JPS58115975A - Picture discriminator - Google Patents

Abstract

PURPOSE:To discriminate an original picture whether it is a line drawing using the picture of an original, that is, lines as major picture constituents, a gradation picture having continuous gradation of contrast, or a mixed picture with the line drawing and the gradation picture, by using an original picture density signal obtained through the scanning of the original. CONSTITUTION:In Figure, 31 is an original, 32 is a transparent original stand reciprocating to the arrow and 33 is a light source. The original is irradiated with a halogen lamp uniformly. 34 is a mirror, 35 is a photoelectric conversion section, a light amount signal So being a reflected light corresponding to the density of picture from the original is bent for the optical path with the mirror 34, enters the section 35 and is outputted as an electric picture signal Se. The picture signal Se outputted from the section 35 is discriminated for the picture at a signal processing section succeedingly. Since adjacent small minimum values being noise are erased for a picture having much density change, the discrimination of picture is done surely.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention uses an original image density signal obtained by scanning an original to obtain an original image, that is, a line drawing with lines as the main image component, or a continuous gradation of shading. The present invention relates to an image discriminating device that determines whether the image is a gradation plane having a gradation plane or a hybrid image in which a line drawing and a gradation plane are mixed.

Copying technology has improved in recent years, and originals consisting only of line drawings or gradation planes can now be copied according to their image quality. Specific Ki:! Line drawings and gradation planes are visually determined, and independent operation buttons are provided to specify the recording conditions for line drawings or gradation drawings, and depending on the visual results, the operator selects the appropriate operation button and copies accordingly. is being carried out. Therefore, unnecessary copying may occur due to a judgment error or an operation error.

Furthermore, in today's world where copying is becoming faster and more efficient, and copying machines equipped with automatic document feeders (ADF devices) are rapidly becoming popular, the above-mentioned visual image discrimination processing operation is becoming faster and more efficient. , obstructing the purpose of efficiency, and special K A
It is practically impossible to apply this method to a copying machine equipped with a DF device.

On the other hand, when transmitting or storing compressed image data, such as in facsimiles and image databases, information compression efficiency is poor unless a coding method is applied that is suitable for both line drawings and gradation planes. Image recognition devices are now required.

One of the conventional image discrimination methods is to scan a document,
Forms a histogram according to the density distribution of each point on the manuscript,
A method for determining an image that has a pattern of the histogram (
Although there is a method (Japanese Patent Application No. 120623/1983), it is mainly effective for line drawing information (5) and is difficult to apply to gradation drawings.

Furthermore, the method using the histogram described above requires a central processing unit such as a microprocessor that performs calculations and judgments such as data statistics and conversion, as well as ROM and R that are essential for the operation of the device.
It required memory such as AM, was expensive, and tended to be complicated.

In view of the above-mentioned demands for higher speed, efficiency, and higher image quality of recording devices such as copying machines, the present invention enables clear discrimination of images such as line drawings, gradation drawings, and mixed drawings through signal processing of image signals. The purpose is to provide an inexpensive image discrimination device η using a simple circuit.

An object of the present invention as described above is to provide an image discrimination device that forms an image signal corresponding to a light intensity signal obtained by scanning a document, processes the image signal, and discriminates the image, based on the light intensity signal. A differentiation circuit that differentiates the time series output of the obtained image signal, a binarization circuit that compares the output signal from the differentiation circuit with a specific level and outputs a binarized pulse signal, and a binarization circuit that outputs a binarized pulse signal. This can be achieved by an image discrimination device that includes a counter that counts output pulses and discriminates an image based on the signal output from the counter.

Next, the principle and process of image discrimination according to the present invention will be explained.

FIG. 1 is a flowchart showing the process of image discrimination, and FIGS. 2(a), (b), and (o) are graphs for explaining the principle of image discrimination.

The document is scanned according to process 1 in the flowchart of FIG.
An electrical image signal se is outputted in a time-series manner from the photoelectric conversion elements such as the following. Further, if necessary, the image signal SE is amplified by an amplifier circuit (processing 2).

The shape and height of the SE curve for each pixel of a CCD or for a discrete photosensor vary greatly depending on the type of original image and the responsiveness of the photoelectric conversion element.In general, line drawings are performed using low-density SI.
It has a signal part consisting of many steep protrusions in the C value region,
On the other hand, in a gradation drawing, the Bv curve has gentle protrusions in the high-density SK value region and a few scattered steep protrusions than in a line drawing. Therefore, by differentiating the 88 curve, S
Image density factors existing within the IC value can be deleted to narrow down the image factors to the protrusion signals (processing 3).

