs signals with no delay, one line delay, and two line delay with respect to the signal from the camera section 1. The output signal itself is the same as the configuration shown in FIG. 302
Generates a Y signal by adding the odd-numbered and even-numbered output signals from the delay circuit 301, and uses the signals for three lines generated by the delay circuit to perform horizontal / vertical aperture correction processing. And a horizontal / vertical aperture correction unit 303 for performing a gamma correction process on the Y signal output from the horizontal / vertical aperture correction unit.
Reference numeral 304 denotes a Y signal, a Cr (RY) signal, and a Cb (BY) signal by adding and subtracting odd-numbered and even-numbered signals of each signal using the signals of three lines from the delay circuit 301.
A synchronous detection unit 305 for generating a signal is an RGB that performs matrix conversion for color-converting the YCrCb signal into an RGB signal.
The matrix conversion unit 306 keeps RGB color reproducibility constant with respect to changes in the color temperature of the light source at the time of imaging,
A white balance adjusting unit that adjusts the RGB signals so that the white level obtained by combining the B signals becomes a reference white level, 307 is a gamma correction processing unit that performs gamma correction processing on the RGB signals, and 308 is , RGB signal is Cr
It is a color difference matrix conversion unit that performs color conversion into an (RY) signal and a Cb (BY) signal. That is, the video input processing unit 3 converts the electric signal of the optical information received from the camera unit 1 into Y,
The image information of the U and V signals is converted and output, and aperture correction, gamma correction, white balance adjustment are performed on the color signals, and the image information is automatically adjusted to an appropriate level.

Next, a method of inputting an image when the high resolution image input device by screen synthesis inputs an image pickup target with high image quality will be described with reference to FIG.

FIG. 9 is a diagram showing image input by screen division according to the first embodiment of the present invention.

In FIG. 9, the high-resolution video input device by screen synthesis is driven by driving the driving unit 2 to move the image pickup area of the camera vertically and horizontally in order to input the image pickup object with high image quality. The image is captured for each screen (in the example of FIG. 9, the case where the image area marked with a circle is captured as the image data). Further, the control of the drive motor that divides the imaging area is easy to realize because the current drive motor has sufficiently high accuracy. However, regarding the composition of divided and input screens inside the device, the boundaries between combined screens are discontinuous, and the time taken to capture each divided screen is slightly different. Becomes Therefore, in order to minimize the influence of the time lag when combining the individual split screens, (1) keep the environment constant during that time, or store the state of the video signal near the boundary of the multiple split screens. , Adjust the video signal so that the areas near the boundary are at the same level. (2) Duplicating the vicinity of the boundaries of a plurality of divided screens, inputting the image, and pattern matching. (3) Perform processing such as correction of screen distortion at the time of capturing an image and synthesize. It is possible to obtain a good still image by solving the discontinuity at the boundary by using such a method.

Next, the image determination unit 30 will be described.
The image determination unit 30 determines whether the imaging region is a region such as a character for each pixel or each image block obtained by dividing the region, and notifies the overall control unit 5 of the determination result. is there. The determination method will be described below.

(1) First, as one of the determination methods, there is a determination method based on the frequency component of the image. For example,
In a natural image, many flat parts are seen, and therefore a tendency to concentrate on low frequency components is seen when comparing frequency components.
In addition, since characters and figures have many contour portions and the contrast is high in order to further enhance the contours, when compared by frequency components, there is a tendency that there are more high frequency components than in the case of a natural image. Furthermore, in the case of fine patterns, etc.,
Since the pattern is finer than that of a character or graphic, more high-frequency components can be seen than the frequency component of a character or graphic. Due to these properties, it is possible to determine whether the image is a natural image or a character or figure by examining the frequency component of the image. Specifically, the image area is decomposed into frequency components by space / frequency conversion (FFT, DCT, etc.), the value of each frequency component is extracted, and the threshold value of the low frequency component is set according to the above-mentioned discrimination criterion. If it is smaller than the threshold value, it is determined that it is a natural image, and if it is larger than the threshold value, it is determined that it is a character or figure. Although this method is excellent as a discrimination method, it requires considerably complicated and large-scale processing.

