Image reading device

Abstract

To reduce a processing time required for edge detection.SOLUTION: An image reading device includes reading means for reading an image of a document, moving means for moving the reading means and the document relatively, first end detection means for detecting a first end in the direction of movement by the moving means from image data, second end detection means for detecting a second end, which is the end opposite the first end, from the image data, and contour detection means for detecting the contour of the document on the basis of the first end and the second end, and the second resolution, which is the resolution of the image data when detecting the second end, is lower than the first resolution, which is the resolution of the image data when detecting the first end.SELECTED DRAWING: Figure 10

Description

[Technical field]

[0001] The present invention relates to an image reading device that reads an image of a document. [Background technology]

[0002] Image reading devices that read images of documents are generally equipped with a contour detection function to detect the contours of the document from the scanned image. By detecting the contours of the document, it becomes possible to calculate the inclination of the skewed document and rotate the scanned image so that it is straight (hereinafter referred to as skew correction).

[0003] For example, a method for detecting the outline of a document from the shadow of the document's edge (hereinafter referred to as the edge) in a scanned image has been conventionally known. Patent Document 1 proposes a technique for detecting the outline of a document by scanning images of both sides of the document and using the top edge of the document detected from the scanned image of the front side and the bottom edge of the document detected from the scanned image of the back side.

[0004] When reading an image of a document, the document is read in the sub-scanning direction from the top edge to the bottom edge. Therefore, the timing of detecting the bottom edge from the image read from the back side is slower than the timing of detecting the top edge from the image read from the front side. In other words, the outline of the document cannot be detected until the detection of the bottom edge is complete, and processing that requires outline information of the document, such as skew correction, cannot be started. Therefore, in order to prevent a decrease in reading speed performance, it is very important to complete the detection of the bottom edge as quickly as possible.

[0005] Here, Patent Document 2 proposes a technology for shortening the processing time required for edge detection. Specifically, after reading an image of a document, edge detection image data with a lower resolution than the image data to be output to the user is generated, and the edges of the document are detected using the edge detection image data. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2010-153975 A [Patent Document 2] Patent No. 6311641 Summary of the Invention [Problem to be solved by the invention]

[0007] However, when detecting the top and bottom edges of a document using edge detection image data with a lower resolution than the image data output to the user, the edge detection accuracy is affected to some extent. In particular, when rotating a scanned image for skew correction, the skew of the document is calculated so that the top edge of the document is aligned with the top edge of the image data to be output. Therefore, if the detection accuracy of the top edge decreases, the accuracy of the skew correction is significantly affected.

[0008] The present invention is intended to solve the above problems, and has an object to provide a technique that can reduce the time required for document edge detection processing as much as possible while suppressing a decrease in the detection accuracy of the document contour. [Means for solving the problem]

[0009] In order to solve the above problems, the present invention provides A reading means for reading an image of a document; a moving means for moving the reading means and the document relative to each other; a first end detection means for detecting a first end in a moving direction of the moving means from the image data; a second end detection means for detecting a second end, which is an end opposite to the first end, from the image data; an edge detection means for detecting an edge of a document based on the first edge and the second edge; Equipped with The present invention is characterized in that a second resolution, which is the resolution of the image data when the second end is detected, is lower than a first resolution, which is the resolution of the image data when the first end is detected. Effect of the Invention

[0010] According to the present invention, it is possible to suppress a decrease in the accuracy of detecting the document contour while shortening the time required for document edge detection processing as much as possible, thereby suppressing a decrease in performance of the reading speed due to the document contour detection processing. [Brief description of the drawings]

[0011] [Figure 1] Schematic cross-sectional view of an image reading device [Diagram 2] Block diagram of the control system of the image reading device [Diagram 3] 1 is a cross-sectional view of an image reading sensor unit in an image reading device according to the present invention; [Figure 4] Schematic diagram of a system in which an image reading device and a host computer are connected [Diagram 5] A diagram for explaining the positional relationship between a light source and an image reading sensor. [Figure 6] A diagram to explain the shadows that appear on an image when a document is scanned. [Figure 7] Diagram and graph for explaining the effect of the shadow of the document on the brightness of the scanned image [Figure 8] FIG. 1 is a diagram for explaining a method for detecting edges of a document. [Figure 9] Flowchart of document edge detection process [Figure 10] FIG. 13 is a diagram for explaining a process for detecting the outline of a document from the edges of the document; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS EXAMPLES

[0012] <Basic configuration of image reading device> 4 shows an example of an image reading system in which an image reading device 1 of the present invention and a host computer 2 are connected by a general-purpose communication cable 3. The image reading device of the present invention is configured in this manner, but is not limited to this. For example, the image reading device 1 may be configured so that the host computer 2 is included inside the housing of the image reading device 1.