Using a differential circuit, the SK concave curve in FIG. 2(a) was differentiated, and the results are shown as the SKK segment curve in FIG. 2(b).

Regarding the number of protrusions on the SE differential curve, for line drawings and gradation drawings, they are distributed around the average numbers isolated from each other, so images are discriminated based on the number of protrusions on the sm differential curve for line drawings and gradation drawings. A certain threshold value can be statistically determined as a standard (discrimination threshold value).

5- Process the count value as n5), compare the magnitude relationship between this n and the threshold value I mentioned above (judgment 1), and use the result as the second
It is shown in figure (c).

Hereinafter, specific embodiments of the image discrimination device of the present invention will be described.

FIG. 3 shows one embodiment of the present invention.

In the figure, 31 is a document, and 32 is a transparent document table, which moves back and forth in the direction of the arrow in the figure. 33 is a light source. The original g4 may be irradiated with laser light or convergent light from a lens,
Alternatively, light can be guided from a uniformly illuminated document to the photoelectric conversion section using a slit or a glass fiber. In the illustrated embodiment, the document is uniformly illuminated by a halogen lamp. The light source can be a laser light source, tungsten lamp, xenon lamp, sodium lamp, fluorescent lamp, LED,
EL or iodine lamps can be used.

6-34 is a mirror, and 35 is a photoelectric converter, in which the optical path of the optical amount signal So, which is reflected light corresponding to the image density from the original, is bent by the mirror 34, enters the photoelectric converter 35, and is electrically converted. It is output as an image signal Ss.

Between the photoelectric conversion parts of the illustrated embodiment, there is a condenser lens 352 that converges the light of the light intensity signal So, a filter 353 that is disposed as necessary to adjust the wavelength spectrum of the converged light, and a light intensity signal So. S.

A photoelectric conversion element 351 is provided for photoelectrically converting the image signal So into an image signal So. The installation position of the photoelectric conversion element 351 is necessary and may even be advantageous. The filter 353 may be installed either before or after the condensing lens 352, and it is not limited to a colored filter, and a prism or grating that separates wavelengths by refraction may be used. Furthermore, an infrared cut filter may be used.

The image signal Se output from the photoelectric conversion section 35 is subsequently processed by the signal processing section as explained in the flowchart above, and the image is discriminated. FIG. 4 does not show an example of a block diagram of the signal processing section.

In the figure, reference numeral 41 denotes a photoelectric conversion unit which receives the light amount signal So and outputs it as an image signal Ss. An amplifier circuit 42 amplifies the signal S6 and outputs an amplified image signal SE. 4
3 is a differential circuit. A differential value is determined from the 81 curve obtained by the amplifier circuit 42. 44 is a binarization circuit which outputs a binarization pulse based on a binarization threshold C. 45 is a pulse number counter. A comparator 46 compares the discrimination threshold l and the number of pulses n obtained by the counter 45, and outputs the image discrimination result as a control factor for display or recording conditions.

An example in which image discrimination is performed using the image discrimination device of the present invention described above will be described.

In FIG. 5(a), if the unit reading size is 2 mφ, which is a character string with many strokes, the output for determining the image can be changed.

By selecting the unit reading size, the SK
Even in an image with a convex curve and a very significant change in density, it is possible to eliminate small local maximum values that become noise in discrimination, so image discrimination can be made more reliably.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the image discrimination process of the present invention, and FIG. 2 is a diagram for explaining the principle of image discrimination. FIG. 3 is a schematic diagram of an embodiment of the present invention, and FIG. 4 is a block diagram of a signal processing section of the embodiment of the present invention. Moreover, FIG. 5 is an example of the S10 curve obtained by putting the present invention into practice. 33... Light source, 35.41... Photoelectric conversion section, 3
51... Photoelectric conversion element, 352... Condenser lens,
9-353...filter, so...light amount signal, 3e.
...Image signal, 42...Amplification circuit, 43...Differential m
l Sardine, 44...Binarization circuit, 45...Counter,・
16... Comparator. Agent Kuwabara d Bi10- -1- 1-r--- Fang-30

Claims (1)

[Scope of Claims] In an image discrimination device that forms an image signal corresponding to a light intensity signal obtained by scanning a document, processes the image signal, and performs image discrimination, an image obtained based on the light intensity signal is provided. A differentiation circuit that differentiates the time-series output of a signal, and a comparison of the output signal from the differentiation circuit with a specific level [1. A binarization circuit that outputs a binarized pulse signal, and an output pulse from the binarization circuit An image discriminating device l is provided with a counter for counting and discriminates an image based on a signal output from the counter.