JPH078005B2 - Image processing device - Google Patents

Description

TECHNICAL FIELD The present invention relates to an image processing apparatus for processing image data for each pixel obtained by reading an original image according to a predetermined processing condition.

(Prior Art) Conventionally, as a method of binarizing an image signal, a threshold level (hereinafter referred to as a threshold level) is calculated by a statistical method from an image signal read from a document, and this threshold value is calculated.
There is a processing method that binarizes at the level.

In the case of such a processing method, the threshold for binarization
When the level is determined in one mode, there is a disadvantage that a necessary area is skipped depending on the original.

(Object) The present invention has been made in view of the above points, and an object of the present invention is to provide an image processing apparatus capable of performing suitable binarization processing according to a document.

Embodiments Embodiments of the present invention will be described in detail below with reference to the drawings.
FIG. 1 is a schematic diagram of a document reading apparatus to which the present invention can be applied. In order to read the image information of the original 102 placed on the original table 101 and pressed by the original cover 110, an image pickup device such as a CCD
103 is used, the illumination light from the light source 104 is reflected on the surface of the original 102, and is imaged on the image sensor 103 by the lens 108 via the mirrors 105, 106 and 107. The light source 104, the mirror 105, and the mirrors 106, 107 move at a relative speed of 2: 1. This optical unit is driven by the DC servomotor 109
Move from left to right at a constant speed while applying control. This moving speed varies from 90 mm / sec to 360 mm / se in the outward path depending on the magnification.
It is variable up to c and is always 630 mm / sec on the return trip. While scanning the main scanning line orthogonal to the moving sub-scanning direction A of this optical unit with the image sensor at a resolution of 16 pel / mm, the optical unit is moved forward from the left end to the right end and then back to the left end again. Finish scanning.

FIG. 2 shows a schematic block diagram of a circuit for processing an image signal from the image sensor 103. The image signal V _D read by the image sensor 103 is converted into a 6-bit digital signal by the A / D converter 201, and is sent to the latch 203, comparator 204, and latch 205 in synchronization with the sampling clock SCL via the latch 202. To be The comparator 204 compares the 6-bit image signal sent from the latch 202 with the 6-bit image signal one clock before sent from the latch 203, and if the latch 202
If the new image signal sent from is smaller, it outputs a comparator output to the AND gate 206. And gate 2
06 sends the comparator output from the comparator 204 to the latch 205 in synchronization with the sampling clock SCL. Upon receiving the comparator output, the latch 205 sends the image signal sent from the latch 202 to the cpu 208. Also, AND gate 20
In addition to the comparator output and the sampling clock SCL, the enable signal EN indicating the effective section of the image signal from the image sensor 103 is input to 6 and the comparison result of the image signal in the predetermined section of each main scanning line and the latch 205 are entered. To send to cpu208. The cpu 208 can detect the lowest density level of each main scanning line, that is, the background level by taking in the image signal from the latch 205 in synchronization with the main scanning line synchronization signal MS.

When the surface of the document cover 110 on the platen side is black or mirror-finished, the image signal of the part other than the document becomes black and is not detected as the background level, so there is no problem with the enable signal EN as the maximum main scanning width. . Further, when there is a possibility that a non-original part is detected as white due to a conveyance belt, such as when equipped with DF and ADF, the enable signal EN is limited to the minimum paper size.

Based on the detected background level, the cpu 208 determines the threshold level for each main scanning line by an algorithm described later, and sends the threshold level to the comparator 207 in synchronization with the main scanning line synchronization signal MS. In the comparator 207, the image signal from the latch 203 and the threshold value from the cpu 208
The levels are compared to generate a binarized signal V _S. As mentioned above
Since the level of each pixel after A / D conversion is 6 bits, it is given as 3F (HEX) for the blackest and 0 for the whitest, and the threshold level and the background level for each main scanning line are 0 to 3F.
Given in (HEX).

Then, the background level of the main scanning line of interest is predicted from the background level of each of the N lines immediately preceding the main scanning line of interest, and the threshold level is determined. The algorithm is explained.
In the following, W _i-1 is the background level detected for the (i-1) th main scanning line, Algorithm to be described later from the i at background level prediction value of main scanning line, N pieces of data from the background level W _iN of the iN main scanning line to background level W _i-1 of the (i-1) th main scanning line Predicted in. Si is the threshold level for the i-th main scan line, Is determined by the algorithm described below. The above algorithm is programmed and stored in the ROM 209. Further, N background levels are stored in the RAM 210.