[0013] 1 is a cross-sectional view for explaining a conveying path of an image reading device according to the present invention. The image reading device has a mechanism for separating originals 101 one by one from a stack of originals on an original table 102 by a paper feed roller 106 or the like, sending the originals to a conveying path 108, and conveying the originals along the conveying path by a conveying roller 107. The image reading device also reads images from originals passing in front of image reading sensor units 104 and 105, and discharges the originals to a paper discharge unit 103.

[0014] <Hardware configuration of image reading device> 2 is a circuit block diagram of an image reading device provided in the image reading device 1 according to the present invention, and the image reading device 1 includes a CPU (Central Processing Unit) 202, a RAM (Random Access Memory) 203, a ROM (Read Only Memory) 204, an external I / F (Interface) 205, a reading control unit 206, a motor driver 207, and further includes a stepping motor 208, an image reading sensor 209, an illumination LED 210, and a DSP (Digital Signal Processor) 220 having a resolution conversion unit 211, all of which are mounted on a control circuit board 200. The CPU 202, the RAM 203, the ROM 204, the reading control unit 206, the motor driver 207, and the external I / F 205 are connected to each other via a system bus 201.

[0015] The CPU 202 controls the entire image reading device 1. The RAM 203 provides a work area for the CPU 202, and at the time of image scanning, functions as a buffer memory for temporarily storing image data obtained from the reading control unit 206. The ROM 204 stores various programs, initial setting values, and the like.

[0016] The reading control unit 206 has an A / D converter and a shading correction processing unit (not shown) for controlling the image reading sensor 209 and processing the image signal from the image reading sensor 209. The reflected light of the light irradiated onto the document surface by the illumination LED 210 is read by the image reading sensor 209, the read analog image signal is A / D converted, shading correction is performed, and the image data is written to the RAM 203.

[0017] The feed roller 106 is driven by a gear, a belt, a motor, and a motor driver (not shown), and is controlled by the CPU 202. The transport roller 107 is connected to a stepping motor 208 by a gear and a belt (not shown). The stepping motor 208 is controlled by the CPU 202 through the motor driver 207.

[0018] <Scanning process> When the user operates the host computer 2 to start image reading, the paper feed roller is driven, and then the registration sensor 109 detects that the document has arrived just before the image reading sensor, and after a certain period of time, the illumination lamps 302 and 307 are turned on, and at the same time, the image reading sensors 303 and 308 are operated to read the image. The image signal output by reading the image is converted into digital image data by the reading control unit 206. Furthermore, the image data that has been subjected to image processing such as shading correction, gamma correction, edge emphasis processing, etc. to make the image more suitable is sent to the host computer 2, which is a generally performed scanning operation.

[0019] <Image reading sensor unit> Fig. 3 is an enlarged cross-sectional view of the front image reading sensor unit and the back image reading sensor unit of this image reading device. As shown in Fig. 3, the front image reading sensor unit 104 is a unit sealed by a housing 301 and glass 304. Inside the image reading sensor unit 104, there are provided an illumination lamp 302 (light source), an image reading sensor 303, and a background plate 305 that serves as a background member when the back side reading sensor unit reads the back side of the document. Meanwhile, the back side image reading sensor unit 105 has a similar configuration. In Fig. 3, the description will be given assuming that an image is read from a document transported from right to left.

[0020] <How shadows on the edges of a document are read> FIG. 5 shows the positional relationship between the light source (illumination LED) and the image reading sensor in this image reading device. In FIG. 5, if the document is transported from right to left as in FIG. 3, the image reading sensor (front image reading sensor) first reads the reflected light from the background member before the document arrives. As the document approaches, the light irradiated from the illumination LED to the background member is blocked, and the shadow of the document is cast on the background member, so the image reading sensor reads the shadow. When the document is further transported, the image reading sensor reads the document surface. That is, when an image is read from the document shown on the left in FIG. 6, a certain amount of dark shadow is formed above the top edge of the document image in the front read image as shown on the right in FIG. 6. The operation of the front reading sensor unit to read an image in FIG. 3 can be explained as above using FIG. 5.

