JPH10137695A - Fruit sorting equipment - Google Patents

Abstract

(57) [Summary] [Problem] The fruit sorting work is easy, the judgment accuracy of the grade of fruits and vegetables is high regardless of the variety or the grade, and it is inexpensive without the need for dedicated equipment for class judgment. To provide a fruit device. SOLUTION: An imaging device 1 for photographing fruits and vegetables to be sorted.
And, among a plurality of types of parameters whose values are calculated by analyzing imaging of fruits and vegetables, the type of parameter having a relatively high correlation with the weight of fruits and vegetables in each variety of fruits and vegetables is associated with the variety of fruits and vegetables. A memory 45 for storing the data, and a data processing unit 4 for receiving the setting of the variety of the fruits and vegetables to be selected.
And, among the parameter values calculated by analyzing the imaging of the fruits and vegetables to be selected, a parameter value of a type having a relatively high degree of correlation with the weight in the set variety is extracted, and based on this parameter value A weight estimating unit 46 for estimating the weight of fruits and vegetables to be selected and a class determining unit 47 for determining the class of fruits and vegetables to be selected based on the estimated weight are provided.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to the production of fruits and vegetables, such as fruits and vegetables, in an attribute of, for example, 4 L, 3 L by weight.
The present invention relates to a fruit sorting device for classifying into classes such as L, 2L, L, M, S, and 2S.

[0002]

2. Description of the Related Art Various sizes brought to a shipping union,
Fruits and vegetables, such as fruits and vegetables, are shipped to the market in several grades depending on their appearance, such as pests, wounds, color, luster, shape, etc. Fruits are classified and classified.

[0003] An apparatus for automatically performing such fruit selection is disclosed in JP-A-57-135076, JP-A-57-197074, and
No. 58-156382, these devices determine the grade of fruits and vegetables based on the roundness and the like obtained from imaging of fruits and vegetables captured by a video sensor. The class of the fruits and vegetables is determined by comparing the cross-sectional area, the maximum diameter, and the like with reference values.

[0004]

However, for example, even if the persimmons are the same, the astringent persimmons such as Hirakaki Mugi and Tone Persimmons have a square flat shape, and the sweet persimmons such as Fuyu Persimmon have a rounded shape. It is a plump shape. Therefore, even when the maximum diameter or cross-sectional area of the two persimmons of the persimmon type and the sweet persimmon type is equal, the weight varies depending on the shape and height.

[0005] Further, the way of deformation of sweet persimmon-based persimmons tends to deform to an elliptical cross section, but the way of deformation of persimmon persimmon-based persimmons tends to protrude like a knob. Therefore, in the persimmon system, the correlation between the cross-sectional area and the weight seems to be relatively unchanged between the deformed persimmon and the persimmon that is close to a perfect circle,
It is considered that the correlation between the weight and the cross-sectional area of the deformed persimmon is relatively low in the astringent persimmon system. Thus, height information is needed to more accurately estimate the weight, and the fruits and vegetables must be photographed from both above and side.

Therefore, there is a variety of fruits and vegetables such as persimmons, which have different shapes in shape, and the types of parameters having a high degree of correlation with weight vary depending on the variety, and the degree of correlation between parameters and weight varies due to deformation. When the class is determined based on the same type of parameter, if the correlation between the parameter and the weight is not high enough to estimate the weight, the difference from the actual weight will increase, and depending on the type and in the same type Also, the accuracy of class judgment varies depending on the class (outer shape).

In order to avoid such variations,
At the fruit sorting site, it is necessary to reset the cross-sectional area and the maximum diameter threshold for classifying according to the variety, for example, so as to be as close as possible to the actual weight of the fruits and vegetables. It is.

[0008] In addition, if the weight of fruits and vegetables is directly measured with a load cell or the like and classified into classes, the classification can be easily and accurately classified into classes. However, dedicated equipment is required for judging the class of fruits and vegetables. Cost is high.

