JPH03170924A - Original size detecting device - Google Patents

Abstract

PURPOSE:To facilitate detection processing and simplify circuit constitution for that by displaying a specific function pattern for an original presser, band by band, and finding original size from the change point of the total number of original area pulses and function pattern pulses in an image memory. CONSTITUTION:An image processing part 5 performs a process wherein image area pulses are generated from a signal inputted to the image memory 4, a process wherein the address (coordinate) at the point of time of the change in the total number of image area pulses appearing in one scan, band by band, and the number of function pattern pulses is stored, and a process wherein the original size is detected. Then the coordinate data at the time of the change in the total number of both pulses appearing in one scan by the bands, the starting and ending changes in the total number are based upon original area pulses and correspond to diagonal points of the original, and the change in the total number obtained between them corresponds to the intersection of the original and function pattern. Then the inclination of the original is calculated from the coordinates of the intersection and the size of the original and the original size is determined according to the inclination and the original size. Thus, the original size is detected by the simple processing method with the simple circuit constitution.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a device for detecting the document size of a placed document, which is suitable for application to an image forming apparatus such as a digital PPC.

[Conventional technology]

Conventionally, in some cases, such as copying, etc., an image is formed from an image signal of a placed original obtained by surface scanning (main scanning and sub-scanning) with an optical system.
If the document is not placed in an appropriate state, there is a problem that the document image may be misaligned or tilted, and furthermore, a portion of the document may be cut off and the image may be imprinted on the transfer paper.

For this reason, in recent years, apparatuses have been provided that detect the document and its placement state, perform image correction, and form the above-mentioned image.

As a method for detecting the size of a document placed on a document, a method has been proposed in which a predetermined pattern is provided on the document holder so that when the document is scanned by an optical system, the document area and the non-document area on which the pattern is drawn are distinguished. (Unexamined Japanese Patent Publication No. 55-490
No. 68, JP-A-56-120264). That is, in the method described in Japanese Patent Application Laid-Open No. 55-49068, a periodic and regular pattern is provided on the document presser, and this pattern is detected by a filter means to recognize the document presser area. This is how it was done.

Also, in the method described in Japanese Patent Application Laid-Open No. 56-120264, a unique pattern is provided for the document presser, and this pattern is detected from the captured image content through image processing to identify the document presser area. After this process is completed, each corner of the document is determined, and the size of the document is detected from this.

[Problem to be solved by the invention]

By the way, in the former technique, although it is possible to obtain the document area, there is no description about detecting the document size.

In addition, in the latter technology, the document size is detected from each corner of the document, so when determining each corner, the image memory must be scanned again to determine each corner from each pixel signal. The processing will take a lot of time, and the circuit configuration for the analysis processing will also become complicated.

Furthermore, in each of the above-mentioned conventional techniques, since the pattern on the document holding surface is relatively fine, similar patterns are often drawn on the document, and in such cases, there is a high possibility of misdetecting the document size. .

The present invention has been made in view of the above, and an object of the present invention is to provide a document size detection device that detects the document size with a relatively simple circuit configuration and processing method.

[Means to solve the problem]

The present invention provides a document size detection device that temporarily stores an image signal read by surface scanning using an optical system in an image memory and detects the document size from this image signal, which has at least a portion that overlaps with the placed document, and n function patterns are displayed on the mirror sheet so that they can be read by the optical system so that one pattern is included in each of the n bands provided in one scanning direction on the scanning surface by the optical system. a first storage means that stores coordinate data of each of the function patterns; an image processing means that creates a document area pulse from the captured image signal; and a document area within each of the bands within one scan. a second storage means for storing coordinate data at a point in time when the total number of pulses and function pattern pulses changes; and a document is stored from coordinate data that matches the coordinates of the function pattern among the coordinate data stored in the second storage means. A first calculation means for calculating the inclination of the document size, coordinate data at the first and last change points of the total number among the coordinate data stored in the second storage means, and a calculated value of the first calculation means. and second calculation means for calculating.

Further, according to the configuration of claim 2, when n is 3 or more, the coordinate data used for calculation by the first calculation means is selected from different bands according to the band in which the total number changes first. I have to.

[Effect]

According to the above configuration, when the optical system performs surface scanning in the main scanning and sub-scanning directions with the original placed thereon, there is a scanning line from which the Rm signal and the function pattern pulse are obtained. This original signal is image-processed into original area pulses using image continuity and stored coordinate data of the function pattern. Therefore, document area pulses and function pattern pulses appear on such scanning lines.

