JPS59189784A - Picture signal processing device - Google Patents

Abstract

PURPOSE:To perform a suitable binarizing processing of a picture signal in accordance with an original by detecting a texture density level of a line to be binarized and determining a threshold level in a variable mode in accordance with this level and binarizing the picture signal with this threshold level. CONSTITUTION:After scanning of the (i-1)th line, a forecast value W'i of the i- the line is obtained as an average value of N-number of texture levels Wi-k (k=1-N) of the (i-N)th and following lines before scanning of the (i-N)th line, and a threshold level Si for the i-th line is obtained on a basis of this value W'i. When scanning of the i-th line is started, it is binarized by the level Si, and the texture level of the i-th line is obtained. In this case, the threshold level Si is so decided that W'i is white when W'i is smaller than a prescribed value P. If W'i is larger than the prescribed value P, the level Si is defined as a level for texture level ''0''. With the threshold level decided in this manner, one is added to (i). This operation is repeated until an optical system reaches an inversion position C.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field The present invention relates to an apparatus for processing image signals, particularly for performing binarization processing.

Conventionally, as a method of binarizing an image signal, there is a processing method of binarizing an image signal read from a document at a threshold level (hereinafter referred to as thread).

In the case of such a processing method, the threshold e for binarization is
When the level is determined using one mode, there is an inconvenience that, depending on the manuscript, necessary areas may be skipped.

- Purpose The present invention has been made in view of the above points, and it is an object of the present invention to provide an image signal processing device that can perform suitable binarization processing of image signals depending on the document.

A further object of the present invention is to provide an image signal processing device that does not require a large amount of memory or preliminary scanning.

Example 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 device to which the present invention can be applied. In order to read the image information of the original 102 held by the original cover 110 and placed on the original table 101, a CCD is used.
The illumination light from the light source 104 is reflected on the surface of the original 102, and the mirrors 105, 1
06 and 107 to the lens 108 and the image sensor 1
06. The light source 104, the mirror 105, and the mirrors 106 and 107 move at a relative speed of 2:1. This optical unit moves from left to right at a constant speed while being subjected to PLL control by a DC servo motor 109. This movement speed varies from 90 rMv/sec to 660rtHrry's on the outward journey depending on the magnification.
Variable up to e c, always 630+m@/s on the return trip
It is ec. The main scanning line perpendicular to the person in the sub-scanning direction in which this optical unit moves is captured by the image sensor16pe17
- The optical unit is moved forward from the left end to the right end while reading at a resolution of -, and then moved back to the left end to complete one scan.

FIG. 2 shows a schematic block diagram of a circuit that processes image signals from the image sensor 106. The image signal read by the image sensor 106 is converted into a 6-bit digital signal by the converter 201, and sent via the launch 202 to the launch 206, comparator 204 . The signal is sent to latch 205.

The comparator 204 compares the 6-bit image signal sent from the launch 202 with the 6-bit image signal sent from the latch 206 one clock ago, and if the launch 2
If the new image signal sent from 02 is smaller,
A comparison output is output to the AND gate 206. AND gate 206 sends the comparison output from comparator 204 to latch 205 in synchronization with sampling clock SCL. When the latch 205 receives the comparator output, the CPU outputs the image signal sent from the latch 202.
Send to 208. In addition to the comparator output and the sampling clock SCL, the AND gate 206 also receives an enable signal B indicating the effective period of the image signal of the image sensor 106.
N is entered, and the comparison result of the image signal in a predetermined section for each main scanning line is sent from the latch 205 to the CPU 208. 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 launcher 205 in synchronization with the main scanning line synchronization signal MS.

If the surface of the document cover 110 on the document table side is black or mirror-finished, the image signal of the non-document portion becomes black and is not detected as the background level, so there is no problem when the enable signal EN is set to the maximum main scanning width. Also D F
, When there is a possibility that a non-document portion may be detected as blank by a conveyor belt or the like, such as when an ADF is installed, the enable signal KN is limited to the minimum paper size.