[0021] On the other hand, the operation of the back side reading sensor unit to read an image in Fig. 3 corresponds to the image reading operation when the document is transported from left to right in Fig. 5. That is, after the document has passed the image reading sensor, the document blocks the light irradiated from the illumination LED to the background member, so that a certain amount of dark area due to a shadow is formed below the bottom edge of the document image in the back side read image.

[0022] In this way, in this image reading device, the front side illumination lamp 302 and image reading sensor 303 are arranged as shown in Fig. 3 and the illumination light is diagonally emitted so that a shadow is read on the top edge of the document on the front side. Also, the back side illumination lamp 307 and image reading sensor 308 are arranged as shown in Fig. 3 and the illumination light is diagonally emitted so that a shadow is read on the bottom edge of the document on the back side.

[0023] <How to detect edges of a document> Fig. 7(a) shows an image with a shadow formed at the upper edge, and Fig. 7(b) shows a density graph at the position indicated by A. A method for detecting the edges of a document will be explained using the graph in Fig. 8 and the flow chart in Fig. 9.

[0024] Fig. 8 shows an example of a density graph of an original like Fig. 7, but this graph shows both an original with a bright surface (light original) such as a white sheet of paper and an original with a dark surface (dark original). The slice level th and the shadow length h are fixed values ​​determined in advance.

[0025] The flowchart in FIG. 9 shows a process for searching for an edge of an original image in a read image by scanning in the sub-scanning direction from the top end position of the read image in a certain row in the main scanning direction.

[0026] First, the CPU 202 performs initialization in step S901. Here, y is the pixel position in the sub-scanning direction, and edge is the top edge position of the detected document image. When the value of edge is -1, it indicates that there is no edge of the document image on that vertical line.

[0027] In order to perform the process of step S903 by shifting one pixel at a time in the sub-scanning direction, a loop is formed in which the processes of steps S902 to S904 are repeated. In step S902, the loop is repeated as long as y is smaller than the image height (if Yes), and the process of this row in the main scanning direction is terminated when y becomes equal to or greater than the image height. In step S903, pix(y) is the image brightness at the y position (the brighter the image, the larger the value). When pix(y) is large and bright, it is determined that the image is a background image (if Yes), and the process proceeds to step S904 to continue the loop process. Then, when pix(y) becomes equal to or less than th (position P in FIG. 8), it is determined that the top end of a shadow formed outside the document image in the read image has been detected, and the value of y is stored in the top end position shadow of the shadow (step S905). In addition, the process of detecting the rear end of the shadow, i.e., the top end edge of the document, is performed in the next step, but the value of shadow+h is stored as an initial value in the top end edge position edge of the document image in case the top end edge of the document cannot be found. Here, h is a preset value as the length of the shadow, and is a fixed value that is geometrically determined from the positional relationship between the image reading sensor of the image reading device and the light source LED, and the distance between the document transport path and the background member.

[0028] The loop that repeats the processing from step S906 to step S908 forms a loop process for repeatedly performing the processing of step S907 by shifting one pixel at a time in the sub-scanning direction, and the processing is continued until y is equal to or greater than the smaller of the height value and the value of shadow+h. When the image brightness pix(y) exceeds th in step S907, it is determined that y has reached the top edge of the document image (position Q in FIG. 8), and the process proceeds to step S909. If the image brightness pix(y) is equal to or less than th in step S908, the process proceeds to step S908 and continues the loop processing. In step S909, y is stored in the top edge position edge, and the process ends.

[0029] 8, the determination in step S907 is YES at the top edge of the document image in the read image, but in the case of a dark document, the determination in step S907 is always NO. In this case, the loop process ends without performing the process in step S909, so the value of the top edge position edge remains shadow+h.

[0030] In other words, in the case of a light original, the top edge of the original is detected as is, and in the case of a dark original, the top edge of the shadow formed on the original image is detected and the distance h from this point is used for calculation, thereby making it possible to identify the top edge position of the original regardless of the brightness of the original.

[0031] <How to detect the outline of a document> Next, the flow of detecting the outline of a document, which is a feature of the present invention, will be described with reference to FIG.

[0032] When images on both sides of a document are scanned, a scanned image of the front side as shown in Fig. 10(a) and a scanned image of the back side as shown in Fig. 10(b) are obtained. Here, it is assumed that the scanned images in Fig. 10(a) and Fig. 10(b) are scanned at a resolution of 300 dpi (first resolution) in both the main scanning direction and the sub-scanning direction according to the scan settings. In addition, the scanned image of the front side in Fig. 10(a) includes a shadow 1000 at the top edge of the document, and the scanned image of the back side in Fig. 10(b) includes a shadow 1001 at the bottom edge of the document.