Further, n function pattern pulses are generated corresponding to each of the set n bands. Then, the coordinate data at the time of change in the total number of both pulses appearing in the scanning line for each band is captured.

The first and last changes in this total number are due to document area pulses and correspond to diagonal points of the document.

On the other hand, the change in the total number obtained during that time corresponds to the intersection between the manuscript and the function pattern.

The inclination of the document is calculated from the coordinates of this intersection point, and the size of the document is further determined from this inclination and the coordinates of the diagonal point of the document.

According to claim 2, when n is 3 or more, three or more of the above-mentioned intersection points can be obtained. The inclination can be calculated from the two intersection points, and in this case, the direction of inclination of the document is determined based on which band the first corner of the document was obtained, and the position is placed on one side of the document according to the determination result. Calculate the inclination of the document by changing two points with high probability. This change ensures that two points on the l side of the document are selected and the inclination of the document is determined.

〔Example〕

FIG. 2 is an overall configuration diagram showing an embodiment of an image forming apparatus to which the present invention is applied.

In the figure, a contact glass 9 on which a document is placed is disposed at the top of an image forming apparatus main body 8, and a document presser 10 is provided on the top surface of the contact glass 9 so as to be openable and closable. The lower surface of the document presser 10 is coated with a mirror surface, for example.

An optical system L is provided below the contact glass 9 and is configured to reciprocate in parallel with the contact glass 9 in the direction of arrow A and scan the placed original during this forward movement.

The optical system L includes, for example, a halogen lamp 11 that emits white light, a reflecting mirror 12, a self-occurring lens 13 that focuses reflected light from the document surface and a function pattern described below, a CCDI that receives the focused light, and the optical system L. It consists of a drive mechanism (not shown) that reciprocates. Although not shown in the figure, one filter is arranged in front of the sensor array constituting CCD 1, with the same dimensions as the pitch of each sensor array and sequentially colored R, G, and B. ing.

This optical system L is driven to start scanning from the home position shown in the figure (indicated by a solid line), scan to a predetermined position (indicated by a chain line), and then return to the home position again. It is done like this.

Note that the signals of the original image captured for each RGB color by this scanning are guided to a signal processing means (not shown), and level-converted there into four colors: cyan, magenta, yellow, and black. There is. Then, the processed color signals are sequentially outputted as optically modulated laser light from the laser emitting means 7 for each color, and guided to the photosensitive drum 14, which will be described later.

The photosensitive drum 14 is arranged inside the main body 8, and a charging device 15 is arranged around the photosensitive drum 14.
, a developing device 16, a transfer device 17, a separating device 18, and a cleaning device 119 are provided. The developing device 16
Four developing devices 16a to 16d are arranged side by side in the vertical direction, and the developing devices 16a to 16d are arranged in, for example, N7, magenta, black, and yellow from the top. These developing devices 16a to 16d are fixed to a movable frame 20 that can be raised and lowered, and developer of each color is sequentially supplied to the photoreceptor drum 14 during color copying by means of raising and lowering means (not shown). ing. Each developer 16a
~16d and the laser emitting means 7 are synchronized,
As a result, the developing devices corresponding to the laser beams of each color are sequentially driven so as to face the photosensitive drum {4.

A paper supply mechanism is provided below the main body 8. The paper supply mechanism includes paper feed cassettes 21a and 2lb of different sizes, a manual feed tray 21c1, transport roller pairs 22a to 22d, and a registration roller pair 23, and the paper transported from the registration roller pair 23 is transferred to a transfer drum 24. is supplied to

The transfer drum 24 is disposed close to the photosensitive drum 14 and is controlled to rotate synchronously.

The transfer drum 24 rotates the conveyed paper integrally during the four rotations of the transfer drum 24.

That is, for each of these four rotations, the transfer process for each color is performed four times.

When the four transfer processes are completed, the paper is separated by a separating device 18 and guided to the conveyor belt 25 side.

As a discharge mechanism, a fixing device 26, a discharge port-ra pair 27, and a paper discharge tray 28 are provided downstream of the conveyor belt 25.

