JP3198810B2 - Image input device, image input method, and image correction method in image input device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image input apparatus, an image input method, and an image input method for inputting images such as characters and figures written on an original or the like by manual operation and correcting distortion of the input image. The present invention relates to an image correction method in an input device.

[0002]

2. Description of the Related Art As an example of image input means (generally called a scanner) for reading an image by manually scanning an image of a character or a figure written on a manuscript or the like, a figure has been conventionally used. The one shown in FIG. 5 has been developed and put into practical use.

This image input means is, as shown in FIG.
Two line sensors 10 as one-dimensional image sensors
1 and 102 (hereinafter, referred to as a first sensor 101 and a second sensor 102), and two light transmission holes 111 and 112 arranged at a predetermined interval. Irradiation is performed on the characters on the original through the holes 111 and 112, and the reflected light is transmitted to the first through the optical system 3.
Of the sensor 101 and the second sensor 102.

Image data read by such an image input means is such that an image signal appearing on a first sensor 101 is transmitted to a second sensor 102 as shown in Japanese Patent Application Laid-Open No. 56-108175. A method has been proposed in which a manual scanning speed is detected by measuring a time until the image appears, a tamping signal for image sampling is controlled in accordance with the speed, and an image is input at a fixed position interval.

In Japanese Patent Application Laid-Open No. 3-286679 previously filed by the applicant of the present invention, the first
An image read at a certain position of the sensor 101 is compared with an image read by the second sensor 102, and when it is determined that the second sensor 102 has come to the reading position of the first sensor 101, The average moving speed of the image reading device at that time is detected from the sensor interval and the reading time, and the overall speed change is estimated by approximating the speed change with a curve using a plurality of average moving speeds. A method of correcting is proposed.

FIG. 6 shows an equivalent block diagram of the image processing means disclosed in Japanese Patent Application Laid-Open No. Sho 56-108175, and the outline of the operation will be described below.

In FIG. 6, a first sensor 1 according to a signal output from a read timing signal generation circuit 103 is provided.
01 and the second sensor 102 input image data.
The image data input from the first sensor 101 has its features extracted by a first feature extraction circuit 104, and the image data input from the second sensor 102 has its features extracted by a second feature extraction circuit 105. Features are extracted. The features extracted by the first feature extraction circuit 104 are stored in a feature buffer 106, and then stored in an image data input from the second sensor 102.
The characteristic is compared with the characteristic of the data by the characteristic comparison circuit 107. If the feature comparison circuit 107 determines that the features match, the speed determination circuit 108 determines the scanning speed of the image input means. That is, the fact that the characteristics of the image data read by the second sensor 102 match the characteristics of the image data read by the first sensor 101 means that the image read by the first sensor 101 2 sensors 10
2 means that the scanning has been performed, and the scanning speed of the image input means is detected from the time difference. Then, an image reading timing corresponding to the speed is sent to the timing signal generation circuit 103. Thereby, the first and second sensors 10
Reference numerals 1 and 102 denote image data at timings corresponding to the scanning speed.
The image data (the first and second sensors 1 and 2)
01 or 102) is sent to an information processing apparatus (for example, a personal computer) 109.

[0008]

However, the method disclosed in Japanese Patent Application Laid-Open No. 56-108175 does not provide the method shown in FIG.
The image 201 shown in FIG.
If the input is made in step 2, the following problem occurs. FIG. 7 shows that an image input unit scans an image 201 (here, characters “A, B, C” are drawn as characters to be read) at a relatively high speed in the vicinity of the character A. It is shown that scanning is performed at a slower speed as the distance from B is approached, and scanning is performed at a higher speed again as the distance from the character C is approached.

The image data obtained by the first sensor 101 by the feature comparison circuit 107 and the second sensor 102
Assume that the source image positions determined to be coincident with the image data obtained in the above are positions P1, P2,..., P6 in FIG.
It is assumed that the scanning speeds determined by the speed determination circuit 108 at that time are the speeds v1, v2,..., V6. For example, the image data read by the first sensor 101 at the position P1 is stored in the feature buffer 106, and the content of the feature buffer 106 is compared with the image data read by the second sensor 102. If the two data coincide with each other, it means that the second sensor 102 has read the same position P1 as the first sensor 101, and the scanning speed v1 at that time is obtained from the time difference. Similarly, at the position P2, the scanning speed v2, the position P3
In, the scanning speed at each position is obtained as the scanning speed v3.

In JP-A-56-108175, the image input timing is controlled based on the speed obtained as described above. Hereinafter, the control will be described with reference to FIG.

FIGS. 8A and 8B are basically the same as FIG.
8A and 8B, and FIG. 8C shows an image read at a predetermined timing from the timing signal generation circuit 103, which will be described later.

That is, the technique disclosed in JP-A-56-108175 controls the image input timing based on the scanning speed at each position.
, V6,..., V6, the input timing is controlled. That is, from position P1 to position P2
During the period, the input timing control based on the speed v1 is performed, and between the position P2 and the position P3, the input timing control based on the speed v2 is performed. is there.

In practice, however, as can be seen from FIG. 7 or FIG. 8, the speed change is a continuous analog change. For example, the actual speed change between the position P1 and the position P2 (region 321) is a speed change like v12 as shown in FIG.
From the position P3 to the position P3 (region 322) is v23
Such a speed change. Therefore, when viewed from the position P1 to the position P2, a large difference occurs between the actual scanning speed at a point close to the position P2 and the detected scanning speed v1.

Further, between the position P4 and the position P5 (region 3
As shown in FIG. 8A, the actual speed change of 24) is v
A speed change like 45 is shown, and a speed change between the position P5 and the position P6 (region 325) is a speed change like v56. Therefore, for example, when viewed between the position P4 and the position P5, a large difference occurs between the actual scanning speed at a point close to the position P5 and the detected scanning speed v4. In other words, in this case, image data is input at a sampling speed lower than the actual scanning speed.

Thus, the detection speeds v1, v
When input timing control is performed using 2,..., V6, a distorted image 202 as shown in FIG. 8C is input.

