Fruit and vegetable shape discrimination device

Abstract

To provide a vegetable / fruit shape discriminating device that is able to easily and accurately discriminate a vegetable / fruit shape.SOLUTION: A vegetable / fruit shape discrimination device includes: an emission unit 5 that emits an X-ray to a vegetable / fruit 1 as an inspection target; a detection unit 6 that detects the X-ray emitted from the emission unit 5 and has penetrated the inspection target; an image formation unit 8 that forms an X-ray penetration image of the inspection target according to an amount of X-ray detected by the detection unit 6; and a discrimination unit 10 that specifies a shape group corresponding to the inspection target by comparing information on the X-ray penetration image formed by the image formation unit 8 with information on a reference image previously stored in a reference image storage unit 9.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This invention relates to a device for discriminating the shape of fresh fruits and vegetables.

Background Art

[0002] As criteria for the quality of fresh fruits and vegetables, in addition to taste, nutritional value, size, etc., visual beauty is also an important factor. Especially for expensive fruits, a regular shape is preferred. Therefore, when classifying fresh fruits and vegetables, shape discrimination becomes an important factor. Conventionally, in the process of transporting fresh fruits and vegetables by a conveyor or the like, veteran workers may visually judge the individual size and shape of the fresh fruits and vegetables and pack them into boxes according to grades. However, due to differences in the abilities of workers, the discrimination results may vary or the work efficiency may be poor. Therefore, a device that photographs fresh fruits and vegetables and discriminates their shapes from the images has been considered.

[0003] For example, a device for discriminating the size and shape of an inspection object being transported by a conveyor during the transportation process is described in Patent Document 1. This device photographs the inspection object from above with a line sensor S having a visual field in the width direction from above the moving inspection object, binarizes the photographed image, and discriminates the size and shape from the data. It has been used for discriminating objects such as eggplants.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] As described above, images of fruits and vegetables taken from above by the line sensor S are similar to those seen with the naked eye from above, and the shape of the shaded area on the opposite side of the line sensor S may not be discernible. For example, as shown in Figure 5, strawberries 1a and 1b have different tip shapes when viewed from the side. However, when these strawberries 1a and 1b are placed on a plane 2, and images 1a' and 1b' are captured by moving strawberries 1a and 1b in the direction of arrow α relative to a line sensor S placed above them, the images are almost identical, as shown in Figure 6.

[0006] Images 1a' and 1b' do not allow us to distinguish the difference in the shape of the tip between strawberry 1a and strawberry 1b. Ideally, the well-formed cone-shaped strawberry 1a should be of a higher grade than the flattened tip of strawberry 1b, but the above device could not accurately distinguish between them. On the other hand, by changing the orientation of the strawberries in various ways relative to the line sensor S so that no shadows are cast, it is possible to obtain accurate three-dimensional images. However, changing the orientation of the strawberries without damaging them would complicate the device and increase the inspection time.

[0007] Furthermore, devices with a greater number of imaging means have been considered to acquire images from various angles simultaneously (see Patent Document 2). However, such devices become structurally complex and expensive.

[0008] The objective of this invention is to provide a fruit and vegetable shape discrimination device that can easily and accurately determine the shape of fruits and vegetables. [Means for solving the problem]

[0009] The present invention provides a transport path that allows X-rays to pass through and transports produce, which is the object to be inspected, in one direction, and the transport RoadThe system comprises: one irradiation unit that irradiates the object to be inspected with X-rays from above; a detection unit located opposite the irradiation unit across the transport path and detecting X-rays irradiated from the irradiation unit and transmitted through the object to be inspected; an image forming unit that forms one X-ray transmission image for each object to be inspected according to the amount of X-rays detected by the detection unit; a reference image storage unit that stores information of reference images corresponding to a plurality of pre-set shape groups; and a discrimination unit that compares the information of the X-ray transmission image formed by the image forming unit with the information of the reference images and identifies the shape group containing the reference image data closest to the data of the X-ray transmission image as the shape group corresponding to the object to be inspected. The reference image storage unit is characterized in that it irradiates samples of fruits and vegetables that should be classified into each shape group with X-rays from various angles to form multiple X-ray transmission images, and stores these image groups in advance as reference images for the corresponding shape groups. [Effects of the Invention]

[0011] Book According to the invention, two-dimensional information and thickness data of fruits and vegetables can be acquired simultaneously. Therefore, more accurate shape identification is possible, and product grading based on shape can be performed without relying on experienced personnel.