(2) The following method is a method for determining based on the variation of the image data, and it is determined by cumulatively adding the absolute value of the difference between the average value of each target image and the value of each image. There is a way. Also in this case, when the value of the cumulative addition result is small, it is determined that there is little abrupt change and that the image is flat and that the image is close to a natural image. If the difference is large, it can be considered that the contrast is large and a sudden change occurs, such as characters, figures, and patterns.

The other method is a method of making a determination using the standard deviation of the image.

Γ = γ + (each pixel value-average of each pixel value) ^
2 (however, ^ 2 represents the square) is used to perform cumulative addition of the square of the difference between the average value of the target pixel and the pixel value for determination. In this case, if the standard deviation is small, it is determined that there is little sudden change and that the image is flat and that the image is close to a natural image. If the standard deviation is large, it can be considered that the contrast is large and a sharp change occurs, such as characters, figures, and patterns. A threshold value is set according to this criterion, and when it is smaller than that value, it is determined as a natural image, and when it is larger, it is determined as a character or figure. In this case,
The arithmetic processing is simple and can be processed at high speed. However, the accuracy is lower than the method using the frequency component.

(3) Furthermore, there is a method of detecting and determining an edge or the like existing in the image. That is, it is possible to determine that a portion having many edges is a character or a figure. vice versa,
When there are few edges, it can be determined that the image is a natural image. As a specific determination method, there is a method of detecting a local mode change by a difference filter or the like, and a weighting process is performed on the image data with a difference filter such as a linear Sobel operator of the first-order difference or a Laplacian of the second-order difference. (Mask) and if the value (absolute value) is large, it is determined that an edge exists in the image area. This allows
The number of edges in the image block is set as a threshold value, and if it is smaller than the threshold value, it is determined to be a natural image, and if it is larger than the threshold value, it is determined to be a character or a figure.

As a method of improving the discrimination accuracy of the natural image area, there are the following methods. First, a boundary where image characteristics change is detected by edge detection, and a contour of a certain image is extracted to clarify the boundary where image characteristics differ by the above determination. For example, when an image or figure is embedded in a sentence, a large change occurs in the characteristics of the image near the embedded boundary, so that the boundary can be clearly defined and divided. There is also a contour detection method by thinning processing.

Next, there is also a method of avoiding a malfunction of discrimination.
For example, if all flat images are identified as natural images, the background color (white background, black background, etc.) may be input with high image quality. To avoid this, the background color should be registered in advance, or the image The image of the part that can become the background color in the area is captured and compared, if it is the same, it is recognized as the background and input processing with high image quality is not performed, otherwise it is recognized as different from the background and input with high image quality Perform processing.

By using a plurality of discrimination methods in the above (1) to (3) together, more accurate and highly accurate image discrimination can be performed. In this way, it is possible to specify a region where an image such as a natural image, a character, or a figure is determined.
Needless to say, methods other than the above-described image discrimination method can be applied to the present invention.

Next, the character recognition section 31 will be described.
First, the image data is binarized, cut out in units of one character, the size of the cut out character is calculated, and the size of the character referred to for the recognition process is normalized. Further, with respect to this normalized character, a character is extracted from a dictionary stored in the storage unit 6 in advance, the difference is compared and evaluated, the character with the smallest difference is found, and its character code is used as a recognition result. Output. Here, regarding the character cut-out, when characters are printed in print at a constant interval such as in books and newspapers, the space area between the characters is extracted, and the characters are separated and extracted in character units. It is possible.