That is, normally, 16 scans are performed during a scan of 1 mm.
When trying to obtain the same resolution as that of taking an image twice, for example, when scanning at a speed of 30 mm / sec, 1
When an image is captured in synchronization with a 480 Hz timing signal per second and scanned at a speed of 60 mm / sec, 1
The image is captured in synchronization with the 960 Hz timing signal per second. Therefore, as the speed increases, the number of samplings increases, and more image data is captured. When the speed decreases, the number of samplings decreases, and the captured image data decreases. For example, in FIG. 8, the region 321 from the position P1 to the position P2 has an actual scanning speed gradually decreasing,
Since the data is read at the timing corresponding to the speed v1, the number of samplings increases, and the position P5 moves from the position P4 to the position P5.
The region 324 is read at a timing corresponding to the speed v4, although the actual scanning speed gradually increases, so that the number of samplings is reduced. As a result, as shown in FIG.
322,... Are input. Note that FIG.
Each area shown in (c) is the respective area 32 for convenience of explanation.
Reference numerals 331, 332 different from 1,322,.
... is attached.

As described above, the technique disclosed in Japanese Patent Application Laid-Open No. 56-108175 is a method of controlling the image reading timing of each section based on the scanning speed obtained at the start point of the section. Therefore, a time delay always occurs in the speed change, and it is impossible to input an image synchronized with the actual speed change. Therefore, even if the scanning speed is detected and the image reading timing is controlled, it is impossible to take out an image without distortion.

Japanese Unexamined Patent Publication No. 3-286679 has been proposed to solve such a problem. JP-A-3-
In 286679, for example, it is suggested that distortion correction is performed by an image processing unit in a handy scanner.
Here, an equivalent configuration diagram using a personal computer as the image processing unit is shown in FIG. 9, and the operation will be described below.

The reading timing signal generating circuit 103 supplies a timing signal of a constant speed to the first sensor 101 and the second sensor 102 so as to correspond to all practically conceivable scanning speeds in order to obtain a sufficient reading resolution. send. The sensors 102 of the first and second sensors 101 and 12 input an image in synchronization with the fixed timing signal and transmit data to a personal computer 109 or the like as an image processing device.

On the personal computer 109 side, the first feature extraction means 104 extracts feature points from the image signal sent from the first sensor 101 and stores them in a feature buffer 106. The second feature extraction unit 105 extracts a feature from the image signal sent from the second sensor 102, and the feature comparison unit 107 compares the feature with the feature point of the first sensor 101, and matches the image. Is detected. Further, based on the result, the speed judgment unit 108 judges the speed, and sends the speed information and the position information to the image correction unit 110. The image correcting means 110 approximates the speed change as a curve based on the sent speed information and speed position information, estimates the entire speed change, and performs distortion correction of the input image based on the estimated speed change.

According to this method, a continuous speed change is estimated and image distortion is corrected, so that an image without distortion can be created. Is possible, the image can be corrected according to the actual speed change.
It solves 75 problems.

However, in this method, it is necessary to input an image at a reading speed sufficient for the resolution of a target image, and a huge amount of image data is sent to a personal computer.

In other words, the technique disclosed in Japanese Patent Application Laid-Open No. 3-286679, as described above, performs sampling at a sampling rate necessary for obtaining a sufficient resolution, and therefore can practically cope with all possible scanning rates. The sampling rate (the sampling rate is constant) is set. This is because if the detection speed is lower than the actual scanning speed, the image data is taken in at a sampling speed lower than the originally required sampling speed, and the image data becomes insufficient and the target resolution is obtained. This is to prevent the situation from being lost.

[0024]

As described above, in the technique disclosed in Japanese Patent Application Laid-Open No. 3-286679, in practice, a fixed sampling rate capable of coping with all possible scanning rates is set. The processing unit (the personal computer 109 in FIG. 9) needs to hold the input image data until the feature point is detected, and therefore requires a huge memory (not shown in FIG. 9). Also,
When the amount of data is enormous, the processing time is extremely long in proportion thereto, and real-time image input processing becomes impossible. Further, since the speed detection is performed on the image processing unit side, it is necessary to always send the image data of the two sensors (first and second sensors 101 and 102) to the image processing unit (PC 109 in FIG. 9). There is also a problem that a high-speed data transfer method is required.

Therefore, the present invention solves these problems, and an object of the present invention is to provide an image input apparatus and an image input method capable of obtaining a clear image at high speed without distortion with a small amount of memory. To provide an image correction method in an image input device.

[0027]

SUMMARY OF THE INVENTION The present invention comprises an image input section and an image processing section for correcting image data read by the image input section. The first and second sensors as a one-dimensional image sensor for reading an image, and the second sensor for moving image data read at a position where the first sensor moves forward in the manual feeding direction at a certain position. In the image input device which detects the scanning speed from the time until the sensor reads, and corrects the image data according to the scanning speed, the feature is extracted from the image data input by the first sensor. First feature extracting means and second feature extracting means for extracting features from the image data inputted by the second sensor, and the feature data extracted by the first feature extracting means are stored. The feature buffer and this feature buffer A feature comparison means for comparing the feature data stored in the feature data with the feature data extracted by the second feature extraction means to determine a correspondence relationship; Speed determining means for determining a scanning speed for the image input position based on the scanning speed based on the scanning speed determined by the speed determining means.
A read timing signal generating means for controlling the read timing of the first sensor, the image data read by one of the first sensor and the second sensor, and the result of the speed determining means. It is characterized by comprising a data synthesizing means and an image correcting means for performing distortion correction of an input image based on the data synthesized by the data synthesizing means.

The first and second sensors, the first and second sensors,
The second feature extracting unit, the feature buffer, the feature comparing unit, the speed determining unit, the reading timing signal generating unit, and the data synthesizing unit are provided on the image input unit side.

Further, the data synthesizing means is provided between the image data for each line of the image data read by one of the first sensor and the second sensor. A control signal indicating that information different from the data exists and scanning speed data determined by the speed determining means are inserted as necessary.

The reading timing signal of the reading timing signal generating means sets in advance a maximum scanning speed change amount with respect to a scanning movement amount from a certain scanning speed detection position to a next scanning speed detection position. It is determined from scanning speed data at the scanning speed detection position in consideration of the maximum scanning speed change amount.

The setting of the maximum scanning speed change amount with respect to the scanning movement amount from the certain scanning speed detection position to the next scanning speed detection position is twice as large as the predetermined scanning movement amount. It is characterized by being.