The present invention has been made in order to solve such a problem, and among the parameters calculated by analyzing the imaging, the types of parameters having a high degree of correlation with the attributes of fruits and vegetables, for example, weight. Is stored in association with the type and variety of fruits and vegetables, or information indicating the degree of correlation with an attribute such as weight is stored in association with the type of parameter, and the weight is referenced with reference to this storage information. By estimating attributes such as the weight of fruits and vegetables to be selected based on parameters having a relatively high degree of correlation with such attributes, fruit selection is easy, and fruits and vegetables are irrespective of varieties or grades. It is an object of the present invention to provide an inexpensive fruit sorting apparatus which has high class determination accuracy and does not require a dedicated facility for class determination.

[0010]

According to a first aspect of the present invention, there is provided a fruit sorting apparatus.
In a fruit sorting device that classifies fruits and vegetables according to their attributes,
Means for photographing fruits and vegetables to be fruit selected, and among a plurality of parameters whose values are calculated by analyzing the imaging of fruits and vegetables,
Means for storing a type of parameter having a relatively high degree of correlation with the attribute of the fruits and vegetables in each type of fruits and vegetables in association with the type of fruits and vegetables, and means for receiving the setting of the type of fruits and vegetables to be selected. Of the parameters calculated by analyzing the imaging of fruits and vegetables to be selected, parameters of a type having a relatively high degree of correlation with the attribute of the fruits and vegetables in a set type are extracted, and And a means for estimating the attribute of the fruit or vegetable to be selected based on the attribute, and a means for determining a class of the fruit or vegetable to be selected based on the attribute estimated by the means.

[0011] The fruit sorting device of the second invention is characterized in that the attribute in the first invention is the weight of fruits and vegetables.

A fruit sorting apparatus according to a third invention is characterized in that the kinds in the first or second invention are varieties belonging to the same kind of fruits and vegetables.

A fruit sorting apparatus according to a fourth aspect of the present invention is a fruit sorting apparatus for classifying fruits and vegetables in accordance with their weights. Means for storing information indicating the degree of correlation between the parameter type and the weight of the fruit or vegetable in association with the parameter type; and a parameter value calculated by analyzing the imaging of the fruit or vegetable to be selected. Means for extracting a kind of parameter value having a relatively high degree of correlation with, and estimating the weight of the fruits and vegetables to be selected based on the parameter value; and fruits and vegetables to be selected based on the weight estimated by the means. Means for determining the class of

According to a fifth aspect of the present invention, there is provided a fruit sorting apparatus according to any one of the first to fourth aspects, wherein the means for photographing fruits and vegetables to be fruit selected comprises: means for photographing from the direction in which the cross-sectional shape of the fruits and vegetables is obtained; Means for photographing from the direction in which the image is obtained.

The fruit sorting apparatus according to the first to third aspects of the present invention, when accepting the setting of, for example, a variety as a kind of fruit to be sorted, takes a picture of the fruit to be sorted, and calculates a parameter calculated by analyzing the image of the fruit and vegetable. Among the values, in the set variety, a parameter value of a kind stored as a parameter having a relatively high degree of correlation with, for example, weight, which is an attribute of fruits and vegetables, is extracted, and based on this parameter value, a fruit selection target is extracted. The weight of the fruits and vegetables is estimated, and the class of the fruits and vegetables to be selected is determined based on the estimated weights.

A fruit sorting apparatus according to a fourth aspect of the present invention captures fruits and vegetables to be sorted and stores the parameter values calculated by analyzing the imaging of the fruits and vegetables as parameters having a relatively high degree of correlation with weight. Are extracted, the weight of the fruits and vegetables to be selected is estimated based on the parameter values, and the class of the fruits and vegetables to be selected is determined based on the estimated weight.