The character recognition method is roughly classified into a pattern matching method and a structure analysis method. The pattern matching method is a method of superimposing a standard character template of each character in a dictionary and a character input as an image. On the other hand, the structure analysis method extracts some features that express the structure such as the direction, size, shape, connection, and intersection of the line elements that make up a character, and inputs the features and images of each character in the dictionary. This is a method of matching with the characteristics of characters. Furthermore, since there is a limit in accuracy and quality in recognizing character by character as described above, a method of incorporating natural language processing technology and recognizing it by the context or context of words or sentences is used. By introducing it, the recognition rate can be dramatically improved. Needless to say, methods other than the character recognition method described above can be applied to the present invention.

Next, the flow of character recognition processing of the character recognition unit 31 will be described with reference to FIGS. 14 and 15.

FIG. 14 (FIGS. 14A to 14D) is a flow chart showing the character recognition processing as the first embodiment of the present invention.

In the figure, in step S1, it is determined whether or not a still image is input. If it is not a still image, a moving image is input from the image pickup unit (step S13), transferred to the monitor 16 and displayed (step S14). In the case of inputting a still image, it is determined in step S2 whether the input request is of high image quality for the entire area.

If YES, the resolution for high image quality input is set (step S3), and the screen division number is also set (step S4). Next, in steps S5 to S11, the drive unit 2 is controlled to align with the first imaging region,
The image data obtained by inputting the image and processed by the image signal is stored in the image synthesis memory unit 18 by the image synthesis processing unit 17.
When storing image data in the video composition memory unit 18,
The memory range is specified by the video composition processing unit 17 so that the plurality of divided drawings to be captured do not overlap. Thus, the first image data is stored in the predetermined area of the video composition memory unit. When this operation is repeated for the divided number of screens and the acquisition is completed, the overall control unit 5 accumulates all the acquired image data in the accumulation unit 6. After that, the process returns to the selection of the moving image mode and the still image mode.

In the case of NO, the image input from the camera unit 1 is subjected to image signal processing and output to the monitor 16 via the image memory unit 4, the graphic composition unit 13 and the image output unit 15 (step S21), and the operation unit. In response to a still image input instruction from 7, the overall control unit 5 displays graphic data indicating a capture area in the still image (for example, a frame surrounding the area) and graphic data indicating an operation method for controlling the imaging range. Memory (eg zoom, pan,
(Operation screen for tilting, etc.), synthesizes with the image being captured, and displays it on the monitor (step S22). Next, on the displayed graphic screen, the lens control unit 12 and the camera unit 1 are controlled under the control of the operation unit 7 to set the image pickup area of the still image (step S23), and the set image pickup area is set. The still image is stored in the video memory unit 4 (step S24). Further, it is transferred to the overall control unit 5 and stored in the storage unit 6 (step S25).

Next, in step S26, it is determined whether character recognition processing is to be performed. In the case of NO, the process directly returns to the selection of the moving image mode and the still image mode. On the other hand, in the case of YES, the above-described image determination processing is performed (step S
27), it is determined whether there is a character area (step S2).
8). In the case of NO, the process directly returns to the selection of the moving image mode and the still image mode. On the other hand, in the case of YES, the above-described character recognition processing is performed on all the character areas, and the graphic data for identifying the character areas is combined with the video and displayed (steps S29 to S31). In step S32, it is determined whether there is a region in which character recognition is impossible. In the case of NO, the character information of the recognition area is registered in the already stored still image (step S33) and stored in the storage unit 6 (step S34), and the process returns to the determination of the moving image and still image mode. Here, the term "registration" refers to additional registration with respect to the image data that has been initially stored, or storage in association with it. The registration information includes position information, resolution information, data type and the like. On the other hand, in the case of YES, the unrecognizable area is extracted and registered, the first imaging area is designated from the areas (step S41), and the center of the imaging area is driven so as to be the center of image capturing. The section 2 is driven (step S42). Next, in step S43, it is determined whether zoom-up is possible. In the case of YES, the lens unit 11 is controlled to increase the magnification and the image is taken (step S44).