Also, the present invention comprises an image input section and an image processing section for correcting image data read by the image input section, and a one-dimensional image for reading characters, figures, etc. on a document illuminated by a light source. It has first and second sensors as sensors, and until image data read at a position where the first sensor moves forward in the hand feeding direction is read by the second sensor that moves later. In the image input method of detecting the scanning speed from the time of (1) and correcting the image data according to the scanning speed, the feature of the image data input by the first sensor is sent to the first feature extracting means. The extracted feature data is stored in a feature buffer, and the features of the image data input by the second sensor are extracted by a second feature extracting means. Stored feature data and previous The feature data extracted by the second feature extracting means is compared by the feature comparing means to determine the correspondence, and then, based on the correspondence determined by the feature comparing means, the speed determining means determines the correspondence to the image input position. A scanning speed is determined, and a reading timing signal generating means controls the reading timing of the first and second sensors in accordance with the determined scanning speed,
The image data read by one of the second sensor and the second sensor is combined with the result of the speed determination means by data combining means, and the input image is synthesized based on the combined data. Distortion correction is performed.

Also in this case, the first and second sensors, the first and second feature extracting means, the feature buffer, the feature comparing means, the speed determining means, the reading timing signal generating means, and the data synthesizing means are: It is characterized in that it is provided on the image input unit side.

Also in this case, the data synthesizing means may be configured to output the image data for each line of the image data read by one of the first sensor and the second sensor. It is characterized in that a control signal indicating that information different from the image data exists and scanning speed data determined by the speed determining means are inserted as necessary.

Also in this case, the reading timing signal of the reading timing signal generating means sets in advance the maximum scanning speed change amount with respect to the scanning movement amount from one scanning speed detection position to the next scanning speed detection position. The scanning speed data at the certain scanning speed detection position is determined in consideration of the maximum scanning speed change amount.

Also in this case, the setting of the maximum scanning speed change amount with respect to the scanning movement amount from the certain scanning speed detection position to the next scanning speed detection position is twice as large as the predetermined scanning movement amount. It is characterized by a scanning speed change amount.

Also, in the present invention, a one-dimensional image for reading characters, figures, etc. on a document illuminated by a light source comprises an image input section and an image processing section for correcting image data read by the image input section. It has first and second sensors as sensors, and until image data read at a position where the first sensor moves forward in the hand feeding direction is read by the second sensor that moves later. In the image correction method in the image input device for detecting the scanning speed from the time of (1) and correcting the image data according to the scanning speed, the characteristics of the image data input by the first sensor are defined as first characteristics. The feature data is extracted by the feature extracting means, the extracted feature data is stored in a feature buffer, and the features of the image data input by the second sensor are extracted by the second feature extracting means. Stored in the feature buffer The feature data obtained and the feature data extracted by the second feature extracting means are compared by a feature comparing means to determine a correspondence, and based on the correspondence determined by the feature comparing means. The scanning speed for the image input position is determined by the speed determining unit, and the reading timing signal generating unit controls the reading timing of the first and second sensors in accordance with the determined scanning speed. The image data read by either one of the second sensor and the second sensor and the result of the speed determining means are combined by data combining means and transferred to an image processing unit. After converting the scanning speed for each image input position obtained by the speed determination means into a scanning speed for each image input time, the scanning speed for each image input time is approximated by a curve. Determines the scanning speed changes in the middle, and performs the distortion correction seeking scan amount from this curve.

The means for converting the scanning speed for each image input position into a scanning speed for each image input time comprises: By matching the image input time for each line in the slowest reading area with the image input time for each line in the area where the fastest reading has been performed, the time axis indicating the image input time is the same scale. And the position of each of the input images, the image input time for each position of the input image, and the scanning speed are made to correspond to each other.

When converting a change in scanning speed for each image input position into a change in scanning speed for each image input time, correspondence data for associating each image input position with an image input time is used. It is characterized in that the input image is not changed but only created.

[0040]

As described above, according to the present invention, the scanning speed is detected while the image is being input, and the image is read at a timing corresponding to the scanning speed. The capacity of the memory for storing the read image data can be made very small as compared with the case where the reading is performed at the timing. Further, the image data read at a timing corresponding to the scanning speed and the distortion data of which the distortion has been reduced to some extent and the speed data are combined and sent to the image processing unit, and the image processing device performs the speed data processing.
By further performing distortion correction based on the data, a clear image without distortion can be obtained.

Further, the scanning speed is detected on the image input unit side,
Since the image is read at a timing corresponding to the scanning speed, the transfer of the image data from the image input unit to the image processing unit sends both the image data read by the two sensors. Since image data from one of the sensors can be transmitted without any need, high-speed transfer means is not required, and the load on hardware can be reduced. Further, the scanning speed data is inserted together with the control signal between the image data of each line of the image data, and is transmitted to the image processing section as the synthesized data. The image data and the speed data can be easily identified on the image processing unit side.

Further, since the image data and the speed data are not separately transmitted but transmitted as synthesized data, it is easy to synchronize the image data and the speed data. −
There is no need to prepare image data and speed data as data transfer means, and the burden on hardware can be reduced. Further, the memory size of the image processing unit can be reduced to half or less.

Further, by controlling the reading timing in consideration of the amount of change in the scanning speed, it is possible to obtain image data corresponding to the target resolution. That is, the reading timing is set in consideration of the amount of change in the scanning speed during scanning from one speed detection position to the next speed detection position. This is because, as described above, if the detection speed is lower than the actual scanning speed, image data is captured at a sampling speed lower than the originally required sampling speed, and there is a disadvantage that sufficient resolution cannot be obtained. That's why. However, in order to solve this problem, if a constant sampling rate that can cope with all possible scanning rates is set in practice as in the past, an enormous amount of memory is required. However, in the present invention, the reading timing is controlled in consideration of the amount of change in the scanning speed, so that the memory can be suppressed to a minimum size and a sufficient resolution can be obtained.

Further, by setting the amount of change in the scanning speed during scanning from a certain speed detection position to the next speed detection position to be twice, the speed change from one speed detection position to the next speed detection position can be achieved. , An image having a sufficient resolution can be obtained, and the quality of the image can be guaranteed. Also, 2
The value of "2" is practically sufficient for a change in the scanning speed. By doubling the value, the memory of the image processing unit can be reduced, and the data transfer speed to the image processing unit can be reduced. Can be slowed down, and the burden on hardware can be reduced.