Now, Cp[1208 is a threshold value for each main scanning line based on the detected background level using an algorithm described later.
Determine the level of the main scanning line synchronization signal MSK synchronization L-
Sends the C threshold level to comparator 207. In the comparator 207, the image signal from the latch 206 and cp
The thread level from u 20B is also given as 0 when it is white, and the threshold level and background level for each main scanning line are also given as 0 to 3F (HEX).

Next, the background level of the main scanning line of interest is predicted from the background level of each of the N lines immediately before the main scanning line of interest, and the threshold level is determined. Explain the algorithm.

In the following, wi-1 is the background level detected for the i-1st main scanning line, Wi is the predicted background level value of the i-th main scanning line, and from the background level Wi of the i-Nth main scanning line to the i-th N up to background level Wi-1 of 4 main scanning lines
It is predicted from the following data using the algorithm described below. 8
4 is the threshold level for the i-th main scanning line;
It is determined by an algorithm described later from the predicted background level Wi of the i-th main scanning line. The aforementioned algorithm is programmed and stored in tOM2D9.

Further, N background levels are stored in the RAM 210.

The basic idea will be explained based on FIG. After the scanning of the i-1th line is completed, and before starting the scanning of the i-th line, the N background levels Wi-k detected from the N lines from the i-Nth line to the i-1th line are =1.・
..., calculate the predicted value Wi of the background level of the i-th line based on N, and then calculate the predicted value Wi of the background level Wi of the i-th line based on the predicted background level Wi.
Find the threshold level/I/Si for the line.

When the i-th line is scanned, the threshold level Si
At the same time as the digitalization, the background level of the i-th line is determined.

Next, FIG. 5 shows a schematic flow of the sequence.

The process will be explained below according to this flowchart with reference to FIG.

In steps 501 to 504, the optical system 200 starts moving forward from its home position point A, and performs initialization in steps 502 and 503 until it reaches the eight leading edges of the image.

In step 502, areas BUF+ to BUFN on the N-byte RAM 210, where N background levels immediately before the line of interest are always stored, are cleared to zero. Step 5
After the optical system 200 reaches point B in step 505
Every time the main scanning line synchronization signal arrives, the background level of the previous line is taken in at step 506, and the background level of the previous line is taken in at step 507.
The oldest background level data stored in UFl is discarded, 13UFj is shifted to BUFj-1, and the latest data taken in is stored in UFN. As a result, the latest N skin level data are always BUF1 to BtJFN.
It will be stored in . Next, in step 508, N
In order to remove peculiar data from individual skin level data, K,
The maximum value and the minimum value are omitted, and the average of the remaining N-2 pieces of data is set as the predicted value W1 of the background level of the line of interest. That is, calculate. Further, if Wi is smaller than the predetermined value P in step 509, the process proceeds to step 510, and the threshold level is determined in the first mode. Immediately, attack + to 5i = (3F + - Wi): 5i = (3F + - Wi) so that the predicted skin level Wi becomes white.
It is determined by calculating Wi. If the skin level prediction Δ measured value Wi is equal to or greater than the predetermined value 2, the process proceeds to step 511, where the threshold level is determined in the second mode. So, when the threshold level is S Mira and the background level is 0, the threshold level 8i=3F+: is set to α. Here, α is the internal division point of a predetermined ratio between the black level 3FH and the background level Wi.
The concentration lever (not shown in Fig. 2) is adjusted depending on the set concentration set by the operator, for example α = upper - 2.
It is also possible to take any value such as -...9io'1o'' io-.

Step 5
12, and 1 is added to i in step 513. Then, steps 505 to 513 are repeated until the optical system reaches the inversion position C point in step 514.

When the optical system reaches the inversion position C, step 515
The optical system 200 is reversed in steps 517 to 517, and when it reaches the home position A, the optical system 200 is stopped.