The basic idea will be described with reference to FIG. After the scanning of the i-1th line is completed, before the scanning of the i-th line is started, the N background levels W _ik , R = N detected from the N-th line from the i-Nth line to the i-1th line are detected. 1, ..., N
Based on the i-th line And then Then, the threshold level Si for the i-th line is calculated based on When scanning the i-th line, the threshold level Si is binarized, and at the same time, the background level of the i-th line is obtained.

Next, FIG. 5 shows a schematic flow of the sequence. It will be described with reference to FIG. 3 according to the following flowchart.

In steps 501 to 504, the optical system 200 starts to move forward from the home position A point and initializes the steps 502 and 503 until the image front end point B is reached. In step 502, N background levels immediately before the target line are always stored.
Bytes RAM 210 area BUF ₁ to BUF _N are cleared to 0. After the optical system 200 reaches the point B in step 504, each time the main scanning line synchronizing signal comes in step 505, step 506 is performed.
Incorporate the background level of the front line with, and BUF with step 507
The oldest background level data stored in ₁ is discarded, BUF _j is shifted to BUF _j-1 , and the latest data taken in is discarded.
Store in BUF _N. As a result, the latest N pieces of background level data are always stored in BUF _{1 to} BUF _N.
Next, in step 508, the maximum value and the minimum value are omitted in order to remove the peculiar data from the N background level data, and the remaining N are left.
-Average of two data And I.e. Is calculated. Further in step 509 Is smaller than the predetermined value P, the routine proceeds to step 510, where the threshold level is determined in the first mode. That is, it is predicted Make sure that the threshold level Si is Is calculated. or, Is greater than the predetermined value P, the process proceeds to step 511 to determine the threshold level in the second mode. That is, the threshold level Si is set to a threshold level Si = 3F _H * α when the background level is 0, and the binary is converted into black without skipping as the background. Where α is black level 3F _H 0 <α <1 is satisfied by a coefficient for setting an internal division point of a predetermined ratio between the two as a threshold level, for example, Is determined empirically. In addition, the coefficient α is, for example, according to the set density set by the operator by a density lever (not shown) of the operation unit 211 (FIG. 2), for example, It is also possible to take any of the values of.

The threshold level determined as described above is set to step 51.
It is output at 2, and at step 513, 1 is added to i. And
Steps 505 to 513 are repeated until the optical system reaches the reversal position C at step 514.

When the optical system inversion position C is reached, steps 515 to 5
The optical system 200 is inverted at 17 and when the home position A is reached, the optical system 200 is stopped.

In the above embodiment, the actual detection data is always stored in BUFi after the (N + 1) th line. However, for the first to Nth lines, the background level detection is performed before the first line as shown in steps 502 and 503. The threshold level is calculated and binarized on the assumption that the provisional N line having a value of 0 is assumed. For example, if N = 16, the first 16 lines will be binarized using temporary blank data, but at 16pel resolution, 1
It is equivalent to mm, and it is possible to assume that the leading edge of 1 mm in an actual document is the background level 0 of the background part without normal information.

By controlling in this way, it becomes possible to separate the background part and the information part even for a document whose background level changes greatly. For example, as shown in FIG. 6 (a), for a document whose background level W greatly changes, as shown in FIG. When the image signal V _D is binarized by, the background part and the information part are not sufficiently separated. On the other hand, the threshold level The image signal V _{D is set} to 2 by following the change in the background level W as
By digitizing, the background part can be surely separated from the information part in white as shown in FIG. 6 (c).

Also, in the case of an original document with white characters as shown in FIG. 7 (a),
As described above, when the threshold level is made to follow the change in the background level and the image signal is binarized, the background level W and the threshold level S are obtained as shown in FIG. 7 (b). In the area having only the black level in the main scanning direction as shown in (4), although it is desirable to binarize with black, it is binarized as white.

In order to eliminate such inconvenience, in the present embodiment, when the predicted background level is equal to or higher than the constant value P, it is determined that the white level does not exist and only the black level exists over the entire one main scanning line. , The threshold level at the background level 0 is output so as to binarize the black level as black as information without skipping it as the background.