The image processing unit converts the scanning speed for each image input position obtained by the speed determining means into a scanning speed for each image input time, and then converts the scanning speed for each image input time. A change in scanning speed in the middle of knotting is obtained by a quadratic or cubic curve, the amount of scanning movement is detected, and distortion correction is performed. This is because, when the speed change is estimated from the relationship between the image input position and the scanning speed, since the image input time for each line in each region between the image input positions is different, the speed estimation corresponding to the actual speed change is performed. Since it cannot be performed, the speed is converted into a scanning speed for each image input time to estimate the speed. According to this, it is possible to estimate the scanning speed close to the actual scanning speed change, and to perform distortion correction using the scanning speed estimated in this way, it is possible to perform extremely accurate distortion correction.

As means for converting the scanning speed for each image input position into the scanning speed for each image input time, the image input time for each line, which is different between each area between the image input positions, is set to the most. By setting the other image input time to the reference image input time based on the high-speed image input time, the image input time for each line is the same in each area, and the input in the area other than the reference time is performed. Since the image is expanded in time in the sub-scanning direction for the sake of convenience, the image input position, the image input time for this image input position, and the scanning speed are made to correspond to each other. The same scale can be used, and the scanning speed can be estimated close to the actual change in the scanning speed. The distortion can be corrected using the estimated scanning speed. More, it can be carried out extremely high-precision distortion correction.

When converting a change in scanning speed for each image input position into a change in scanning speed for each image input time, correspondence data for associating each image input position with the image input time is used. The input image is not changed, only created. This means that if an image of an area input at the fastest speed is actually created by adding an image of another area, a large-capacity memory for storing the image is required. However, according to the present invention, the input image itself is not converted, and only the correspondence data for making each image input position correspond to the image input time is created, so that a large-capacity memory is not required.

[0048]

FIG. 1 is a block diagram of an image input apparatus according to the present invention. In the present embodiment, each processing portion of the image input apparatus of the present invention is realized in hardware, and hence the following description will be referred to as a circuit. However, the present invention is not limited to this. Any other means is possible.

In the figure, reference numeral 100 denotes an image input unit, which includes a first sensor 101, a second sensor 102, a read timing signal generation circuit 103, a first feature extraction circuit 104, a second feature extraction circuit 105, Feature buffer 10
6. It comprises a feature comparison circuit 107, a speed determination circuit 108, a data synthesis circuit 111, and the like. Also, 109
Denotes a personal computer as an information processing apparatus. Here, only an image correction unit 110 necessary for this embodiment is shown. Note that the first sensor 101 and the second sensor 1
02, the first feature extraction circuit 104, the second feature extraction circuit 105, the feature buffer 106, and the feature comparison circuit 107 have the same basic operations as those shown in FIG.

That is, the first sensor 101 and the second sensor 102 input image data in accordance with the timing signal output from the read timing signal generation circuit 103. The image data input from the first sensor 101 is extracted by a first feature extraction circuit 104, and the image data input from the second sensor 102 is extracted by a second feature extraction circuit. At 105, features are extracted. The image data extracted by the first feature extraction circuit 104
The feature of the data is stored in the feature buffer 106 and compared with the feature of the image data input from the second sensor 102 by the feature comparison circuit 107. If the feature comparison circuit 107 determines that the features match, the speed determination circuit 108 determines the scanning speed of the image input means. That is, the image data read by the first sensor 101
The characteristics of the data include the image data read by the second sensor 102.
That the characteristics of the first sensor 101 match.
Means that the image read by the second sensor 102 has been read, the scanning speed of the image input means is detected from the time difference, and an image reading timing corresponding to the scanning speed is output from the timing signal generation circuit 103. .

Although the above operation is almost the same as that of the configuration shown in FIG. 6, in the present invention, the amount of change in the scanning speed during scanning from one speed detection position to the next speed detection position is considered. To set the read timing.

This is because, as described above, if the detection speed is lower than the actual scanning speed, image data is captured at a sampling speed lower than the originally required sampling speed, and sufficient resolution cannot be obtained. This is because, in order to eliminate the inconvenience, it is necessary to set a sampling speed that can correspond to all possible scanning speeds in practical use.

It is known that the speed of a normal image can be estimated at a rate of about once every 2 mm. Therefore, in the present invention, until the next speed detection,
That is, assuming that the scanning speed does not change more than twice while the scanner is moved by 2 mm, the reading timing is set to be twice the frequency and the image is read. This is performed by issuing an instruction from the speed determination circuit 108 to the read timing signal generation circuit 103.
That is, when a speed at a certain position is detected, image data is input at a reading timing corresponding to the detected speed. , So that the reading timing is twice as long as the reading timing corresponding to the detection speed.

That is, in FIG. 8, when the scanner is at the position P1, for example, and moves from the position P1 in the direction of the position P2, in this embodiment, the scanning speed is reduced during the movement of 2 mm. The change is set to be at most twice the scanning speed at the position P1. Therefore,
In this case, if the reading timing at the position P1 is double the reading timing of the frequency, sufficient image data can be obtained even if the scanning speed after moving 2 mm from the position P1 is doubled. The setting of doubling the change in the scanning speed while moving the scanner by 2 mm is a practically sufficient setting.

As a result, if the speed is constant before the next speed detection, image data having an information amount twice as large as the target image data is obtained. Even if the scanning speed changes twice, since the image data is input at twice the sampling speed, the required amount of image information can be obtained. As a result, an image twice as large as the image input by the method disclosed in Japanese Patent Application Laid-Open No. 56-108175 is input. As a result, even if the size of the finally obtained image data is the same, the information amount of the image contained therein becomes the information amount according to the target resolution, and a more accurate image can be obtained. become. This method does not produce a noticeable difference when simply inputting an image and viewing the image with human eyes, but a large difference occurs when the method is used for OCR or the like that recognizes and codes the input image. Is very effective.

By the way, the number of image lines of the final character image is set to 400 dpi, and the reading timing is set to twice the speed as in the present invention, and the timing according to the scanning speed as in the prior art. When the input image was subjected to OCR processing in the case where reading was performed, the recognition rate was
98% when the reading timing is set to double, 1
In the case of double (the timing according to the detection speed as in the conventional case), it was 92%. As is clear from this,
As in the present invention, it is effective to set the read timing in consideration of the amount of change in the scanning speed during the scanning distance (2 mm in this embodiment) from one speed detection position to the next speed detection position. You can see clearly.

In this embodiment, the change in the scanning speed during the movement of 2 mm is set to at most twice the scanning speed at the position P1, but the numerical value setting is not limited to this.

According to the present invention, the speed determined by the speed determination circuit 108 is transmitted to the read timing signal generation circuit 103 at the same time as the data synthesis circuit 111.
Also sends speed data. In addition to the speed data, the data synthesizing circuit 111 always synchronizes an image from one of the first sensor 101 or the second sensor 102 (the second sensor 102 in this embodiment) in synchronization with the image reading timing signal. Will be sent.