Therefore, in the fruit sorting apparatus of the present invention, for example, there are varieties having different shapes such as persimmons, and parameters having a high degree of correlation with attributes such as weight vary depending on the varieties. Uses imaging of fruits and vegetables used to determine the grade of fruits and vegetables, without the need for special equipment, for the class of fruits and vegetables whose correlation degree between the parameter for estimating attributes and this attribute varies depending on the deformation of fruits and vegetables And can be determined with high accuracy.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing the structure of a fruit sorting device according to the present invention. In the drawing, OB is a persimmon to be selected, and is placed on a plate arranged on a conveyor CO at an appropriate interval and transported in the arrow direction in the drawing.
In the figure, reference numeral 1 denotes an image pickup apparatus including two cameras disposed above and on the side of the conveyor CO, respectively, for photographing the upper surface and the side surface of a persimmon to be selected.

Conveyor C upstream of the imaging device 1 in the transport direction
On the side of O, the passage of persimmons conveyed by the conveyor CO is detected, and the detection signal is obtained based on the data collection unit 3 that captures the image of the imaging device 1 and the grade and grade of the fruits and vegetables to be sorted. A rotary encoder 2 is provided for giving a command to a sorting device 20 for sorting fruits and vegetables and outputting the command to a sorting control device 10 for controlling sorting of fruits and vegetables.

The data collection unit 3 includes the rotary encoder 2
In synchronization with the detection signal given by the imaging device 1, the image of the plane and the side of the persimmon taken by the imaging device 1 is taken in, and the image is processed and analyzed. The data processing unit 4 receives the setting of the variety to be selected by the variety setting unit 5, calculates the measurement parameter values corresponding to the set variety from the image processing result of the data collection unit 3, and estimates the weight of the fruits and vegetables. It consists of the following cores:

The area calculating section 41 calculates the cross-sectional area of the persimmon from the analysis result of the persimmon plane image taken in by the data collecting section 3. The height calculation unit 42 calculates the height of the persimmon from the analysis result of the side image of the persimmon captured by the data collection unit 3. The maximum diameter calculation unit 43 and the minimum diameter calculation unit 44 are based on the center of gravity obtained from the analysis result of the persimmon planar image captured by the data collection unit 3,
Calculate the maximum and minimum diameter of the persimmon cross section.

The memory 45 stores measurement parameters having a high degree of correlation with the weight in each type of astringent persimmons such as Hirakaki Mugi and Tone persimmons and sweet persimmons such as Fuyu persimmons, for calculating the weight. Is stored in association with the type together with the arithmetic expression of (1). The memory 45 stores the upper and lower limits of the weight of each class as thresholds for class determination.

FIG. 2 is a conceptual diagram showing an example of measurement parameters calculated by analyzing an image of a persimmon and the storage state of the memory 45. As shown in the figure, for example, in astringent persimmons such as Hirakaki No Persimmon and Tone Persimmon, area and height are measurement parameters with a high degree of correlation with weight, and in sweet persimmons such as Tomi Persimmon, flat cross-sectional areas and heights are used. In addition to the above, the minimum diameter, the maximum diameter, and the height of the flat section are measurement parameters having a high degree of correlation with the weight. Therefore, the memory
In 45, an arithmetic expression for calculating the weight is stored together with the type of each measurement parameter in association with the type.

The weight estimating unit 46 refers to the type of the measurement parameter stored in the memory 45 in association with the type,
From the parameter values calculated by the area calculation unit 41, the height calculation unit 42, the maximum diameter calculation unit 43, and the minimum diameter calculation unit 44,
The type of the measurement parameter value of the type stored in association with the type set by the type setting unit 5 is extracted to estimate the weight of the fruits and vegetables. The class determining unit 47 determines the class of the fruit or vegetable based on the weight estimated by the weight estimating unit 46 and the weight threshold of each class stored in the memory 45.

The sorting control device 10 includes a rotary encoder 2
Class determination unit 47 in synchronization with the aforementioned detection signal given from
And outputs a command signal to the sorting device 20 that sorts persimmons according to the grade and grade of the persimmon conveyed on the conveyor CO.