FIG. 1 shows a block diagram of a portion that detects the document size. In the figure, 1 is the CCD, 2 is the CCDI
It is a line memory that A/D converts the original signal taken in, temporarily stores it, and sequentially reads it out at a predetermined cycle. 3 is a determination unit in which a predetermined threshold value for comparison is set;
It compares the level of the output signal from the line memory 2, and when the output level is equal to or higher than the threshold value, a high signal is output, and when the opposite occurs, a low signal is output. This determination unit 3 outputs high signals for the entire document surface excluding characters and graphic portions in the document and a portion of a predetermined function pattern on the document presser 10, and outputs low signals for other portions.

Reference numeral 4 denotes an image memory having a capacity corresponding to the scanning surface of the optical system L, which receives high and low signals from the determination section 3. This image memory 4 processes discrete high signals from the document surface that appear within one scanning line into continuous high signals to form a document area signal (hereinafter referred to as document area pulse) from the captured signal. Then, signals are exchanged with the image processing section 5 described below. Then, the document size (including tilt) is determined by performing the processing described below from the image-processed signal.
Based on this data, the image signal in the image memory 4 is updated and rewritten to correct it. The rewritten image signal is read out to the laser emitting means 7. Reference numeral 6 denotes a write/read means for specifying addresses (coordinates) in the main scanning direction and the sub-scanning direction for writing and reading signals to and from the image memory 4.

5 is the process of forming image area pulses from the signals taken into the image memory 4, and the time point at which the total number of image area pulses and function pattern pulses appears within one scan for each band (to be described later) changes. This is an image processing unit that performs processing to store the address (coordinates) of the document and detect the document size. The image processing means 51 executes each of the above-mentioned processes, and has an image memory (not shown) having the same capacity as the image memory 4 for these processes. 52 is a document size ROM in which data regarding the document size is stored in advance; 53 is a function pattern ROM in which main scanning and sub-scanning addresses (coordinates) representing each function pattern or the function formula of the function pattern is stored in advance. It is.

In the above configuration, the document area pulse is formed from the document signal by processing the stored contents of the image memory 4, for example, as follows. That is, first, function pattern pulses are distinguished from one scan's worth of signals sequentially read out from the image memory 4.

The pulses corresponding to this function pattern are distinguished by comparing the address of each function pattern stored in the function pattern ROM 53 with the address of the signal for one scan, and determining that a signal that matches is not an original signal. It is. Then, the remaining discrete signals are treated as original signals and processed to be continuous. Continuous processing gives redundancy to the above discrete degree,
Continuity can be achieved by converting all low levels within that range to high levels. Note that the image processing method is not particularly limited to this method, and any conventional method may be used.

In this way, the continuously processed original area pulses and the original function pattern pulses are written into the image memory in the image processing means 51.

After such image processing, processing for detecting the document size is performed. This process will be explained below with reference to FIGS. 3 to 5 showing the first embodiment and FIG. 6 showing the second embodiment.

The first example is a case where there are three function patterns (n=3). In this function pattern, when the main scanning direction is the y-axis and the sub-scanning direction is the X-axis, y=bSy=ax+b, y=-ax+b, where b is a straight line at the center value in the main scanning direction.

Note that among the above function patterns, the patterns indicated by broken lines have been deleted. This is to ensure that the original area pulse is detected first, even if the placed original is placed somewhat away from the y-axis or placed accurately on the y-axis, as shown in Figures 3 and 4. It is.

Each of these patterns is stored in advance in the function pattern ROM 53 as an address {coordinates (x, y)}.

The image processing means 5 also processes the document area pulses described above for each band separated into three areas in the main scanning direction on the scanning surface by the optical system L, as shown in FIGS. 3 and 4. It is programmed to detect a change point in the total number of pulses and the number of function pattern pulses.

In addition, each band is Band■...y<b-db Band■...b-db<y<b+db Band■...
b+ab<y However, db can be changed by the slope a, and is set to about several cm on the actual scanning plane.

As a result, most of the function pattern of Y=-ax+b is in Band E, the function pattern of y=b is in Band ■, and y=
Most of the function patterns of ax+b are included in band (2). Therefore, changes in the number of function pattern pulses occur within each corresponding band.

FIG. 3 shows the case where the document is tilted counterclockwise, and FIG. 4 shows the case where the document is tilted clockwise.

(1) About Figure 3 Figure 3 (a) is a diagram explaining the relationship between the placed original and the intersection between the placed original and the function pattern, and Figure 3 (b) is a waveform diagram of the output pulse of each band. be.