In the data synthesizing circuit 111, the second sensor 1
02 and the speed judgment circuit 108
And sends the data to the image correction means 110 of the personal computer 109. On the personal computer 110,
In response to the combined data (image data and velocity data),
The distortion correction according to the scanning speed is performed.

When synthesizing the speed data and the image data in this manner, a control signal indicating that the speed data is between the data of each line of the image data from the second sensor 103. And then the speed judgment circuit 108
Insert speed data from. Therefore, the personal computer 109
By judging this control signal, the image data from the second sensor 103 and the scanning speed data can be obtained as one data.

In this embodiment, for example, "0xaaa" is used as the control signal. The “aaaa” of this control signal is a hexadecimal number, and becomes “10101010101101010” when corresponding to a binary number. This is an array of dots for an image,
Such an image rarely exists as an image to be read, and even if it exists, it is set based on the idea that it does not have much information as an image.

As described above, when there is a change in the scanning speed, the control signal "0xaaa" is inserted after the image data obtained by reading a certain line. It has a data structure that continues the determined speed data. Therefore, in order to ensure that "0xaaa" does not exist in the image data sent from the sensor (the second sensor 103 in this embodiment), the head of the image data for each line is always checked. Check that the two bytes are not "0xaaaa", and if the first two bytes are "0xaaaa", forcibly convert the first two bytes to "0xffff"
It is set to send to the side.

This is because if the first two bytes of the image data are “0xaaa”, the personal computer 109
This is because "0xaaa" is determined as a control signal. In other words, since it is assumed that "0xaaa" is not usually included in the image data to be read, if the first two bytes are "0xaaa", the two bytes are used. To "0
xffff "forcibly. This “ffff” is “111111” when corresponding to a binary number.
11111111111 ". This means "1" is white,
When "0" is black (character portion), the image is an image that is followed by blanks, and as described above, such an image is usually an image that has little meaning as an image to be read.

As a result, on the personal computer 109 side, “0”
If "xaaa" comes, it is determined that it is a control signal, and thereafter it is determined that speed data is always continued.

Note that “0x
When "aaa" has meaning as image data and is important data, other data rarely present in the image is used as a control signal instead of "0xaaa" as a control signal. I just need.

Then, in the personal computer 109, the synthesized data (image data and speed data) transmitted in this manner are transmitted.
Image correction by the image correction means 110 based on
This image correction processing will be described later.

As described above, in the present invention, the scanning speed is detected on the image input unit 100 side, and the first and second timing signals are doubled in frequency with respect to the reading timing signal based on the detected scanning speed. A personal computer which controls reading of the sensors 101 and 102 and converts image data read from one of the sensors and detection speed data into a data structure in which a control signal is inserted as described above. It is sent to the 109 side. Therefore, the image data sent to the personal computer 109 is image data that has been corrected to some extent, and image data from one sensor only needs to be transferred.
In order to obtain a sufficient resolution, in practice, two sensors read at a constant reading timing signal having a sampling speed that can correspond to all possible scanning speeds, and image data read by these two sensors is read by a personal computer. After transferring these image data to the
Compared with the method of detecting the speed and correcting the image, the amount of information of the image sent to the personal computer 109 is less than half. As a result, a low-speed data transfer method with the personal computer 109 can be used. Further, if the maximum length of the target input image is also known on the personal computer 109 side, it is possible to determine the maximum memory amount to be prepared for image information. No longer needed
This can be extremely advantageous in commercialization.

FIG. 2 shows the result of the above-described processing.
FIG. 4 is a diagram showing a relationship between an image input to the image forming apparatus and a scanning speed.
As shown in FIG. 2, the personal computer 109 has different timing speeds for each area, that is, the positions P1 and P shown in FIG.
Scan speed data v1, v2,.
., Image 20 read at a timing corresponding to v6
2 and scanning speed data v1, v2,.
v6 is input. The speed data here is exactly the same as the speed data in FIG. 8, but the speed data v1, v in FIG.
,..., V6 are positions P1, P2,
.., P6, and FIG.
, V6 are the positions P11, P12,..., P1 of the input image (FIG. 2B) 202.
2A and FIG. 8A do not always coincide with each other. Further, as described above, in the present invention, the reading timing is set to the reading timing corresponding to the scanning speed (Japanese Patent Laid-Open No. 56-108175).
Since the reading timing is twice as large as that of FIG. 2, a double amount of image data is input. Therefore, FIG.
8C is an image enlarged twice in the sub-scanning direction as compared to FIG. 8C. Accordingly, FIG. 2A is originally an image enlarged twice in the horizontal direction of the drawing. For convenience of the drawing, FIGS. 2A and 2B are halved in the horizontal direction of the drawing.
It has been reduced to.

Next, the image correction means 11 of the personal computer 109
The correction method at 0 will be described. Here, this figure 2
The case of correcting the area 332 to the area 334 of the input image 202 shown in (b) will be described with reference to FIGS.

FIG. 3 is a view showing a region 3 from the region 332 in FIG.
The portion indicated by a dotted line in FIG. 3 corresponds to the same section (region 332 to region 3) in FIG.
34).

As described above, as information for correcting the input image 202 in the area 332 to the area 334, the speed data v
2, v3, and v4. Therefore, the speed change between the respective regions is estimated using the speed data, and the image is corrected.

As a method of estimating the speed change between the respective regions, the method described in the above-mentioned Japanese Patent Application Laid-Open No. 3-286679 is basically used. That is, the technique described in Japanese Patent Application Laid-Open No. 3-286679 has a scanning speed (here, v
2, v3, v4) are connected by a quadratic or cubic curve, and the overall speed change is a continuous change.

However, when the speed is estimated based on the speed data v2 to v4 shown in FIG. 3 by the method of Japanese Patent Laid-Open No. 3-286679, a speed change 401 shown by a solid line in FIG. 3 is estimated. This is the actual speed change (indicated by the dotted line) 40
2, it can be seen that the estimated speed change is largely shifted. This corresponds to the input image position (in this figure, P1
This is because the speed was estimated on the basis of (2) to (P15), and the speed was estimated without considering that the temporal scale of the image input time of each area was not the same.