[0012] moreover Furthermore, since the information obtained from X-ray transmission images of a sample irradiated with X-rays from multiple directions is stored as reference image information, it is possible to find a reference image that matches the object being inspected and determine its shape, regardless of how the fruit or vegetable being inspected is placed. Furthermore, even if there are many reference images, they can be stored in advance, so it does not take time, as would be the case when changing the orientation of the object being inspected and acquiring multiple image information during inspection. [Brief explanation of the drawing]

[0013] [Figure 1] Figure 1 is a block diagram of a shape discrimination device according to an embodiment of the present invention. [Figure 2]Figure 2 is an X-ray transmission image created by the shape discrimination device of the embodiment. (a) is an X-ray transmission image of strawberry 1a. (b) is an X-ray transmission image of strawberry 1b. [Figure 3] Figure 3 is an example of the display of the shape discrimination result of the embodiment. [Figure 4] Figure 4 is an explanatory diagram of the method for forming a reference image in the embodiment. [Figure 5] Figure 5 is an example of the appearance of a strawberry that is the object to be inspected. [Figure 6] Figure 6 is an example of a determination image by the conventional shape discrimination method.

Mode for Carrying Out the Invention

[0014] [Embodiment] The embodiment of this invention will be described below using the example of discriminating the shape of strawberry 1. Figure 1 is a block diagram of the shape discrimination device of the embodiment. Figure 2 is an X-ray transmission image formed by the image forming unit of the embodiment, (a) is an X-ray transmission image of strawberry 1a, and (b) is an X-ray transmission image of strawberry 1b. Figure 3 is an example of the display of the shape discrimination result of the embodiment. Figure 4 is an explanatory diagram of the method for forming a reference image.

[0015] As shown in Figure 1, the shape discrimination device of this embodiment includes a conveying means 4 such as a conveyor that places a plurality of strawberries 1 as the inspection object on a tray 3 and conveys them in the direction of arrow α. A plurality of strawberries 1 are placed on the tray 3 in rows vertically and horizontally. Both the tray 3 and the conveying means 4 are formed of a material that transmits X-rays. In addition, an irradiation unit 5 that irradiates X-rays toward the strawberry 1 is provided above the conveying means 4, and a detection unit 6 that detects X-rays is provided at a position facing the irradiation unit 5 across the conveying means 4. A data processing unit 7 is connected to the detection unit 6.

[0016] The above-described irradiation unit 5 is capable of irradiating X-rays in a direction orthogonal to the conveyance direction α of the strawberry 1, and has a function of covering the entire width of all the strawberries 1 passing under the irradiation unit 5. Further, the detection unit 6 is a line-shaped sensor having a length corresponding to the irradiation range of the X-rays irradiated from the irradiation unit 5.

[0017] The above-described detection unit 6 detects the X-rays transmitted through the strawberry 1, the tray 3, and the conveyance means 4 with cells arranged in a line, and causes the intensity data of the detected X-rays to be input to the data processing unit 7. Based on the intensity data of the X-rays input from the detection unit 6, the data processing unit 7 includes an image forming unit 8 that forms an X-ray transmission image of the inspection object (strawberry), a reference image storage unit 9 that stores information on the reference image, and an image forming unit 8. And a discrimination unit 10 that discriminates the shape of the inspection object by comparing the X-ray transmission image of the inspection object formed with the information of the reference image.

[0018] The reference image stored in the reference image storage unit 9 is an image serving as a reference for discriminating the shape of the inspection object. Although details will be described later, it is an X-ray transmission image of a sample classified for each shape group corresponding to a predetermined grade. For example, if there are five predetermined grades, it is an X-ray transmission image of a sample classified for each of the five shape groups.

[0019] The above-described discrimination unit 10 has a function of specifying the shape group of the strawberry 1 by comparing the information of the X-ray transmission image of the inspection object formed by the image forming unit 8 with the information of the reference image. Then, the result of the shape discrimination of the strawberry 1 by the discrimination unit 10 is output to the output unit 11. The output unit 11 outputs information on the grade corresponding to the specified shape group as the result of the shape discrimination. In FIGS. 1 and 4, the flow of data is indicated by a dashed arrow.