As described above, according to the present embodiment, a simple circuit for detecting the background density level of each main scanning line is provided, and the background level of the attention line is predicted from the background level of the N line immediately before the attention main scanning line, If the threshold level is determined to make the predicted background level white, and the background level is blacker than a certain value, the threshold level is determined so that the predicted background level is black. In addition, a suitable binarization process can be performed, and neither preliminary scanning nor a memory for storing image signals is necessary, and proper ground erasing can be performed with simple birdware and software.

It should be noted that in the above-mentioned embodiment, the threshold is selected from N background levels.
Although the level is determined, it is of course possible to determine the threshold level from N = 1, that is, the main scanning line immediately before the target line.

Further, in the present embodiment, the background level of the attention line is predicted from the background level before the attention line, but the present invention is not limited to this. For example, the threshold level may be calculated from the read image signal and stored in the memory, or the threshold level of each line may be obtained in advance by performing preliminary scanning of the document. Good.

As described above, according to the embodiment of the present invention, the threshold level is determined in different modes according to the background density level, and the image signal 2
Since the binarization is performed, the binarization of the image signal can be performed at the optimum threshold level according to the original image.

Furthermore, by configuring the background density level of the line to be binarized according to the background density level before this line, a large amount of memory and preliminary scanning become unnecessary, and the time for scanning the original can be shortened. And it can be provided at low cost.

(Effect) As described above, according to the present invention, an image reading unit that reads a document image line by line and outputs digital image data for each pixel (in the embodiment, corresponds to the document reading apparatus in FIG. 1), White peak data of the density of the original image (also in FIG.
Wi) detection means (also latch 205 in FIG. 2)
And the background density level of the original image based on the white peak data detected by the detecting means. (Also step 508 in FIG. 5) and compares the predicted background density level of the original image with a predetermined fixed density level (also P in FIG. 5) (also step 509 in FIG. 5). ), If the background density level is lower than the fixed density level, a threshold value dependent on the background level based on the background density level (also step 510 in FIG. 5)
On the other hand, when the background density level is higher than the fixed density level by the first threshold value determining mode,
Threshold value determining means for deciding a threshold value for binarizing the digital image data in a second threshold value deciding mode for deciding a predetermined threshold value (also step 11 in FIG. 5) so as to binarize the digital image data with black. (Similarly, CPU 208 in FIG. 2) and binarizing means (also comparator 207 in FIG. 2) for binarizing the digital image data by the threshold value determined by the threshold value determining means The original image can be binarized with an appropriate threshold value according to the background level of the original image predicted based on the peak data of the density, and the background density level of the predicted original image is lower than the predetermined density level. When it is high, the threshold value for binarization can be determined without being affected by the predicted background density level. It is possible to faithfully binarize an original image as shown in FIG. 7A, which has a low density level.

[Brief description of drawings]

1 is a schematic diagram of a document reading apparatus to which the present invention can be applied, FIG. 2 is a block diagram of an image signal processing circuit, FIG. 3 is a diagram showing a positional relationship between a document table and an optical system, and FIG. FIG. 5 is a diagram showing a procedure for determining a threshold level for a line, FIG. 5 is a flowchart showing a control procedure according to the present invention, and FIG. 6 is a diagram showing a relationship between the threshold level and a binary signal.
The figure is a diagram showing an example of binarization of a document having a solid black area. In the figure, 101 is a platen, 102 is a document, 103 is an image sensor, 201 is an A / D converter, 202, 203 and 205 are latch circuits, 2
04 and 207 are comparators, 208 is cpu, 209 is ROM, 210 is RAM
Is.

Claims (1)

[Claims] 1. An image reading means for reading a document image line by line and outputting digital image data for each pixel, a detecting means for detecting white peak data of density of the document image, and a detecting means for detecting the white peak data. Based on white peak data, the background density level of the original image is predicted, the predicted background density level of the original image is compared with a predetermined fixed density level, and the background density level is the fixed density. If the background density level is lower than the level, the first threshold value determination mode for determining a threshold value depending on the background level is used based on the background density level. On the other hand, if the background density level is higher than the fixed density level, The second for determining a predetermined threshold value for binarizing the digital image data to black
Of the digital image data according to the threshold value determination mode of
An image processing apparatus comprising: a threshold value determining unit that determines a threshold value for binarization; and a binarizing unit that binarizes the digital image data according to the threshold value determined by the threshold value determining unit.