That is, the area 332 is, for example, 60 mm / s
Assuming that reading is performed at a reading timing corresponding to a scanning speed of ec and that the area 333 is read at a reading timing corresponding to a scanning speed of 30 mm / sec, for example, to obtain a resolution of reading 16 lines with a scanning movement amount of 1 mm,
The area 332 reads 960 lines per second, and the area 333 reads 480 lines per second. That is, the area 332 reads one line in about 1 msec, and the area 333 reads one line in about 2 msec.

However, in the speed estimation shown in FIG. 3, since the speed estimation is performed on the assumption that the image input time for each line of each area is constant, a deviation occurs from the actual speed curve. This is because, for example, in the case of the region 332 and the region 333, as described above, the temporal scale is different such that the image input time for each line is approximately 1 msec in the region 332 and approximately 2 msec in the region 333. It is.

Therefore, the speed change with respect to the input image position in FIG. 3A is converted into a time change with time as shown in FIG. 4A. To do this, another image input time later than this is set to the fastest image input time based on the fastest image input time for each line. In this case, the image input time for each line of the area 332 (1 ms)
The image input time for each line of the other areas 333 and 334 is adjusted to ec). In this way, for example, the region 333
When the image input time for each line is set to the image input time for each line in the area 332, the area 333 becomes as shown in FIG.
As shown in FIG.
0, whereby the image input time for each line of the area 332 and the area 333 becomes the same. Similarly, the image input time of the other area is set to the image input time of one line of the reference area. As a result, FIG.
The time axis in (a) can be on the same scale.

As a result, the input image times t2, t3, t4 and t5 at the start points of these areas and the scanning speeds v2, v3, v4 and v5 for these times are as shown in FIG. When a speed change is estimated based on these speeds v2, v3, and v4, a speed change 501 is indicated by a solid line in FIG. 4A, and this speed change is very close to an actual speed change 502 indicated by a dotted line in FIG. Things.

Next, the image is corrected based on the estimated speed curve 501. That is, in this case, distortion correction is performed on the image 202 with the remaining distortion as shown in FIG. 2B using the curve of the speed change 501 in FIG. Here, as described above, the input image times t2, t3, t4, and t5 at the start point of each area and the speed v
In order to find the relationship between 2, 3, v4, and v5, an input image 530 in which the image input time in each region is matched is shown in FIG.
Although illustrated as (b), in the present invention, the image 530 is not actually created, and the image 530 in FIG.
Are merely used for convenience of explanation. In other words, the personal computer 109 actually stores the correspondence data between the input image 202 and the speed change 501 (the image input times t2, t3, t4, and t5 at the start point of each area and the speeds v2, v3, v4, and v4 at each time). v5).

This is because if the image 530 of FIG. 4B is actually created, this image 530 is exactly the same as the image read with the image reading timing kept constant. 4) and the region 333 in FIG.
Compared with the region 333 of FIG. 4B, since the region 333 of FIG. 4B has twice as much image data as the region 333 of FIG.
This would require a large amount of memory. However, in the present invention, since the image 530 in FIG. 4B is not actually created, the memory can be minimized.

The speed change 50 with respect to time obtained here.
When the regions 332 to 334 are corrected with reference to FIG. 1, the respective regions become images of the regions 541 to 543 as shown in FIG. 4C, and an image 540 without distortion can be created. . This image correction method is the same as the method described in JP-A-3-286679, in which the amount of movement of the image input device is determined by time-integrating the curve of the speed change 501, and a corresponding image is moved each time the image input device moves by a certain amount. Use the extraction method. As a result, as shown in FIG. 4C, it is possible to obtain an image 540 without distortion according to the target resolution.

As described above, according to the present invention, an image is always read at the read timing corresponding to the actual scanning speed, and processing is performed on the image. Can be reduced. In other words, when the reading process is performed at a fixed timing, an image is read at a sampling speed that can support all scanning speeds in order to obtain a fixed resolution. Etc.), and the image processing apparatus requires a large-capacity memory to hold it. Further, in the present invention, the read image is a minimum image corrected to some extent, so that a low-speed data transfer method to the image processing apparatus is sufficient. In addition, the correction method does not perform the overall speed estimation based on the speed change for some image input positions, but temporarily converts the speed change for some image input positions to the speed change for some image input times. After the conversion, the entire speed is estimated, so that it is possible to estimate the speed very close to the actual speed change. By performing correction using the estimated speed curve, image correction with less distortion is performed. be able to.

[0082]

As described above, according to the present invention, the following effects can be obtained.

First, according to the first aspect of the present invention, the scanning speed is detected while the image is being input, and the image is read at a timing corresponding to the scanning speed.
The capacity of the memory for storing the read image data can be made very small as compared with the case where reading is performed at a constant timing that can correspond to all scanning speeds. Further, the image data and the speed data which are read at a timing corresponding to the scanning speed and have a small distortion to some extent are read.
By combining the data with the image data and sending it to the image processing unit, and further correcting the distortion based on the speed data by the image processing apparatus, a clear image without distortion can be obtained.

According to the second aspect, the scanning speed is detected on the image input unit side, and the image is read at a timing corresponding to the scanning speed. The high-speed transmission means is required because the image data from either sensor can be transmitted without transmitting both the image data read by the two sensors. The load on the hardware can be reduced.

According to the third aspect, one of the image data
Since the scanning speed data together with the control signal is inserted between the image data for each line and transmitted as synthesized data to the image processing unit, the image data and the speed data are transmitted.
Can be easily identified on the image processing unit side. Also, since the image data and the speed data are not separately transmitted but are transmitted as synthesized data, the image data and the speed data are transferred.
It is easy to synchronize with data, and there is no need to prepare image data and speed data as data transfer means, and the burden on hardware can be reduced. Also,
The memory size of the image processing unit can be reduced to half or less.

Further, by controlling the reading timing in consideration of the amount of change in the scanning speed, image data corresponding to the target resolution can be obtained. That is, the reading timing is set in consideration of the amount of change in the scanning speed during scanning from one speed detection position to the next speed detection position. This is because, as described above, if the detection speed is lower than the actual scanning speed, image data is captured at a sampling speed lower than the originally required sampling speed, and there is a disadvantage that sufficient resolution cannot be obtained. That's why. However, in order to solve this problem, if a constant sampling rate that can cope with all possible scanning rates is set in practice as in the past, an enormous amount of memory is required. However, in the present invention, the reading timing is controlled in consideration of the amount of change in the scanning speed, so that the memory can be suppressed to a minimum size and a sufficient resolution can be obtained.