Here, the lens control section 12 is controlled to change the magnification of the image pickup area to the telephoto side so as to fit within the image pickup range. The method will be described with reference to FIG.

FIG. 15 is a diagram showing the scaling processing of the image pickup area according to the first embodiment of the present invention.

In the figure, the areas of the still image that are first captured and accumulated are (X0, Y0), (X1, Y1), (X2, Y
2), (X3, Y3), and the area in which the input with high image quality is designated is (x0, y0),
It is indicated by (x1, y1), (x2, y2), (x3, y3). Here, first, in order to make the area designated for high image quality input the center of the imaging area, (X0, Y0), (X1, Y
1), (X2, Y2), (X3, Y3) to (X'0, Y '
0), (X'1, Y'1), (X'2, Y'2), (X'3,
The image pickup area is moved so as to become Y'3). This movement amount is (X0, Y0), (X1, Y1), (X2, Y2),
Center position of (X3, Y3) and (x0, y0), (x1, y
1), (x2, y2), (x3, y3) are calculated by the deviation of the center position. Next, in order to scale the area designated for high image quality input according to the set resolution, (X'0,
Y'0), (X'1, Y'1), (X'2, Y'2),
From (X'3, Y'3) to (X "0, Y" 0), (X "1,
Y "1), (X" 2, Y "2), (X" 3, Y "3) The image area is scaled. The scaling amount is the specified resolution and the still image stored first. The scaling ratio is calculated by the scaling factor with the resolution of 1. It is of course possible to collectively perform the series of operations of this scaling process.

Scaled (X "0, Y" 0), (X "1,
The image data in the areas Y "1), (X" 2, Y "2), (X" 3, Y "3) is stored in the video memory unit 4. The overall control unit 5 stores the image data. In the image data, further specified areas are (x0, y0), (x1, y1), (x
Only image data of 2, y2) and (x3, y3) is extracted.
Then, the image of the image pickup area picked up by the above procedure is stored in the image memory, and character recognition processing is performed on all the stored areas (steps S45 to S47). As a result, if there is no unrecognizable area, the character information of the area is registered in the accumulated still image (step S4).
9) and stores in the storage unit 6 (step S50), and when all unrecognized areas already registered are completed,
The process returns to the selection of the moving image mode and the still image mode (step S51). If there is still an unrecognizable area, the imaging area is designated as the next unrecognizable area (step S
52), the driving unit is driven so that the center of the image pickup area becomes the center of image capturing, the magnification is increased, the image is stored, and the process up to character recognition is repeated. Repeat the above steps,
If the lens cannot be controlled and the magnification cannot be increased, it is determined whether to shift the pixel (step S6).
1).

If YES, it is determined whether the resolution can be increased by shifting the pixels. If YES, the resolution is set, the screen division number n is calculated, and pixel shift control is performed by the number of screen divisions. The video information is input and stored, the screens are combined, the still image processing is performed on the n pieces of combined video information, and the character recognition processing is performed on all of the still images (steps S65 to S73). As a result, if there is no unrecognizable area, the process proceeds to step S49, the character information of the recognized area is registered in the already stored still image and is accumulated in the accumulating unit 6, and the already recognized unrecognizable area is recognized. When all the areas are completed, the process returns to the selection of the moving image mode and the still image mode. If there is a region that cannot be recognized, the process returns to step S62. If the resolution cannot be increased by repeating the above procedure and if NO in step S61 (if the pixel is not shifted), the video data of the unrecognizable area is a still image if the image is shifted by pixel. In the case of a normal still image, the video data is read from the video memory from the memory (step S75), the read video data as it is is registered in the accumulated still image (step S76), and the process proceeds to step S50. .

<Second Embodiment> Next, a high-resolution video input device based on pixel shifting according to a second embodiment of the present invention will be described. Since the high resolution video input device based on the pixel shift is basically the same as the high resolution video input device based on the screen synthesis in the first embodiment, only the different points will be described below.