[0033] First, the document edge detection process described above is performed on all columns of the entire image using the flowchart of FIG. 9 for the scanned image of the front side of FIG. 10(a), and the top edge 1002 is detected as shown in FIG. 10(c).

[0034] On the other hand, for the process of detecting the bottom edge from the scanned image of the back side, image data with a lower resolution than the scanned image of the front side in Fig. 10(a) used for detecting the top edge is used. Therefore, the scanned image of the back side in Fig. 10(b) is subjected to a resolution conversion process by the resolution conversion unit 211, and is converted into image data with a resolution (second resolution) of 150 dpi from 300 dpi in both the main scanning direction and the sub-scanning direction as shown in Fig. 10(d). In other words, this image data in Fig. 10(d) becomes the image data for detecting the bottom edge.

[0035] Next, a process similar to the above-described upper edge detection process is executed on the resolution-converted rear side image data of FIG. 10(d), thereby detecting a lower edge 1003 as shown in FIG. 10(e).

[0036] Furthermore, the data of the bottom edge 1003 in FIG. 10(e) is again subjected to resolution conversion processing by the resolution conversion unit 211, and is converted into data of the bottom edge 1004 equivalent to 300 dpi in both the main scanning direction and the sub-scanning direction, as shown in FIG. 10(f).

[0037] Next, by matching the top edge 1002 in Figure 10(c) detected from the scanned image of the front side with a mirror image of the bottom edge 1004 in Figure 10(f) detected from the scanned image of the back side, image data containing both the top edge and bottom edge as shown in Figure 10(g) is generated.

[0038] Furthermore, by connecting the ends of the upper and lower edges of Fig. 10(g) with an auxiliary line, the outline of the entire document can be obtained as shown in Fig. 10(h). To create image data representing the outline of the document image on the back side, the image data of Fig. 10(h) may be inverted, or the image of Fig. 10(c) may be inverted horizontally and combined with the image of Fig. 10(f) and the ends of the upper and lower edges may be connected with an auxiliary line.

[0039] Based on the outline (corresponding coordinates) of the entire document obtained in this manner, an image is cut out and the inclination is corrected for the image to be output as read image data.

[0040] <Summary> As described above, image data for detecting the bottom edge, which has a lower resolution than the image data for detecting the top edge, is generated from the image data for detecting the back side. Then, the image data for detecting the bottom edge is used to detect the bottom edge of the document.

[0041] As a result, by maintaining the resolution of the scanned image of the front side used for upper edge detection at the first resolution, it is possible to suppress a decrease in the detection accuracy of the document outline, while by performing lower edge detection with image data for lower edge detection whose resolution is reduced to the second resolution, it is possible to reduce the time required for document edge detection processing as much as possible. In other words, it is possible to suppress a decrease in the performance of the reading speed due to the document outline detection processing.

[0042] As an example of a method for detecting the bottom edge using image data having a lower resolution than the image data used for detecting the top edge, the resolution of the read image of the back side is reduced by the resolution conversion unit 211 before detecting the bottom edge, but the present invention is not limited to this.

[0043] For example, in S904 and S908 of the flowchart in Fig. 9, the edge position is detected while shifting the pixel position y in the sub-scanning direction by one pixel at a time, but edge detection may be performed while shifting this by two pixels at a time, thereby performing processing equivalent to edge detection processing at half the resolution in the sub-scanning direction. Also, with regard to the main scanning direction, instead of performing the edge detection processing of the flowchart in Fig. 9 for all columns of the entire image, it is also possible to perform processing equivalent to edge detection processing at half the resolution in the main scanning direction by performing it for every other column, for example.

[0044] In other words, if a lower end edge detection process equivalent to a lower resolution than the upper end edge detection process is also executed, the same effect can be obtained.

[0045] In addition, in this embodiment, a device configuration in which a shadow occurs at the top edge of the front image and at the bottom edge of the back image has been described, but the present invention is not limited to this. For example, even in a device configuration in which a shadow occurs at the bottom edge of the front image and at the top edge of the back image, the same effect can be obtained by detecting the bottom edge using image data of the front image that has a lower resolution than the back image used for detecting the top edge.