In the above embodiment, the actual detection data is always stored in BUFi after the N+1 line, but from the 1st line to the Nth line, the background data is stored before the 1st line as shown in steps 502 and 506. Assuming a virtual empty N line with a level detection value of 0, calculate the threshold level 2
Perform value conversion. For example, if N=16, the first 116
The line will be binarized with temporary empty data, but 1
With a resolution of 6 pels, this corresponds to 1 pixel, and it is safe to assume that the leading edge of the actual manuscript is 0 in the number of backgrounds with no information.

By controlling in this way, it is 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)K, conventionally, for a document in which the background level W changes greatly, as shown in FIG.
As shown in b), when the image signal VD is binarized using Threshold Rubel SA = (3FH-0)X-, the background part and the information part are not sufficiently separated.

Koreni vs. Shi, Thresh Level 5B = (3FH-W
) By binarizing the image signal Vo by following the change in the background level W as XY+W, the background part can always be separated from the information part in white as shown in FIG. 6(C).

In addition, in the case of a manuscript with white characters as shown in Figure 7(a),
When the image signal is binarized by making the threshold reference level follow the change in the background level as described above, the background level W and the threshold level S are obtained as shown in FIG. 7(b). As shown in C), an area having only a black level in the main scanning direction is binarized as white even though it is desirable to binarize it as black.

In order to eliminate such inconvenience, in this embodiment, when the predicted background level is equal to or higher than a certain value P, it is determined that there is no white level and only black level exists over all main scanning lines. For example, the threshold level when the background level is O is output so that the black level is not skipped as the background but is treated as black and binary information.

In this way, according to this embodiment, a simple circuit is provided to detect the background density level of each main scanning line, and the background level of the line of interest is predicted from the background level of the N lines immediately before the main scanning line of interest. Preliminary III to the background and Belwo Shiranicho <
When the threshold reference level is determined and the background level is blacker than a certain value, the threshold level is determined so that the predicted background level is black, so that the binary value is suitable even for originals such as peta black. Furthermore, there is no need for preliminary scanning or image signal storage memory, and appropriate ground skipping can be performed using simple hardware and software.

In the above embodiment, N background level force/rathreshold levels are determined, but it is of course possible to determine the threshold level from N=1, that is, from the main scanning line immediately before the line of interest.

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

Effects As described above, according to the present invention, the threshold level is determined in different modes depending on the background density level and the image signal is binarized. 2
It is possible to perform value conversion.

Furthermore, by configuring the system so that the background density level of a line to be binarized is predicted according to the background density level of the line before this line, a large amount of memory, preliminary scanning, etc. are not required, and the time required for document scanning can be shortened. It can be provided at low cost.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram of a document reading device to which the present invention can be applied, FIG. 2 is a block diagram of an image signal processing circuit, FIG. 6 is a diagram showing the positional relationship between the document table and the optical system, and FIG. FIG. 5 is a flowchart showing the control procedure according to the present invention; FIG. 6 is a diagram showing the relationship between the threshold level and the binarized signal; FIG.
The figure shows an example of binarization of a document having a peta black area. In the figure, 101 is a platen, 102 is a document, and 103 is a platen.
is an image sensor, 201 is a ~Φ converter, 202, 203
205 is a latch circuit, 204 and 207 are comparators, 208 is CPU, 209 is ROM, 210 is RA
It is M.

Claims (1)

[Claims] (1) In an image signal processing device that reads an original image line by line and binarizes the read image signal according to a threshold level, detects the background density level of the line to be binarized. An image signal processing apparatus characterized in that a threshold level is determined in different modes according to a detected background density level, and an image signal is binarized according to this threshold level. (2) The detected background density level is compared with a predetermined density level, and if the background density level is whiter than the predetermined density level, the threshold level is determined according to the background density level. An image signal processing device according to claim 1. (6) The detected background density level is compared with a predetermined density level, and if the background density level is blacker than the predetermined level, the threshold level is set to a predetermined level. An image signal processing device according to items 1 to 2. (4) Set the background density level of the line to be binarized to 2.
2. The image signal processing device according to claim 1, wherein the image signal processing device detects the image signal by predicting it according to the background density level before the line to be converted into a value.