FIG. 2 is a block diagram of a high-resolution video input device using pixel shifting according to a second embodiment of the present invention.

In the figure, reference numeral 20 is a pixel shift data synthesis processing section for inputting color shift data of pixels shifted by a parallel plate or the like from the camera section 1 and synthesizing, and 21 is pixel shift data for storing the synthesized data. Memory part, 22
Is a still image processing unit for converting the color filter data of the pixel shift data memory unit 21 into image data.

In the case of the still image mode, the image area is shifted by half a pixel or by one pixel to capture an image for each filter to improve the resolution and filter information (Cy, Ye) having the same optical axis. , Mg, Gr) are obtained, and there is a mode for generating a still image with excellent color reproducibility. In this case, unlike the field read mode in the first embodiment, the frame read mode in which the upper and lower filter images are not added and the filter images are read as they are (therefore, 209 in the camera unit 1 in FIG. 3 is used). , In the case of data acquisition by pixel shift, it becomes a frame read control unit that performs frame read). This frame read mode is shown in FIGS.
Shown in

FIG. 7 is a diagram showing a filter array in frame reading according to the second embodiment of the present invention, in which the filter array is sequentially read as it is.

FIG. 8 is a diagram showing a filter array in frame reading according to the second embodiment of the present invention, in which only the odd-numbered columns of the filter array are read out first, and then only the even-numbered columns are read out. It is a read for each type.

Next, a supplementary explanation of the still image processing unit 22 will be made. The internal structure is almost the same as that of the video input processing unit 3, and the filter information (Cy, Ye, Mg, Gr) is sent from the pixel shift data memory unit 21. Read, Cy + G, Y
After the addition processing of e + Mg, Cy + Mg, and Ye + G, the same processing as the video input processing unit 3 is performed.

Next, the pixel shifting method will be described with reference to FIGS.
It shows in FIG.

FIG. 10 is a diagram showing image input by pixel shifting according to the second embodiment of the present invention.

In the figure, by slightly shifting the optical axis from the optical axis b to the optical axis a in the lens section 11, the image area picked up by the camera section 1 is slightly shifted. Therefore, by capturing an image each time while slightly shifting the optical axis, the same effect as that when the number of pixels of the image sensor of the camera unit 1 is increased and the resolution is improved can be obtained (in the example of FIG. 10, It is an example of capturing the image of the mark). The mechanism for slightly changing the optical axis is a mechanism for changing the apex angle of the prism lens, and the silicone oil 4 is provided between the glass plates 40 arranged in parallel with each other.
1 and the periphery thereof is sealed, and the lens controller 12 changes the inclination between both glass plates to make the apex angle variable. Here, since the lens unit 11 is slightly moved, control of the lens control unit 12 requires considerable precision. In the screen combination, the discontinuity of the boundary between the combined screens is eliminated, so that the unnaturalness after the combination is eliminated. In particular, since continuity is not lost even when the resolution changes, this is an optimal method for inputting a high-quality still image. However, since there is a time lag in capturing multiple screens, keep the environment constant during that time to minimize the effect of time lag, or use the average video signal of multiple screens. Memorize the state of
By adjusting the video signal so that the average level of the entire screen becomes the same level, it becomes possible to solve the problem of time shift and obtain a good still image.

FIG. 11 is a diagram showing image input by pixel shifting using a parallel plate as the second embodiment of the present invention.

By tilting the parallel plate 42 obliquely, the angle deviation of the incident light caused by the refractive index when light passes through the substance is used to capture an image while slightly shifting the optical axis. Thus, the resolution and color reproducibility can be improved.

FIG. 12 is an explanatory diagram of the deviation of the optical axis due to the parallel plate as the second embodiment of the present invention.