As in the above embodiment, in addition to storing the measurement parameters suitable for estimating the weight in each type in the memory 45 in association with the type, the degree of correlation with the weight is also determined. For example, if the correlation coefficient is stored for each measurement parameter, for example, even in a fruit sorting device that captures only a flat image of fruits and vegetables, the correlation with weight can be calculated from among the measurement parameters calculated from the imaging. The class can be accurately determined by using the measurement parameter having a higher degree.

For example, FIGS. 3 to 5 are diagrams showing the correlation between the weight of a persimmon-based house sword and various measurement parameters. FIG. 3 shows the relationship between weight and area × height (FIG. 3) and weight and area (FIG. 4). ) And weight and the correlation coefficient obtained from the regression line of the perimeter of the plane section (FIG. 5), the correlation between weight and area × height is the highest among these.

However, only one camera is installed above the conveyor CO, and in a fruit sorting device that captures only a planar image by this, a height is not required as a measurement parameter. By comparing the weight and the respective correlation coefficients such as the perimeter of the plane section, it can be seen that the measurement parameter having a higher correlation coefficient or a combination thereof may be used for weight estimation. Further, in this case, the correlation coefficient between the weight and the area is quite close to the correlation coefficient between the weight and the area × the height and is high enough to estimate the weight, so that the class can be determined with high accuracy.

In this case, either the correlation coefficient or the standard deviation may be stored in the memory 45 as information indicating the degree of correlation with the weight.

The procedure of class determination by the fruit sorting apparatus of the present invention as described above will be described with reference to the flowchart of FIG. The data collection unit 3 binarizes the planar image of the fruits and vegetables photographed by the camera for photographing the upper surface of the imaging device 1 (S1), tabulates the grayscale value of each pixel of the image, and coordinates the periphery (pixels) of the plane cross section of the fruits and vegetables Is detected (S2). The area calculation unit 41
The number of pixels is counted to calculate the area (S4, S5).

The maximum diameter calculating section 43 and the minimum diameter calculating section 44
Calculates the coordinate data of the detected surrounding coordinates (S
6) The center of gravity of the persimmon in the plane section is calculated from the X moment and the Y moment (S3, S7). Next, the length of the line segment between the center of gravity and each point of the peripheral coordinates is obtained (S8), and the maximum diameter and the minimum diameter are calculated (S9).

On the other hand, the data collection unit 3 binarizes the side image of the fruits and vegetables photographed by the camera for photographing the sides of the image pickup apparatus 1 (S10), sums up the grayscale values of each pixel of the image, and surrounds the periphery of the fruits and vegetables. Detect coordinates. The height calculator 42 calculates the height of the fruit or vegetable from the difference between the maximum value and the minimum value of the y coordinate in the detected surrounding coordinates (S11, S12).

The weight estimating section 46 is stored in the memory 45 as an optimal measurement parameter for weight estimation of the set varieties among the basic data for weight estimation calculated by the respective calculating sections as described above. The weight of the fruits and vegetables is estimated by using the measurement parameter values (S13). The class determining unit 47 compares the estimated weight with the weight threshold of each class stored in the memory 45 to determine the class of the fruits and vegetables (S1).
4) The determination result is given to the selection control device 10.

In the present embodiment, the case where the fruits and vegetables to be selected are persimmons has been described. However, the fruit selection apparatus of the present invention is not limited to persimmons, but citrus fruits (citrus orange, orange) and apples (fuji, wanglin). Vegetables and fruits such as pears (twentieth century pears, pears), eggplants (long eggplants, round eggplants), etc. that contain multiple varieties with different appearance shapes, and their deformations correlate with measurement parameters for weight estimation and weight The present invention is also applicable to class determination of fruits and vegetables having different degrees, and the same effects as those of the present embodiment can be obtained.

The means for detecting the passage of fruits and vegetables conveyed on the conveyor CO is not limited to the rotary encoder 2,
It may be a load cell, a proximity switch, or the like.

The criteria for classifying fruits and vegetables are not limited to weight. Even if other attributes are stored, by storing measurement parameters having a high degree of correlation with this attribute, the same effect as in the present embodiment can be obtained. can get.