[0020] [Operation, effect, etc.] An example of discriminating the shape group of the strawberry 1 using the shape discrimination device of the embodiment will be described. As shown in Figure 1, the strawberries 1 placed on the tray 3 are transported, and as they pass directly beneath the irradiation unit 5, the strawberries 1 are irradiated with X-rays. Although multiple strawberries 1 are arranged in the width direction of the transport means 4 on the tray 3, this explanation will focus on only one strawberry 1.

[0021] X-rays irradiated onto strawberry 1 pass through strawberry 1 and tray 3 and are detected by detection unit 6. The transmitted X-ray intensity detected by detection unit 6 changes according to the thickness of the material in the area through which the X-rays have passed. Specifically, the transmitted X-ray intensity decreases exponentially with increasing thickness of the absorber interposed between the irradiation unit 5 and the detection unit 6, as shown by the following equation.

[0022]

number

[0023] Here, I0 is the X-ray intensity before passing through the absorber, I is the transmitted X-ray intensity after passing through the absorber, μ is the X-ray absorption coefficient, and t is the thickness of the absorber. Therefore, the thickness of the absorbent material for each pixel can be calculated from the transmitted X-ray intensity I detected by the detection unit 6 for each pixel.

[0024] Then, ignoring the influence of the bottom surface of tray 3 and assuming that strawberry 1 is made almost entirely of water, the equivalent water thickness can be calculated based on the water absorption rate μ and the detected transmitted X-ray intensity I. Therefore, the image forming unit 8 determines the above equivalent water thickness as the thickness t of strawberry 1. The image information, including the thickness data for each pixel calculated by the image forming unit 8, is the X-ray transmission image of strawberry 1.

[0025] Figures 2(a) and 2(b) show X-ray transmission images X1a and X1b formed by the image forming unit 8 for strawberries 1a and 1b, which have different shapes. These X-ray transmission images X1a and X1b are images formed by the image forming unit 8 based on the transmitted X-ray intensity that passed through the strawberries 1a and 1b when they were placed on the tray 3 and passed through the irradiation unit 5. Note that strawberries 1a and 1b are the same as those shown in Figure 5 used in the explanation of the conventional example, but are different strawberries with different shapes at the tip.

[0026] Then, the X-ray transmission images X1a and X1b in Figures 2(a) and 2(b) are color-coded according to the thickness data calculated by the image forming unit 8 based on the intensity data of the transmitted X-rays detected pixel by pixel by the detection unit 6, as described above. As shown in Figure 5, strawberries 1a and 1b, which were the subjects of the examination, have different tip shapes, but it was difficult to distinguish between them by visual inspection from above (see Figure 6). However, the X-ray transmission images X1a and X1b shown in Figures 2(a) and (b) clearly show that the tip shapes of the two strawberries are different.

[0027] On the other hand, the reference image storage unit 9 stores X-ray transmission images of strawberries as samples, classified according to shape groups corresponding to their grade, as reference images. Therefore, the discrimination unit 10 compares the X-ray transmission images X1a and X1b, each formed by the image forming unit 8, with the reference image information stored in the reference image storage unit 9. The procedure for comparing the X-ray transmission image information of the object to be inspected with the reference image information is pre-set in the discrimination unit 10.

[0028] The determination unit 10 then identifies the shape group of the reference image that is closest to each X-ray transmission image X1a, X1b, and outputs the grade of the identified shape group to the output unit 11. The output unit 11 displays an image, for example as shown in Figure 3, in which grade marks A, B, C, etc., indicating the shape grade, are placed at the positions of the strawberries 1 on the tray 3.

[0029] Next, the method for forming the reference image stored in the reference image storage unit 9 will be explained using Figure 4. As described above, the reference image is an X-ray transmission image of a sample included in the shape group corresponding to the grade. This reference image can be formed using the shape discrimination device of this embodiment. First, a strawberry 1s, which is a sample of a specific shape group G1, is irradiated with X-rays while being transported in the direction of arrow α between the irradiation unit 5 and the detection unit 6, and the transmitted X-rays are detected by the detection unit 6.

[0030] The detection unit 6 sends the intensity data of the detected transmitted X-rays to the image forming unit 8. Upon receiving this intensity data, the image forming unit 8 creates an X-ray transmission image including thickness data based on the intensity data. Up to this point, the procedure is the same as during the inspection described above. Subsequently, the created X-ray transmission image is stored in the reference image storage unit 9 as the reference image for shape group G1. Similarly, a reference image is formed for each shape group sample corresponding to all grades and stored in the reference image storage unit 9.