If the parallel plate 42 is perpendicular to the optical axis, the optical axis is not displaced, but as shown in the figure, when light is incident from the oblique direction of the parallel plate 42, the refractive index peculiar to the object is obtained. Causes refraction with respect to the incident angle. This refraction itself is always constant if the material is uniform and unchanged, but it changes accordingly as the thickness of the object increases. Further, when the light passes through the object, the opposite refraction occurs and becomes light parallel to the optical axis when the light enters the object. Therefore, the length d shown in FIG. 12 is the deviation of the optical axis. The length d can be calculated by the following formula.

N = sini / sin θ (n: refractive index) x = L · (tani-tan θ) d = cosi · x As a result, d = cosi · L · (sini / cosi-tan θ) = L · [sini- cosi · tan {arcsin (sini / n)}].

If the length d is the same as the length between the pixels of the image pickup device, the image can be shifted by one pixel, and if it is ½, the image can be shifted by a half pixel. FIG. 13 is a diagram showing an image of the image pickup by the pixel shift.

FIG. 13 is an explanatory diagram of image pickup by pixel shifting according to the second embodiment of the present invention.

Here, a11 is the home position, and b11, c11, and d11 are the case where the home position of a11 is shifted by one pixel by the pixel shift by the parallel plate 42. That is, the objects being imaged are the same, a11 is Cy, b11 is Ye, and c11 is M.
g and d11 are imaged with each filter image of Gr.

In the case of shifting by half a pixel, a position shifted by half a pixel in the horizontal right direction from the home position in the vertical direction is set as a new home position a12, and one pixel is shifted by b12, c12 and d12. Similarly, a
21 is a new home position and a half pixel shift is performed, so that b21, c21, d21, and a22 are new home positions, and a half pixel shift is performed to b22, c22,
It is possible to generate a high-quality still image by sequentially capturing images with d22.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device as in this embodiment. Further, it goes without saying that the present invention can be applied to the case where it is implemented by supplying a program to a system or an apparatus. In this case, the storage medium storing the program according to the present invention constitutes the present invention. Then, by reading the program from the storage medium to the system or device, the system or device operates in a predetermined manner.

<Effects of the Embodiment> (1) Image data previously captured as a still image, or a character region of a still image captured in real time by the operation unit 7 while observing the monitor 16 is subjected to image determination processing. Can be extracted, the characters can be recognized and converted into text data, and the character information can be registered / stored in association with the still image again. This makes it possible to significantly reduce the amount of data to be processed, the storage capacity, and the processing time. (2) When the input characters are too small to be recognized, the resolution can be improved by zooming or pixel shifting, and the camera drive unit 2 can finely control the camera unit 1 to the image pickup area. It is possible to improve the accuracy of only the portion that requires the accuracy of character recognition.

[0077]

As described above, according to the present invention,
It is possible to provide a video input device capable of converting character information included in a captured image into text data.

[0078]

[Brief description of drawings]

FIG. 1 is a block configuration diagram of a high resolution video input device by screen synthesis according to a first embodiment of the present invention.

FIG. 2 is a block configuration diagram of a high resolution video input device by pixel shifting as a second embodiment of the present invention.

FIG. 3 is an internal block diagram of a camera unit 1 according to the first embodiment of the present invention.

FIG. 4 is an internal block diagram of a video input processing unit 3 according to the first embodiment of the present invention.

FIG. 5 is a diagram showing a filter array on the elements of the photoelectric conversion element 204 according to the first embodiment of the present invention.

FIG. 6 is a diagram showing a filter array in field reading according to the first embodiment of the present invention.

FIG. 7 is a diagram showing a filter array in frame reading according to a second embodiment of the present invention.

FIG. 8 is a diagram showing a filter array in frame reading according to a second embodiment of the present invention.

FIG. 9 is a diagram showing image input by screen division according to the first embodiment of the present invention.

FIG. 10 is a diagram showing image input by pixel shifting according to a second embodiment of the present invention.

