JPH0556257A - Image reader - Google Patents

Abstract

PURPOSE:To obtain an image reader which reads an image in a good condition regardless of the character of the original of a reading object. CONSTITUTION:The reader is equipped with a CCD 2-4 to detect the transmission quantity of the light of an original, a CCD 2-5 to detect the concentration of an original image and a neuro IC l-3 to receive the data from two these CCDs and prepare the threshold level loaded to the processing of an image processing part 1-4.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading device, and more particularly to an image reading device which constitutes a portion for reading a document image in a copying machine or the like.

[0002]

2. Description of the Related Art Normally, in a copying machine, an original image to be read is processed on the assumption that it has a predetermined material or thickness.

The original is placed with the reading surface down on the original table (platen glass), fixed with an original pressing cover, and then copied.

[0004]

By the way, when the original document to be read is made of a material having high transparency unlike ordinary recording paper, and images are recorded on both sides of the original, the image on the back side is also included. There is a problem that the image is read and copied in a state where the front image and the back image are combined.

Further, even if there is a read image on only one side of the original, if the pressing surface of the original pressing cover is dirty, it is also a factor that deteriorates the quality of the output image by reading the dirt. It was

The present invention has been made in view of the above-mentioned prior art, and regardless of the nature of the document to be read,
An object is to provide an image reading device that reads an image in a good state.

[0007]

In order to solve this problem, the image reading apparatus of the present invention has the following structure.
That is, the first to detect the property of the original for the transmission of light.
Of the original image, the second detecting unit for reading the original image from the reading surface and detecting the density thereof, and the correction for correcting the original image reading density based on the detection results of the first and second detecting units. And means.

[0008]

In the structure of the present invention, the reading density of the original image is corrected by the correcting means on the basis of the state quantities detected by the first and second detecting means.

[0009]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram of the first embodiment. In the figure, 1-1 is a paper quality detection sensor, 1-2 is a character density detection sensor, 1-3 is a neuro IC, and 1-4 is an image processing unit. The character density detection sensor is also used when reading a document image. Also, the image processing unit 1-
An image recording unit 4 records an image on recording paper.

FIG. 2 shows an example of the paper quality detection sensor 1-1 and the character density detection sensor 1-2. In the figure, 2-1 is a document illumination lamp, 2-2 is a document, 2
Reference numeral -3 is a pressure plate (original pressing cover), 2-4 is a CCD used as a paper quality detection sensor, and 2-5 is a CCD used as a character (original) density sensor.

The connection and operation of each circuit shown in FIG. 1 will be described below.

The signal output terminal of the paper quality detection sensor 1-1 is connected to the input terminal of the neuro IC 1-3 through the signal line 1-5, and the signal output terminal of the character density detection sensor 1-2 is connected to the signal line 1-6. Through the signal input terminal of the neuro IC 1-3 and the density level input terminal of the image processing section 1-4. Further, the signal output terminal of the neuro IC 1-3 is connected to the threshold level input terminal of the image processing unit 1-4 through the signal line 1-7, and the image processing unit 1-4 receives the density given by the signal line 1-7. It operates so as to perform the image recording process only on the image having the value or more.

In the neuro IC used in this embodiment, data representing the amount of light transmitted through the paper detected by the paper quality detection sensor 1-1 and data representing the character density of the original document detected by the character density detection sensor 1-2 are neural. The weighting is determined by a learning method such as a back-propagation method so that data representing a density level that does not show through can be obtained from the output layer by giving the weight to the input layer of the net and changing the weight information through the intermediate layer. It is a thing.

The image processing unit 1 receives data representing the density level obtained by the neuro IC 1-3 and data representing the character density of the original detected by the character density detection sensor 1-2.
-4 and input the character density data of the document to Neuro IC1
By masking with the data representing the density level obtained by -3, the density below a certain level is cut.