In this embodiment, the types of measurement parameters having a high degree of correlation with attributes such as weight are stored in association with varieties such as astringent persimmons and sweet persimmons, and the weight is stored using the stored information. Although an example of estimating and classifying with high accuracy has been described, it is also possible to classify not only varieties but also a plurality of kinds of fruits and vegetables such as persimmons, apples and citrus fruits by using measurement parameters suitable for each.

[0038]

As described above, the fruit sorting apparatus according to the present invention can determine the types of parameters having a high degree of correlation with the attributes of fruits and vegetables, for example, weight, from among the parameters calculated by analyzing the imaging.
The information indicating the degree of correlation with attributes such as weight is stored in association with the type and variety of the fruits and vegetables, or stored in association with the type of parameter. Since attributes such as the weight of fruits and vegetables for fruit selection are estimated based on parameters having a relatively high degree of correlation with the attributes, fruit selection work is easy, and regardless of the variety or grade, This provides an excellent effect that the judgment accuracy is high, the class judgment does not require special equipment, and the cost is low.

[Brief description of the drawings]

FIG. 1 is a block diagram of a fruit sorting device of the present invention.

FIG. 2 is a conceptual diagram of measurement parameters of each type and a storage state thereof in the fruit sorting apparatus of the present invention.

FIG. 3 is a diagram showing a correlation between weight and area × height of a house blade.

FIG. 4 is a diagram showing a correlation between the weight and area of a house blade.

FIG. 5 is a diagram showing a correlation between the weight of a house blade and a perimeter.

FIG. 6 is a flowchart of a processing procedure of class determination by the fruit sorting device of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image pickup device 2 Rotary encoder 3 Data collection part 4 Data processing part 41 Area calculation part 42 Height calculation part 43 Maximum diameter calculation part 44 Minimum diameter calculation part 45 Memory 46 Weight estimation part 47 Class judgment part 5 Type setting part OB Selection Target fruits and vegetables CO conveyor

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Shoji Tanaka 64 Ishizukita-cho, Sakai-shi, Osaka Inside Kubota Sakai Works

Claims (5)

[Claims] 1. A fruit sorting apparatus for classifying fruits and vegetables according to their attributes, wherein: a means for photographing fruits and vegetables to be fruit selected; and a plurality of parameters of which values are calculated by analyzing the imaging of fruits and vegetables. Means for storing a parameter type having a relatively high degree of correlation with the attribute of the fruit and vegetable in each type of fruit and vegetable in association with the type of fruit and vegetable, and means for receiving the setting of the type of fruit and vegetable to be selected Of the parameters calculated by analyzing the imaging of fruits and vegetables to be selected, a parameter value of a type having a relatively high degree of correlation with the attribute of the fruits and vegetables in a set type is extracted. A fruit sorting device, comprising: means for estimating the attribute of a fruit to be selected based on the attribute; and means for determining a class of the fruit or fruit to be selected based on the attribute estimated by the means. 2. The method according to claim 1, wherein the attribute is a weight of the fruits and vegetables.
A fruit sorting device as described. 3. The fruit sorting device according to claim 1, wherein the types are varieties belonging to the same kind of fruits and vegetables. 4. A fruit sorting device for classifying fruits and vegetables according to their weight, means for photographing fruits and vegetables to be sorted, types of parameters whose values are calculated by analyzing imaging of fruits and vegetables, Means for storing information indicating the degree of correlation with the weight in association with the type of parameter; and, among the parameter values calculated by analyzing the imaging of the fruits and vegetables to be selected, the degree of correlation with the weight is relatively high. Means for extracting a parameter value of a higher kind, estimating the weight of the fruits and vegetables to be selected based on the parameter values, and means for determining the class of the fruits and vegetables to be selected based on the weight estimated by the means. A fruit sorting device comprising: 5. The means for photographing fruits and vegetables to be selected has means for photographing from the direction in which the cross-sectional shape of the fruits and vegetables can be obtained, and means for photographing from the direction in which the height of the fruits and vegetables can be obtained. 5. The fruit sorting device according to any one of 4.