FIG. 11 is a diagram showing image input by pixel shifting using a parallel plate as the second embodiment of the present invention.

FIG. 12 is an explanatory diagram of an optical axis shift due to a parallel plate as the second embodiment of the present invention.

FIG. 13 is an explanatory diagram of imaging by pixel shifting according to a second embodiment of the present invention.

FIG. 14A is a flowchart showing a character recognition process according to the first embodiment of the present invention.

FIG. 14B is a flowchart showing a character recognition process as the first embodiment of the present invention.

FIG. 14C is a flowchart showing a character recognition process according to the first embodiment of the present invention.

FIG. 14D is a flowchart showing a character recognition process according to the first embodiment of the present invention.

FIG. 15 is a diagram showing a scaling process of an imaging region according to the first embodiment of the present invention.

FIG. 16 is a block configuration diagram of a video input device as a conventional example.

FIG. 17 is a flowchart showing still image input processing as a conventional example.

[Explanation of symbols]

 1 Camera Section 2 Driving Section 3 Video Input Processing Section 4 Video Memory Section 5 Overall Control Section 6 Storage Section 7 Operating Section 8 Monitor 11 Lens Section 12 Lens Control Section 13 Graphic Synthesis Section 14 Graphic Memory Section 15 Video Output Processing Section 16 Monitor 17 Image synthesis processing unit 18 Image synthesis memory unit 20 Pixel shifted data synthesis processing unit 21 Pixel shifted data memory unit 22 Still image processing unit 30 Image determination unit 31 Character recognition unit 40 Plate glass 41 Silicon oil 42 Parallel plate 101 Camera unit 102 Driving unit 103 Image input processing unit 104 Image memory unit 105 Overall control unit 105 Storage unit 107 Operation unit 108 Monitor 201 Aperture shutter 202 Optical low-pass filter 203 CCD driver 204 Photoelectric conversion element 205 Automatic gain control unit (AGC) 206 Aperture Photometric circuit 207 Iris drive unit 208 Parallel plate 209 Frame (field) readout control unit 210 Low-pass filter drive unit 211 Parallel plate drive unit 301 Delay circuit 302 Horizontal / vertical aperture correction unit 303 Gamma correction unit 304 Synchronous detection unit 305 RGB matrix conversion unit 306 White balance adjustment unit 307 Gamma correction processing unit 308 Color difference matrix conversion unit

Claims (7)

[Claims] 1. A camera unit capable of moving an imaging region,
In a video input device that captures an image captured by the camera unit as a still image, an image determination unit that determines an area having character characteristics among the image characteristics of the still image, and a character area that divides the area having the character characteristics. Dividing means, character recognizing means for recognizing a character for each character area divided by the character area dividing means, and code converting means for performing code conversion of the character recognized by the character recognizing means. A video input device characterized in that 2. An extraction area scaling means for extracting an area where character recognition is not possible by the character recognition means, and increasing a scaling ratio until a character can be recognized in the extracted area. 2. The video input device according to claim 1, wherein the character recognition section newly recognizes characters in the scaled area. 3. A resolution increasing means is provided for extracting an area where character recognition is not possible by the character recognizing means, and increasing the resolution until the character can be recognized in the extracted area. 2. The video input device according to claim 1, wherein the character recognition unit re-recognizes the different area. 4. The extraction area scaling means and the resolution increasing means are provided, and the area after the scaling and / or resolution enhancement processing is newly recognized by the character recognition means. The video input device according to claim 1. 5. The video input according to claim 1, further comprising a first synthesizing unit that associates the character coded by the code converting unit with the still image or the character region to synthesize the character. apparatus. 6. The video input device according to claim 1, further comprising a second synthesizing unit for synthesizing a region in which the character recognizing unit cannot recognize the character with the still image or the character region. . 7. The image input device according to claim 3, wherein a pixel shift is used as the resolution increasing means.