JPS63200279A - Hue deciding device - Google Patents

Abstract

PURPOSE:To automatically evaluate the hues of colored materials like fruits, etc., that show complicated color tones by suppressing the variance in deciding results, by recognizing a processing object as a color picture and dividing this picture into fine picture elements for evaluation. CONSTITUTION:Apples 15 put on a pallet 13 are carried by a carrying means 11 and photographed by a color TV camera 17. Picture signals sent from the camera 17 are inputted to a processor 23. The processor 23 divides those photographed pictures for each picture element and evaluates the tinges of each picture element based on the absolute values of the hue elements of red, green and blue components as well as the ratio obtained among those absolute values. Then the integrated evaluation is carried out for overall tinges.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for determining the hue of an object exhibiting a complex hue, such as a fruit such as an apple.

First right to disaster Conventionally, judgments regarding the color of fruits such as apples have been made.

It was done manually. However, the determination based on human senses has been problematic because the determination results vary depending on individual differences, environment, etc.

A known method for automating the determination of colored objects is to use a photodiode or the like. However, in this method, the detection range is a spot, so it has been impossible to determine the entirety of objects with complex color combinations, such as fruits such as apples.

Japanese Utility Model Application No. 57-136485 describes a device that uses a color television camera to capture the color of fruits and select the quality of the fruits. However, this document only describes that a signal from a color television camera and a preset reference value are compared and comprehensively determined, but does not provide a specific determination method.

Also, JP-A-61-41905, JP-A No. 61-419
Publication No. 10 describes an apparatus for sorting the size of fruits, etc. from image information obtained by imaging fruits, etc., but does not disclose determination of color.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for automatically determining colored objects exhibiting complex color tones, such as fruits.

The color determination device of the present invention comprises: a conveying means for conveying spherical or block-shaped objects to be processed having different hues; an imaging means for taking an image of the object to be processed conveyed by the conveying means; The image captured by the imaging means is divided into pixels, and the absolute values of the red, green, and blue hue elements of each pixel and their ratios are evaluated to evaluate the hue of each pixel. and a hue determination processing means for performing a comprehensive evaluation of the overall hue.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

FIG. 1 is an explanatory diagram showing an embodiment of the present invention in the case of sorting apples. Apples 15 placed on a pallet 13 by a conveyor 11 such as a belt conveyor
is moving on the fruit sorting line. When the pallet 13 enters the imaging range, the detector 21 detects this, and the color television camera 17 captures the image. The captured image is processed by the processing device 23 to evaluate the hue of the apple. Reference numeral 19 denotes a lighting device that illuminates the fruit in order to maintain constant lighting conditions for color photography.

FIGS. 2 and 3 are explanatory diagrams showing a method of processing this image information. The captured image is divided into minute pixels, and the video signals of the red image, recorded image, and blue image for each pixel are sent, and converted into digital information by an A/D converter.
It is latched and recorded in each image memory. The timing of imaging by the camera is determined by the gate circuit 2 based on the signal from the detector 21.
The imaging timing is determined by outputting the imaging timing from 5.

Based on the information from the image memory, the color of the fruit is determined by an arithmetic device such as a microcomputer. The data for the same pixel obtained from each image memory, that is, the red image component value r obtained from the red image, the recorded image component value g from the recorded image, and the blue image component value from the blue image, are processed by the arithmetic unit as r, g
, b is used to evaluate the hue. Depending on the type of fruit, the ratio of two colors such as r and g, r and g and b is sufficient.1For example, in the case of apples, the relationship between red, green, and red is important.

It can also be evaluated using r and g.

As is clear from the above explanation, "the ratio of the red, green, and blue color components of each pixel" in the claims of this application does not necessarily mean the ratio of these three components. Instead, it may be a ratio of two components, such as a red component and a blue component, and this can be easily understood if we consider that the remaining one component is zero.

After completing the evaluation for all pixels, the color of the entire fruit is evaluated from the distribution of evaluation values for each pixel. Furthermore, the overall hue can be evaluated simply from the average value of all pixels.

If it is necessary to classify the evaluation results into grades, etc., it is also possible to set a threshold in advance and perform the ranking.

In the captured image, unnecessary pixels corresponding to areas that do not accurately reflect the color of the fruit, such as the background, reflective parts of the light source (shiny parts), dirt attached to the surface of the fruit, foreign substances such as pesticides, etc. Since this information may cause misjudgments, this information is deleted as unnecessary data. Fruits vary in size and shape, and the positions on the pallet are often uneven, so it is practically impossible to uniformly determine the positions of unnecessary pixels in advance. However, since the color of the fruit is roughly known in advance, upper or lower limits are set for the absolute values and ratios of each of r, g, and b, and these set values are compared to identify those that exceed the limits. It can be removed as an unnecessary pixel.