That is, since the show-through character is transmitted through the original paper, the reading surface (front surface)
It is noted that the original image is read with the show-through character removed, paying attention to the fact that the density is lower than the density of the character.

FIG. 3 shows a neural network in the embodiment.

The operation of the circuit shown in FIG. 2 will be described below.

At the time of prescan, the document illumination lamp 2-1 illuminates the document 2-2 and the CCD 2-5 measures the density of the document. At the same time, the CCD 2-4 built in the pressure plate 2-3 measures the amount of transmitted light of the document. Then, data representing the document density and the amount of transmitted light is sent to the neuro IC.

Specifically, the original image is pre-scanned to check the density distribution of each read pixel viewed from the surface of the original image. At this time, assume that two peaks A and B are detected as shown in FIG. In this case, when the light transmission amount of the document is large, it is highly possible that the show-through is reflected in the image of the mountain B. Similarly, in the case of an original having a small light transmission amount and no show-through, the image of the mountain B is likely to be an image drawn on the front surface. Therefore,
In the former case, the threshold level for ignoring the read density is high (the position of the valley of peaks A and B), and in the latter case, it is kept low.

Thus, the original density and the amount of transmitted light (and the degree of light transmission) are obtained, and the original density at the time of original reading is corrected based on these two parameters. Since the density level to be cut is detected in the density correction processing of the embodiment as described above, any image below the density level is ignored.

<Description of Second Embodiment> FIG. 4 is a view showing a second embodiment. In FIG. 4, 4-1 is a document density detecting sensor, 4-2 is a CPU, and 4-3 is a neuro IC. Four
-4 is an image processing unit.

The connection and operation of the second embodiment will be described in detail.

The data output terminal of the document density detecting sensor 4-1 is connected to the CPU 4-2 and the image processing section 4 through the signal line 4-5.
-4 connected to the concentration signal input terminal,
The data output terminal of No. 2 is connected to the neuro I through the signal line 4-6.
It is connected to C4-3. Also, this neuro IC4
-3 data output terminal is connected to the threshold level input terminal of the image processing unit 4-4 through the signal line 4-6.

The CPU 4-2 determines the maximum value and the minimum value of the data representing the document density detected by the document density detection sensor 4-1 and inputs the maximum value and the minimum value to the neuro IC 4-3. This neural network is learned by the back propagation method so that the maximum value and the minimum value of the document density are input to the input layer and the data representing the optimum density is output to the output layer.

Since the operation of the other parts has been described in the first embodiment, it will be omitted.

By using the present invention, not only the show-through can be removed, but also the optimum density according to the paper quality can be determined.

As described above, according to this embodiment, the proper original document density is determined by the neural network having three or more hierarchical layers that receive the amount of transmitted light of the paper and the original document density. This makes it possible not only to prevent the show-through of the document but also to set the document density according to the paper quality.

[0029]

As described above, according to the present invention, it is an object of the present invention to provide an image reading apparatus for reading an image in a good condition regardless of the nature of the original to be read.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of an image processing apparatus according to a first embodiment.

FIG. 2 is a diagram illustrating details of a sensor unit according to the embodiment.

FIG. 3 is a diagram illustrating a neural network according to the first embodiment.

FIG. 4 is a block configuration diagram of an image processing apparatus according to a second embodiment.

FIG. 5 is a diagram showing an example of a density distribution of a read image.

[Explanation of symbols]

 1-1 Paper quality detection sensor 1-2 Character density detection sensor 1-3, 4-3 ... Neuro IC 2-1 Original illumination lamp 2-2 Original 2-3 Crimping plate 2-4 CCD 2-5 Character density detection sensor 4 -1 Document Density Detection Sensor 1-4, 4-4 Image Processing Unit

Claims (1)

[Claims] 1. A first detecting means for detecting a light transmission property of an original, a second detecting means for reading an original image from a reading surface and detecting its density, the first and second detecting means. An image reading apparatus comprising: a correction unit that corrects a reading density of a document image based on a detection result of the unit.