As mentioned above, the signals written in the image memory in the image processing means 5 are processed by the writing/reading means 6, for example, from address (0, O) to (0.1), (0,2), . . . (0,2
As shown in b), one scanning line is read out, and each scanning line is (1.0)... (1, 2b), (2.0) ・= (
2.2 b) ..., (Xend.0) -(
Xend. It will be done.

The process will be explained below using the flowchart shown in FIG.

In addition, J...the number of changes from the second time onwards in the above total number from 0 to
Takes a value of 3.

h...Takes the value "1" when the first change in the total number occurs in band I, and takes the value "0" when it occurs in band ■.

Further, the above numerical value is set to J = h = 0 by initial setting.

First, in steps SE and S2, during the X-axis scan, determine whether the first change in the above total number occurs in band ■.
A determination is made as to whether this occurs in band ■.

In the scanning of the X=O axis, the entire region is the document presser 21 (the document is somewhat separated from the y axis), and J=h=o.

Then, sequential scanning is performed and when x=x○, ? Mark (
A change in the number of pulses first occurs at Xo, To). This point ■ appears in the band because the original is placed counterclockwise as shown in Figure 3 (YE in step S1).
S). The coordinates (x'0, yo) of this point (2) are stored in a memory (not shown) (step S3). Subsequently, in step S4, h is set from O to 1.

Thereafter, within the range of xO <x<x1, no change occurs in the number of pulses in any band. Therefore, in each scan during this period, the value of J remains O (No in step S7, step SIO, and step S03), and
<x end (NO in step S16)
), the process of moving to step S7 via step S tV is repeated. Waveforms ``■'' to ``■'' in FIG. 3(b) show the state of the output pulse in band ``during this period. In this figure, the shaded pulses indicate document area pulses.

When x=x■ is reached, a new function pattern pulse is generated on the scanning line, as shown by waveform ■ in FIG. 3(b). This function pattern pulse is a function y=? It appears at a point on ax+b, that is, in band I (YES in step S7). Therefore, the process moves to step S8, and the coordinates (Xe, y■)=(xt)-aXl+b) of this point (2) are stored in the memory, and further, J is incremented from 0 to 1 in step S9.

Thereafter, within the range of xl<x<x3, no change occurs in the number of pulses in any band. Therefore, in each scan during this period, the value of J remains 1 (step S7, step S
10 and step S13), x <
Since it is x end (NO in step S tS), the process of moving to step S7 via step S17 is repeated.

When X = X 3 is reached, a new function pattern pulse is generated on the scanning line, and the number of pulses changes, as shown by waveform 3 in FIG. 3(b). This function pattern pulse is the function y=
It appears at a point on ax+b, that is, at band ■ (YES at step S tO). Therefore, the process moves to step S1■, and the coordinates of this point ■ (X3, 73) = (X3
, aX3 +b) is memory? stored and further step S
J is incremented from 1 to 2 in L2.

Thereafter, within the range of X3<X<X2, no change occurs in the number of pulses in any band. Therefore, in each scan during this period, the value of J remains 2 (step S7, step S
10 and step S13), and X<x
! Since it is lId (No in step Sta), the process of moving to step S7 via step S17 is repeated.

When x=x2 is reached, a new function pattern pulse is generated in the scanning line, and the number of pulses changes, as shown by the waveform (2) in FIG. 3(b). This function pattern pulse appears at a point on the function y=b, that is, at band H (YES at step S3). Therefore, step S. The coordinates (x2, y2)=(x2, b) of this point (2) are stored in the memory, and J is incremented from 2 to 3 in step S15.

At this point, x < x end (step S
ta), and J=3 (Y in step S17).
ES), move to step 8 18. h is step S
Because it is 1 as obtained in step 4 (NO in step SE8)
, and then detect the change in the number of pulses in band (■). That is, in the range of X2 <

When x=x4 is reached, the document area pulse is no longer obtained and the number of pulses changes (YES in step SE9), as shown by waveform (2) in FIG. 3(b). Therefore, the coordinates of this point (X4%V4) are stored in the memory (step S20).

In this way, the coordinates of points ■ to ■ are captured.

Next, in step S24, ta is calculated from the coordinates of points ■ and ■.
nθ= (X2 - Xi) / 0'2 Vx)
is calculated to calculate the tilt angle θ. If θ is not O (
(NO in step S26), the process moves to step S27. In step S27, the original signal written in the image memory 4 is corrected so that the tilt angle θ becomes O. Furthermore, in step 328, the dimensions of the original are determined from the obtained inclination angle θ and the coordinates of points ■ and ■, which are the diagonal points of the original. X and Y are determined, and the document size is determined from these dimensions and the document size ROM 52.