[0031] In this embodiment, a single sample is irradiated with X-rays from various angles to form multiple X-ray transmission images, and these image sets are used as reference images for a single shape group. For example, a strawberry 1s, which is a sample placed on the transport means 4, is rotated by an angle θ around axis O from the state shown in Figure 4, and an X-ray transmission image is formed and stored for each angle. Axis O intersects the direction of X-ray irradiation and passes through approximately the center of the strawberry 1s. However, the arrangement of the sample when forming the reference images is not limited to the above. Furthermore, for samples with shapes that are slightly deformed but still fall within the same grade, X-ray transmission images are formed in the same manner as described above and stored in the reference image storage unit 9 as reference images for the corresponding shape group.

[0032] In this embodiment, by using the X-ray transmission image information described above, it became possible to distinguish shapes that could not be distinguished using only two-dimensional information. Furthermore, since X-ray transmission images from multiple angles are stored as reference images, the system can identify the reference image corresponding to the X-ray transmission image of the object being inspected, even without strictly controlling the orientation of the object.

[0033] If X-rays were irradiated onto the object to be inspected from various angles to form an X-ray transmission image, the inspection would take too long. However, in this embodiment, reference images can be stored in advance, so even if the amount of information is large, the time required for the inspection will not increase significantly. Moreover, compared to systems requiring multiple imaging methods to obtain accurate three-dimensional images, this system allows for accurate shape identification with a simpler configuration.

[0034] In the above embodiment, the image forming unit 8 creates an X-ray transmission image with color-coded thickness data, but color-coding of the image is not essential. The X-ray transmission image information compared with the reference image information only needs to correspond to the thickness data of the object being inspected, which is an X-ray absorber, and numerical values ​​showing the thickness distribution may be compared as image information. Alternatively, the intensity data of transmitted X-rays may be used directly as image information.

[0035] The above example describes how to calculate thickness data from the X-ray intensity passing through the object being inspected. However, since the amount of X-ray absorption depends on the thickness of the absorbing material, the thickness data can be calculated based on either the amount of transmission or the amount of absorption. Furthermore, this shape discrimination device can simultaneously measure the weight of the object being inspected, strawberry 1, based on the amount of X-ray transmission or absorption.

[0036] Furthermore, the shape group grade can be output from the output unit 11 using AI learning (machine learning). That is, the AI ​​engine (equivalent to the judgment unit and output unit) is trained based on the reference image information and the shape group grade. For example, the reference image information and the shape group grade are used as the training data feature data and the ground truth data, respectively, to train the AI ​​engine. After training, the AI ​​engine outputs the shape group grade corresponding to the X-ray transmission image information of the input object to be inspected.

[0037] Furthermore, the shape discrimination device described above can be applied to fruits and vegetables other than strawberries. It is particularly suitable for objects that exhibit little variation in X-ray absorption rates. [Industrial applicability]

[0038] It is effective for identifying the shape of various fruits and vegetables. [Explanation of Symbols]

[0039] 1,1a,1b (Item to be tested) Strawberry 1 sample (strawberry) 5. Irradiation area 6. Detection Unit 7. Data Processing Unit 8 Image forming unit 9. Reference Image Storage Unit 10 Discrimination part 11 Output section X1a,X1b X-ray transmission image θ angle

Claims

[Claim 1] A transport path is used to carry the produce to be inspected in one direction, and X-rays can pass through it. One irradiation unit that irradiates the object to be inspected with X-rays from above the transport path, A detection unit is provided at a position opposite the irradiation unit across the transport path, and detects X-rays that have been irradiated from the irradiation unit and passed through the object to be inspected. An image forming unit that forms one X-ray transmission image for each object to be inspected according to the X-ray dose detected by the above detection unit, A reference image storage unit that stores information on reference images corresponding to each of several pre-set shape groups, The system includes a discrimination unit that compares the information of the X-ray transmission image formed by the image forming unit with the information of the reference image, and identifies the shape group containing the reference image data that is closest to the data of the X-ray transmission image as the shape group corresponding to the object to be inspected. The above reference image storage unit includes: For the fruits and vegetables of the sample that should be classified into each shape group, This system is characterized by irradiating X-rays from various angles to form multiple X-ray transmission images, and pre-storing these image sets as reference images for corresponding shape groups. A system for recognizing the shape of fruits and vegetables.

Images