[0046] In this embodiment, both the upper and lower shadows may be detected from only one of the front and back images. For example, in the explanation of Fig. 6, it was explained that a certain amount of dark shadow is formed only at the upper part of the top edge of the document image in the front read image due to the shadow being read, but depending on the configuration of the front read sensor unit, a light shadow may also be formed on the downstream side. In this case, it is possible to detect the contour using the downstream shadow, and in that case, the contour may be detected using an image with low resolution using the method described above.

[0047] Furthermore, if the resolution (first resolution) specified in the read settings when reading an image is lower than the preset third resolution, the downstream shadow may be detected from the image data of the first resolution as is, without converting the resolution to image data of the second resolution and then detecting the downstream shadow. For example, if 300 dpi is preset as the third resolution, and 150 dpi is specified as the first resolution in the read settings, the downstream shadow may be detected at the original resolution.

[0048] In this embodiment, the outline detection process is described as being executed by the CPU 202 based on the processing program in the ROM 204, but the present invention is not limited to this. For example, the outline detection process may be executed by a hardware module other than the CPU 202. The same effect can also be obtained with an external device such as the host computer 2 executing the outline detection process.

[0049] 2, the embodiment has been described in which the resolution conversion unit 211 is provided in a DSP provided separately from the control circuit board 200, but the present invention is not limited to this. For example, the DSP may be configured to be mounted on the control circuit board 200, or the resolution conversion unit 211 may be provided as a configuration separate from the DSP. For example, the resolution conversion unit may be provided as a function of the CPU 202.

[0050] Furthermore, other processes may also be configured to be performed by the DSP. For example, a contour detection processing unit that executes the contour detection process described in this embodiment may be provided in the DSP, and the contour detection process may be performed after the resolution conversion process is performed in the DSP. Note that other post-stage processes such as skew correction may also be configured to be executed by the DSP, and these post-stage processes may be executed by hardware modules other than the CPU 202 and the DSP 220.

[0051] In this embodiment, the conveying roller 107 conveys the document to the front image reading sensor unit 104 and the back image reading sensor unit 105 (reading means), but the present invention is not limited to this and may be an image reading device in which the document is placed on a document table such as transparent glass and the image reading sensor unit is moved to read the image of the document. In other words, it is sufficient if the document is moved relatively to the reading means by the moving means, and in either case, the side that is read first is the upstream side (first end in the moving direction), and the side that is read later is the downstream side (second end in the moving direction). [Explanation of symbols]

[0052] 1 Image reader 2. Host Computer 3 Communication Cable 101 Manuscripts 104 Surface image reading sensor unit 105 Back side image reading sensor unit 202 CPU 203 RAM 204 ROM 205 External Interface 206 A reading control unit that controls the image reading sensor 209 Image reading sensor 210 Light source (lighting LED) 211 Resolution conversion unit

Claims

1. A reading means for reading an image of a document; a moving means for moving the reading means and the document relative to each other; a first end detection means for detecting a first end in a moving direction of the moving means from the image data; a second end detection means for detecting a second end, which is an end opposite to the first end, from the image data; an edge detection means for detecting an edge of a document based on the first edge and the second edge; Equipped with An image reading device characterized in that a second resolution, which is the resolution of the image data when the second end is detected, is lower than a first resolution, which is the resolution of the image data when the first end is detected.

2. the reading means includes a first reading means for reading first image data which is image information on a first surface of the document, and a second reading means for reading second image data which is image information on a second surface which is a reverse surface of the first surface of the document, 2. The image reading device according to claim 1, wherein the first end detection means detects the first end from the first image data, and the second end detection means detects the second end from the second image data.

3. The first reading means and the second reading means each have a light source that irradiates light obliquely and a light receiving unit that receives light that is reflected by a document from the light source to obtain an image, 3. The image reading apparatus according to claim 2, wherein the first reading means and the second reading means irradiate the light from the light source in opposite directions in the transport direction.

4. The contour detection means 2. The image reading device according to claim 1, wherein the contour of the document is detected by connecting an end of the detected first end with an end of a mirror image obtained by inverting the detected second end.

5. A resolution conversion means for converting the resolution of image data, The image reading device according to claim 2, characterized in that the second end detection means generates second end detection image data by converting the second image data to a resolution lower than that of the first image data using the resolution conversion means, and detects the second end from the second end detection image data.

6. The image reading device according to claim 2, characterized in that, when detecting the second end from the second image data, the second end detection means processes every other column or row to be detected, thereby detecting the second end from image data that substantially corresponds to a lower resolution than the first image data.

Images