Then, a transfer paper cassette having a transfer paper matching the determined original size is selected, and the corrected storage contents of the image memory 4 are read out to the laser emitting means 7 to form an image (step S29).

(2) About Figure 4 Figure 4 (a) is a diagram showing the relationship between the placed original and the intersection of the placed original and the function pattern, and Figure 4 (b) is a waveform diagram of the output pulse of each band. be.

The process will be explained below using the flowchart shown in FIG.

When scanning starts, since the original is placed clockwise as shown in Figure 4, the point ■ where the pulse number first changes appears in the band ■ (YES in step S2).
Then, the coordinates (Xo -.yo) of this point (2) are stored in the memory (step Ss). On the other hand, h remains at 0 since the point ■ is in the band ■ (step Ss).

Subsequently, as the X-axis scan is repeated, the points ■, ■, and ■ appear in the same band as in the case of FIG. 3, respectively, as described above (steps 87 to S7).
.

Then, since h=0 (YES in step Sd.), the process moves to step S21. After this, when the document area pulse is no longer obtained in band I and the number of pulses changes (YES in step S21), the coordinates of this changing point (2) are stored in the memory (step S22).

The waveform ■ in FIG. 4(b) shows the output state of the function pattern pulse and document area pulse at the point ■ in the band. The waveform ■ also represents the output at the time when the document area pulse is no longer obtained at the point ■ in the band ■. show.

Further, waveforms ``■'' to ``■'' indicate the output state of document area pulses between points ``■'' and ``■'' in band ``■'', and waveform ``■'' similarly indicates an output pulse at point ``■'' in band ``■''. The waveform (■) shows the output of the document area pulse and the function pattern pulse at the point (■) in the band (■).

In this way, the coordinates of points ■ to ■ are captured.

Next, in step S25, ta is calculated from the coordinates of points ■ and ■.
The inclination angle θ is calculated by calculating nθ=(X3−X2)/0′3Y2). If θ is not 0 (NO in step S26), the process moves to step S27.

Then, as described above, the original signal written in the image memory 4 is corrected so that the tilt angle θ becomes O, and the dimensions X and Y of the original are determined from the tilt angle θ and the coordinates of points ■ and ■. , the document size is determined from the dimensions and the document size ROM 52. Then, a transfer paper cassette having a transfer paper matching the determined original size is selected, and the corrected storage contents of the image memory 4 are read out to the laser emitting means 7 to form an image (step S29).

Note that when the original is placed parallel to the y-axis (tilt angle θ=O), the intersection points ■, ■, and ■ in each band
In both cases, there is no change in the total number (step S7).
, step S10 and step? Both No on S 13)
, J=O remains (NO in step S7), and therefore scanning is performed until x = x end (step S46). That is, when scanning is performed until X=x end (YES in step 816), it is determined that θ is 10 in step S23, and the image signal in the image memory 4 is not subjected to tilt correction (YES in step S26). , the document size is determined from points ■ and ■ and θ=0, and image formation is performed as is (step 32B, step S29).

In this way, as can be seen from Figures 3 and 4, the point ■
appears as the first change point in the total number, and point ■ appears as the last change point.

Note that in the figure, points ■ to ■ are output in the order of ■, ■, ■, but this differs depending on the document size, document placement status, and function gradients a and -a. Yes, it is not fixed.

Further, the gradients of the functions in bands I and (2) may be different.

Next, FIG. 6 is a diagram for explaining the second embodiment, and FIG. 6(a) is a diagram showing the relationship between the placed original and the intersection between the placed original and the function pattern, and FIG. 6(b) is a waveform diagram showing output pulses in each band.

That is, the second example is y=ax+bSy=-ax+b It consists of two function patterns (n=2).

Further, in this embodiment, the slope a is determined so that the angle formed by the two function patterns is narrower (more preferably approximately 30 degrees or less) than in the first embodiment.

As a result, regardless of the tilted state of the document (except when it is extremely tilted), it is possible to always select two specific points (points ■ and ■ in Figure 6).
), the inclination angle θ of the document can be found.

In this second embodiment, since the bands are I and ■, there is no particular need to identify the band in which the first change point ■ in the total number appears.

Therefore, sequential scanning is performed, and first, the first change point of the total number appears in band I (waveform 2 in FIG. 6(b)), and the coordinates of this point 2 are stored in the memory.