Further, as set forth in claim 5, by setting the amount of change in the scanning speed during scanning from a certain speed detection position to the next speed detection position to be twice, the change from the certain speed detection position to the next speed detection position is achieved. Thus, it is possible to obtain an image having a sufficient resolution with respect to the speed change up to the speed detection position, and to guarantee the quality of the image. Further, the value of twice is a practically sufficient value for a change in the scanning speed.
The memory on the image processing unit side can be reduced, the data transfer speed to the image processing unit side can be reduced, and the load on the hardware can be reduced.

Further, according to the method of the sixth aspect, similar to the effect of the first aspect, compared with the case where reading is performed at a fixed timing that can correspond to all scanning speeds, the read image data is used. The capacity of the memory for storing can be made very small. In addition, the image data read at a timing corresponding to the scanning speed and the distortion data of which distortion is reduced to some extent and the speed data are combined and sent to the image processing unit.
The image processing apparatus further performs distortion correction based on the speed data, whereby a clear image without distortion can be obtained.

According to the seventh aspect, similarly to the effect of the second aspect, the transfer of the image data from the image input unit to the image processing unit is performed without sending both the image data read by the two sensors. Since image data from either one of the sensors only needs to be sent, high-speed transfer means is not required, and the load on hardware can be reduced.

According to the eighth aspect, similarly to the effect of the third aspect, the image data and the speed data can be easily identified on the image processing unit side. Further, since the image data is transferred as synthesized data, it is easy to synchronize the image data with the speed data.
There is no need to prepare speed data, and the load on hardware can be reduced. Further, the memory size of the image processing unit can be reduced to half or less.

According to the ninth aspect, similarly to the effect of the fourth aspect, since the reading timing is controlled in consideration of the amount of change in the scanning speed, the memory can be suppressed to a minimum size. Sufficient resolution can be obtained.

According to the tenth aspect, similarly to the effect of the fifth aspect, it is possible to obtain an image having a sufficient resolution with respect to a speed change from one speed detection position to the next speed detection position. Quality can be guaranteed. The value of 2 is practically sufficient for a change in the scanning speed. By making the value of 2 times, the memory of the image processing unit can be reduced, and the data to the image processing unit can be reduced. Data transfer speed can be reduced, and the burden on hardware can be reduced.

According to the method of claim 11, the image processing section converts the scanning speed for each image input position obtained by the speed determining means into a scanning speed for each image input time, The scanning speed for each image input time is linked by a quadratic or cubic curve to determine a change in the scanning speed, and the amount of scanning movement is detected to perform distortion correction. This is because, when the speed change is estimated from the relationship between the image input position and the scanning speed, since the image input time for each line in each region between the image input positions is different, the speed estimation corresponding to the actual speed change is performed. Since it cannot be performed, the speed is converted into a scanning speed for each image input time to estimate the speed. According to this, it is possible to estimate the scanning speed close to the actual scanning speed change, and to perform distortion correction using the scanning speed estimated in this way, it is possible to perform extremely accurate distortion correction.

As means for converting a scanning speed for each image input position into a scanning speed for each image input time, one line different between each area between the image input positions may be used. The image input time for each line is set to the same image input time for each line in each area by matching the other image input times to the reference image input time based on the image input time of the fastest speed. Since the input image in an area other than the reference time is temporally stretched in the sub-scanning direction for convenience, the image input position, the image input time for this image input position, and the scanning speed are made to correspond. The temporal scale can be made the same between the respective regions, and the scanning speed can be estimated close to the actual change in the scanning speed. By performing distortion correction it has, can be performed very accurate distortion correction.

Further, when converting the scan speed change for each image input position into the scan speed change for each image input time, each image input position is made to correspond to the image input time. For this purpose, the input image is not changed, only the correspondence data is created. This means that if an image of an area input at the fastest speed is actually created by adding an image of another area, a large-capacity memory for storing the image is required. However, according to the present invention, the input image itself is not converted, and only the correspondence data for making each image input position correspond to the image input time is created, so that a large-capacity memory is not required.

As described above, according to the present invention, the scanning speed of the image input device is determined, an image is input at a reading timing corresponding to the scanning speed, and input from one sensor together with the scanning speed data. Since the transmitted image data is sent to the image processing unit, the data transfer speed can be low, and
Memory can also be minimized. Furthermore, since the image is read at the reading timing according to the scanning speed, if the length of the input image can be specified, the amount of memory to be prepared can be determined, and the memory use efficiency can be improved. . Further, since the data amount of the image to be processed can be reduced, high-speed distortion correction processing can be performed.

Further, since the change of the scanning speed is estimated and the image reading timing is set, it is possible to secure a necessary information amount of the image, and it is possible to accurately obtain an image having a desired resolution.

Further, by converting the change in scanning speed with respect to the position of the input image into an image input time and then correcting the distortion of the image, image distortion correction corresponding to the actual change in scanning speed becomes possible. Images can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image input device according to an embodiment of the present invention.

FIG. 2 is a view for explaining a relationship between an image input position and a scanning speed in the embodiment.

FIG. 3 is a diagram showing a relationship between an input image position and a scanning speed for explaining a distortion image correction method in the embodiment.

FIG. 4 is a view showing the relationship between time and scanning speed for explaining a distortion image correction method in the embodiment.

FIG. 5 is a configuration diagram illustrating an example of a two-line sensor type image input unit.

FIG. 6 is a configuration diagram of a conventional image input device.

7 is a view for explaining the relationship between an image and a scanning speed in the conventional image input device shown in FIG.

FIG. 8 is a diagram specifically illustrating a reading operation in the conventional image input device shown in FIG.

FIG. 9 is a configuration diagram of another conventional image input device.

[Explanation of symbols]

 101 first sensor 102 second sensor 103 read timing signal generation circuit 104 first feature extraction circuit 105 second feature extraction circuit 106 Buffer 107: Feature comparison circuit 108: Speed determination circuit 109: Personal computer 110: Image correction means 111: Data synthesis circuit 201: Original image 202: Input image

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-60-5383 (JP, A) JP-A-3-286679 (JP, A) JP-A-3-132888 (JP, A) JP-A-56-58 108175 (JP, A) JP-A-63-273175 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06K 9/00-9/82 G06T 1/00 H04N 1/107

Claims (13)

(57) [Claims] An image input section and an image processing section for correcting image data read by the image input section, as a one-dimensional image sensor for reading characters, figures, etc. on a document illuminated by a light source. The image sensor has first and second sensors, and reads the image data read at a position where the first sensor moves forward in the manual feeding direction from the time until the second sensor reads the image data. An image input device that detects a scanning speed and corrects image data according to the scanning speed, wherein: a first feature extracting unit for extracting a feature from the image data input by the first sensor; A second feature extraction unit for extracting a feature from the image data input by the second sensor, a feature buffer for temporarily storing the feature data extracted by the first feature extraction unit, Stored in the feature buffer A feature comparing unit that compares the feature data with the feature data extracted by the second feature extracting unit to determine a correspondence; and an image input position based on the correspondence determined by the feature comparing unit. Speed judgment means for judging a scanning speed with respect to, reading timing signal generating means for controlling the reading timing of the first and second sensors in accordance with the scanning speed judged by the speed judgment means, and the first sensor A data synthesizing means for synthesizing the image data read by one of the second sensors and the result of the speed judging means; and an input image based on the data synthesized by the data synthesizing means. An image input device, comprising: image correction means for performing distortion correction. 2. The image input unit according to claim 1, wherein said first and second sensors, first and second feature extracting means, feature buffer, feature comparing means, speed determining means, reading timing signal generating means, and data synthesizing means are provided on an image input side. 2. The device according to claim 1, wherein
The image input device according to the above. 3. The data synthesizing means according to claim 1, wherein said image data is read between image data for each line of image data read by one of said first sensor and said second sensor. 2. The image input device according to claim 1, wherein a control signal indicating that information different from the data exists and scanning speed data determined by the speed determining means are inserted as necessary. 4. The reading timing signal of said reading timing signal generating means sets in advance a maximum scanning speed change amount with respect to a scanning movement amount from a certain scanning speed detection position to a next scanning speed detection position. 2. An image input apparatus according to claim 1, wherein said image input device is determined from scanning speed data at a scanning speed detecting position in consideration of said maximum scanning speed change amount. 5. The setting of the maximum scanning speed change amount with respect to the scanning movement amount from a certain scanning speed detection position to the next scanning speed detection position is twice the scanning speed change amount with respect to a predetermined scanning movement amount. The image input device according to claim 4, wherein 6. A one-dimensional image sensor which comprises an image input section and an image processing section for correcting image data read by the image input section, and reads characters, figures, etc. on a document illuminated by a light source. The image sensor has first and second sensors, and reads the image data read at a position where the first sensor moves forward in the manual feeding direction from the time until the second sensor reads the image data. In an image input method for detecting a scanning speed and correcting image data according to the scanning speed, a feature of the image data input by the first sensor is extracted by a first feature extracting unit. Then, the extracted feature data is stored in a feature buffer, and the features of the image data input by the second sensor are extracted by a second feature extracting means and stored in the feature buffer. Feature data and the second feature The feature data extracted by the output means are compared by the feature comparison means to determine the correspondence, and then the scanning speed for the image input position is determined by the speed determination means based on the correspondence determined by the feature comparison means. The reading timing signal generating means controls the reading timing of the first and second sensors in accordance with the determined scanning speed, and further, one of the first sensor and the second sensor. An image input method comprising: synthesizing image data read by a sensor with a result of the speed determination means by data synthesizing means; and performing distortion correction of an input image based on the synthesized data. . 7. The image input unit side, wherein the first and second sensors, the first and second feature extracting means, the feature buffer, the feature comparing means, the speed determining means, the reading timing signal generating means, and the data synthesizing means are provided. 7. The device according to claim 6, wherein
Image input method described. 8. The image data synthesizing means according to claim 1, wherein said image data is read between one line of image data read by one of said first sensor and said second sensor. 7. The image input method according to claim 6, wherein a control signal indicating that information different from the data exists and scanning speed data determined by the speed determining means are inserted as necessary. 9. The reading timing signal of the reading timing signal generating means sets in advance a maximum scanning speed change amount with respect to a scanning movement amount from a certain scanning speed detection position to a next scanning speed detection position. 7. The image input method according to claim 6, wherein the determination is made from scanning speed data at the scanning speed detection position in consideration of the maximum scanning speed change amount. 10. The setting of the maximum scanning speed change amount with respect to the scanning movement amount from a certain scanning speed detection position to the next scanning speed detection position is the double of the scanning speed change amount with respect to a predetermined scanning movement amount. 10. The image input method according to claim 9, wherein: 11. A one-dimensional image sensor for reading characters, figures, etc. on a document illuminated by a light source, comprising an image input unit and an image processing unit for correcting image data read by the image input unit. It has first and second sensors,
The scanning speed is detected from the time until the second sensor, which reads the image data read at a position where the first sensor moves forward in the hand-feed direction, is read, and the scanning speed is determined according to the scanning speed. In the image correction method in the image input device for performing the correction processing on the image data, the characteristics of the image data input by the first sensor are extracted by the first characteristic extraction means, and the extracted characteristic data is extracted. The second data is stored in a feature buffer, and the feature of the image data input by the second sensor is extracted by a second feature extracting means, and the feature data stored in the feature buffer and the second feature data are extracted. After comparing the feature data extracted by the second feature extracting means with the feature comparing means to determine the correspondence, the speed determining means scans the image input position based on the correspondence determined by the feature comparing means. Speed The reading timing signal generating means controls the reading timing of the first and second sensors in accordance with the determined scanning speed, and further, any one of the first sensor and the second sensor The image data read by the sensor and the result of the speed determination means are combined by a data combining means and transferred to an image processing unit. The image processing unit inputs the data at a read timing based on the scanning speed. The scanning speed for each position of the input image is converted into the scanning speed for each image input time, and then the scanning speed for each image input time is approximated by a curve to find a change in the scanning speed on the way. An image correction method in an image input device, wherein distortion correction is performed by obtaining a movement amount. 12. A means for converting a scanning speed for each image input position into a scanning speed for each image input time, wherein an image input time for each line, which differs between regions between the image input positions, is set as: By matching the image input time for each line in the slowest reading area with the image input time for each line in the area where the fastest reading has been performed, the time axis indicating the image input time is the same scale. 12. The image correction method according to claim 11, wherein each position of the input image is associated with an image input time for each position of the input image and the scanning speed. 13. When converting a change in scanning speed for each image input position into a change in scanning speed for each image input time, correspondence data for associating each image input position with an image input time is provided. 13. The image correction method according to claim 12, wherein the input image is only created and the input image is not changed.