Next, we will explain in more detail the case of deleting unnecessary pixels from red fruits such as apples and sorting them into several groups from greenish to reddish according to the degree of redness. .

The background was black, and a white light source was used as the light source. The captured image is schematically shown in FIG. 4, where the pixel pk captures the background 25, and the pixel P captures the fruit surface 27.
3. Pixel P imaging the reflection part 29 of the light source on the fruit surface;
There is.

The r value (=rb) of the pixel that images the background 25 is.

Pb is deleted as an unnecessary pixel by utilizing the fact that it is smaller than the r value (r, ) of the pixel imaging the fruit surface 25.

That is.

A pixel (Pn) where rn<α (α is a predetermined constant) is identified as a pixel that captures an image of the background 25, and is excluded from the evaluation target.

The P response is removed as an unnecessary pixel by taking advantage of the fact that the ratio of the b value to the r value is increased in the pixel that images the reflective part 29 of the light source compared to the pixel that images the fruit surface 27. .

That is, for the remaining N pixels after removing unnecessary data, the obtained r value, 1g value, and b value are used to evaluate the hue from the ratio thereof. As in this example, when red fruits are classified in order from greenish to greenish, it is also possible to evaluate only by the ratio of the r value and the g value as shown in the following equation.

The hue of the entire fruit can be determined from the distribution of fn obtained from each pixel. Furthermore, if only the average color of the entire fruit is to be determined, it is also possible to evaluate using the arithmetic mean value of fn.

CF=average value of all fruits) Based on the obtained F value, the fruits can be grouped according to a preset threshold value. The smaller the F value, the greener the image, and the larger the F value, the more red the image appears.

In the above description, the background and the reflection part of the light source have been described as unnecessary pixels, but other unnecessary pixels such as parts to which foreign substances such as pesticides and dust are attached can also be processed in the same way.

FIG. 5 is a flowchart showing the above-mentioned hue determination method.

Note that although the above description has focused on fruits, especially apples, the same applies to other objects to be processed.

According to the present invention, by capturing the object to be processed as a color image and further dividing it into micropixels for evaluation, it is possible to evaluate the color of fruits, etc., which have complex combinations of colors, and to reduce the variation in judgment results. can be evaluated while suppressing Furthermore, since unnecessary data such as the reflective portion of the light source can be deleted before making the determination, even more accurate evaluation is possible.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram showing a hue determination device of the present invention. FIGS. 2 and 3 are explanatory diagrams showing a single image processing method. FIG. 4 is an explanatory diagram showing a method of deleting unnecessary data. FIG. 5 is a flowchart showing an example of the determination method of the present invention. 11... Conveyance means 13... Pallet 17...
・Television camera 19...Lighting device 21...Detector 23...Processing device Patent applicant Sumitomo Heavy Industries, Ltd. Figure 1 Figure 2

Claims (1)

[Scope of Claims] 1. A transport means for transporting spherical or block-shaped objects to be processed having different hues; an imaging means for imaging the objects to be processed transported by the transport means; and imaging by the imaging means. The resulting image is divided into pixel by pixel, and the hue of each pixel is evaluated by taking the absolute values of the red, green, and blue hue elements of each pixel and their ratios. From this evaluation, the overall hue is calculated. 1. A hue determination device comprising: a hue determination processing means for performing comprehensive evaluation. 2. The hue determination processing means does not accurately reflect the hue of an object to be processed, such as a pixel that captures an image of the background behind the object to be processed, or a pixel that captures an image of a reflective part of a light source on the surface of the object to be processed. 2. The hue determination device according to claim 1, further comprising a removal means for removing unnecessary pixels from the evaluation target. 3. The removal means compares a red, green, or blue component value from each pixel or a ratio thereof with a setting value set in advance according to the non-processing object, and removes pixels exceeding the setting value as unnecessary pixels. 3. The hue determination device according to claim 2, wherein the hue determination device removes as follows. 4. The hue determination processing means removes the unnecessary pixels,
Pixel red component value (r) for each removed remaining pixel,
The hue evaluation value (fn) of each pixel is obtained from the pixel green component value (g) and the pixel blue component value (b), and the evaluation value (fn)
The patent claim is characterized in that it has a means for calculating the overall evaluation value (F) of the overall hue from the average value of The hue determination device according to range 3.