Next, point ■ (waveform ■ in Figure 6(b)) on band I, point ■ (waveform ■ in Figure 6(b)) on band ■, point ■ (waveform ■ in Figure 6(b)) appears, and its coordinates are sequentially stored in memory.

Then, from the obtained points ■ and ■, the inclination angle θ of the original is calculated as tanθ= (X2 −Xl )/(yz −y>
), and the document size is determined from this θ and the coordinates of points ■ and ■.

On the other hand, when the original is placed clockwise, the dot ■ appears on the band ■, and the dot ■ appears on the band. Also,
Although the timings at which dots ■ and dots ■ appear are different, the processing procedure for determining the tilt angle and document size is the same.

In the first embodiment, if the three function patterns are set so that the intersections ■, ■, ■ between the function pattern and the document appear on the l side of the document when the document is normally placed. , it is not particularly necessary to select coordinate data for determining the inclination angle of the document from different bands depending on the inclination direction of the document. However, in order to take into account the possibility that the document is placed at an extremely tilted angle, or to improve the calculation accuracy of the tilt angle, the slopes a and -a should be set larger, and the slopes should be adjusted according to the document placement condition. It is preferable to select coordinate data for calculating the tilt angle from different bands.

Furthermore, in the first and second embodiments, for convenience of explanation, the coordinates of the point at which the total number of output pulses in each band changes are stored in the memory, but in actual processing, the total number changes. It is arranged that the coordinates of the scan instant immediately preceding the scan that occurred are stored in the memory.

Further, in this embodiment, a straight line is shown as an example of the function pattern, but it may be a log function or other curved pattern, for example.

〔Effect of the invention〕

As described above, according to the present invention, a predetermined function pattern is displayed in advance on the document holder for each band, and the document size is determined from the change point of the total number of document area pulses and function pattern pulses in the image memory. Therefore, the document size can be detected by simply detecting a change in the number of pulses and calculating from the detection result, and the detection process and the circuit configuration therefor can be simplified.

In addition, the coordinate data for determining the tilt angle of the manuscript is selected from different bands depending on the tilt direction of the manuscript.
Two points on one side of the document can be selected reliably, and the document size can be determined without error.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a portion that detects the document size according to the present invention, FIG. 2 is an overall configuration diagram showing an embodiment of an image forming apparatus to which the present invention is applied, and FIG. 3 is a first embodiment. This figure (a) shows the relationship between originals placed counterclockwise and the intersections between the placed originals and the function pattern, and (b) shows the output pulses in each band. FIG. 4 is also a diagram for explaining the first embodiment, and FIG. 4 (a) shows the relationship between the manuscript placed counterclockwise and the intersection of the placed manuscript and the function pattern. , the same figure (b
) is a waveform diagram showing output pulses in each band, FIG. 5 is a flowchart explaining the operation of the first embodiment, FIG. 6 is a diagram explaining the second embodiment, and (a) A diagram showing the relationship between originals placed in a counterclockwise direction and the intersections between the originals placed and the function pattern, and (b) of the same figure is a waveform diagram showing output pulses in each band. 1...CC. D, 3... Judgment unit, 4... Image memory, 5... Image processing unit, 5l... Image processing means, 5
2... Original size ROM, 53... Function pattern R
OM, 6... Writing/reading means, 7... Laser emitting means, 8... Image forming device.

Claims (1)

[Claims] 1. In a document size detection device that temporarily stores an image signal read by surface scanning using an optical system in an image memory and detects the document size from this image signal, at least a portion that overlaps with the placed document is detected. and n function patterns are read by the optical system on the mirror sheet so that one pattern is included corresponding to each of the n bands provided in one scanning direction on the scanning surface by the optical system. a first storage means that stores the coordinate data of each of the function patterns; an image processing means that creates a document area pulse from the captured image signal; a second storage means for storing coordinate data at a point in time when the total number of the document area pulse number and the function pattern pulse number changes; A first calculation means for calculating the inclination of the document from the coordinate data; coordinate data at the first and last change points of the total number among the coordinate data in the second storage means; and a calculated value by the first calculation means. A document size detection device comprising: second calculation means for calculating a document size from 2. In the document size detection device according to claim 1, when n is 3 or more, the coordinate data used for calculation by the first calculation means is selected from different bands according to the band in which the total number changes first. A document size detection device characterized in that: