A device for rapidly grading the hardness of spherical fruits and its detection method

Abstract

The present invention discloses a rapid grading device for the hardness of spherical fruits and its detection method. It includes a fixed frame, an elastomer layer, and an optical system layer and a graphic acquisition device installed in the fixed frame; the fixed frame is a hexahedron structure, and each surface of the fixed frame is equipped with a light-shielding and closed plate; a through hole is opened in the middle of the light-shielding and closed plate on the top surface of the fixed frame, and the elastomer layer is embedded in the through hole, and a position fixing layer that is arranged circumferentially along the elastomer layer and protrudes is provided on the upper surface of the light-shielding and closed plate on the top surface; the graphic acquisition device, the optical system layer, and the elastomer layer are coaxially arranged from bottom to top, the elastomer layer is located on the upper surface of the optical system layer, and the graphic acquisition device acquires the image information of the elastomer layer through the optical system layer. The device of the present invention has the advantages of simple operation and low cost. After the model is established, a simple pressing process can complete the hardness grade classification, and non-destructive detection of spherical fruits can be realized.

Description

Technical Field

[0001] The present invention belongs to the field of fruit quality detection, and particularly relates to a device for rapidly grading the hardness of spherical fruits and a detection method therefor. Background Art

[0002] China is a major fruit producer and consumer. However, due to the backwardness of technologies related to fruit grading, the value of China's fruit industry is difficult to further increase, and its competitiveness in the international market is not high.

[0003] Hardness is an important indicator for measuring fruit quality and has a great relationship with fruit maturity, playing a certain role in guiding the storage period and shelf life. For some fruits, such as apples and pears, hardness can not only reflect maturity but also the taste of the fruit. Fruits with higher hardness often have a crisper taste. Currently, a considerable part of research has listed hardness as one of the important indicators for evaluating the quality grades of spherical fruits such as apples and pears.

[0004] The hardness distribution range of the highest-grade pears in the national standard is 4 - 9 kg / cm 2 , and the hardness distribution range of the highest-grade apples is greater than 6 kg / cm 2 , and the specific range will fluctuate according to different varieties. Therefore, the hardness grades of apples and pears can be divided according to hardness. Currently, the detection methods for fruit hardness are mainly divided into two categories. One category is destructive, and such hardness detection technologies will damage the fruits; the other category is non-destructive, and such technologies mainly include spectral technologies, mechanical technologies based on dielectric characteristics and acoustic characteristics. However, these technologies face problems such as large interference in the process, high cost, large equipment volume, and strict operation requirements. Summary of the Invention

[0005] In order to solve the problems mentioned in the background art, the present invention proposes a device for rapidly grading the hardness of spherical fruits and a detection method therefor.

[0006] The technical solution adopted by the present invention is as follows:

[0007] I. A device for rapidly grading the hardness of spherical fruits

[0008] It includes a fixed frame, an elastomer layer, and an optical system layer and an image acquisition device installed in the fixed frame; the fixed frame is a hexahedron structure, and each surface of the fixed frame is installed with a light-shielding and sealing plate; a through hole is opened in the middle of the light-shielding and sealing plate on the top surface of the fixed frame, and the elastomer layer is embedded in the through hole. A position fixing layer that protrudes and is arranged circumferentially along the elastomer layer is provided on the upper surface of the light-shielding and sealing plate on the top surface; the image acquisition device, the optical system layer, and the elastomer layer are coaxially arranged from bottom to top. The elastomer layer is located on the upper surface of the optical system layer, and the image acquisition device acquires the image information of the elastomer layer through the optical system layer.

[0009] The elastomer layer includes a transparent flexible layer, an intermediate reflective layer, and a top light-shielding layer from bottom to top. The preparation process is as follows:

[0010] 1.1) Prepare the bottom transparent flexible layer:

[0011] Mix the transparent silica gel and silica gel curing agent evenly and stir well. Place it in a low-temperature (about 24°C) vacuum environment and let it stand for half an hour. Then pour the mixture into a metal mold, heat and cure it, and take it out to obtain a cylindrical bottom transparent flexible layer;

[0012] 1.2) When preparing the intermediate reflective layer,

[0013] Sieve the metal powder (aluminum powder or copper powder) through a sieve three times. Mix the sieved metal powder with silica gel, silica gel curing agent, and silica gel thinner, stir well, and then use an air gun to spray it evenly on the bottom transparent flexible layer, and then heat and cure it to obtain the intermediate reflective layer;

[0014] 1.3) When preparing the top light-shielding layer,

[0015] Mix the black silica gel dye with silica gel, silica gel curing agent, and chloroform, stir well, pour it onto the intermediate reflective layer, transfer it to a spin coater for spin coating, and then heat and cure it to obtain a top light-shielding layer with a circular arc surface on the surface of the intermediate reflective layer.

[0016] In the step 1.1), the thickness of the transparent flexible layer is 2 mm to 5 mm;

[0017] In the step 1.2), sieve the metal powder through a 100-mesh sieve with a mesh diameter of 0.01 micrometers three times;

[0018] In the step 1.3), the thickness of the top light-shielding layer is about 0.4 mm.

[0019] The position fixing layer is an annular columnar structure obtained by digging an elliptical cylinder from a cylinder. The centers of the outer ring circle and the inner ring ellipse are the same; the elastomer layer is located within the area enclosed by the inner ring of the position fixing layer;

[0020] The position fixing layer is used to fix the contact position between the fruit and the elastomer layer. The preparation method is as follows:

[0021] Mix the silica gel and silica gel curing agent evenly and stir well. Place it in a low-temperature (about 24°C) vacuum environment and let it stand for half an hour. Then pour the mixture into a metal mold, heat and cure it, and take it out to obtain the position fixing layer. Use a silica gel adhesive to fix the position fixing layer on the light-shielding closed plate.

[0022] The optical system layer is composed of a transparent light guide plate and LED lights. The cross-section of the transparent light guide plate is square, and a plurality of LED lights are arranged equidistantly around the peripheral surface of the transparent light guide plate. The LED lights on the four circumferential surfaces have different colors, namely white light, yellow light, magenta light, and cyan light.

[0023] The irradiation direction of the LED lights is perpendicular to the edge of the light guide plate and irradiates towards the center of the light guide plate, rather than directly towards the elastomer layer.

[0024] By adjusting the number and light intensity of each LED light around the transparent light guide plate, overexposure or underexposure in the fruit pressing area on the transparent light guide plate is prevented.

[0025] The hardness of the position fixing layer is close to the hardness of the spherical fruit to be measured;

[0026] The hardness of the position fixing layer should be adapted to the hardness of the spherical fruit, that is, the hardness of the fixing layer should be slightly greater than or equal to the lowest hardness of the spherical fruit, and should not be too high to cause damage to the spherical fruit during operation.

[0027] The hardness of the elastomer layer is close to the hardness of the spherical fruit to be measured.

[0028] The hardness of the elastomer layer should be adapted to the hardness of the spherical fruit. That is, in the scenario of fruit objects with lower hardness, an elastomer layer with lower hardness should be selected, and it should not be too hard so that the spherical fruit cannot show its deformation when pressed. In the scenario of fruit objects with higher hardness, an elastomer layer with higher hardness should be selected, and it should not be too soft so that the deformation of the spherical fruit is too large to show the difference when pressed.

[0029] II. Detection method of a spherical fruit hardness rapid grading device

[0030] It includes the following steps:

[0031] Step 1: When pressing the spherical fruit on the position fixing layer, keep the axis of the spherical fruit coincident with the major axis of the inner ring ellipse of the position fixing layer to ensure that the middle part of the spherical fruit can be fully in contact with the top light-shielding layer;

[0032] Define the side of the spherical fruit close to the fruit stalk as the top, and the side far from the fruit stalk as the bottom. The line connecting the top and the bottom is called the axis of the spherical fruit.

[0033] Step 2: During the process of pressing the spherical fruit, use the graphic acquisition device to acquire multiple images of the middle reflective layer under the illumination of the four-color light source at equal time intervals, and perform image processing on the acquired pictures to remove noise;

[0034] Step 3: Collect images of the middle reflective layer when multiple spherical fruits are pressed through Step 1 and Step 2. Then, calibrate the hardness of all spherical fruits using a texture analyzer and divide them into three hardness grades according to national standards, namely, relatively soft, relatively hard, and the optimal grade.

[0035] Step 4: Use all the collected pictures marked with fruit hardness grade labels as a dataset, and divide the dataset into a training set, a validation set, and a test set. Input the training set data into a convolutional neural network model for training, adjust the network parameters through the validation set to optimize the model, and finally evaluate the model performance through the test set to obtain a convolutional neural network model that can classify the fruit hardness after training.

[0036] Step 5: For spherical fruit samples with unknown hardness grades, collect picture data through Step 1 and Step 2 and input it into the model in Step 4 to obtain the hardness grade of the spherical fruit samples.

[0037] Advantages of the present invention:

[0038] The present invention realizes a device for rapid grading of the hardness of spherical fruits and its detection method. The proposed device has the advantages of simple operation and low cost. After the model is established, a simple pressing process can complete the hardness grade classification, and non-destructive detection of spherical fruits can be achieved. Description of the Drawings

[0039] Figure 1 It is a schematic external view of the device for rapid grading of the hardness of spherical fruits

[0040] Figure 2 It is a schematic internal structure view of the device for rapid grading of the hardness of spherical fruits

[0041] Figure 3 It is a schematic view of the elastomer structure

[0042] Figure 4 It is a schematic view of the optical system structure

[0043] Figure 5 It is a picture after data processing

[0044] Figure 6 It is a picture with overexposure in the area caused by an unreasonable light source layout

[0045] In the figure: fixed frame (1), light-shielding and closing plate (2), elastomer layer (3), position-fixing layer (4), optical system layer (5), graphic acquisition device (6), top light-shielding layer (7), middle reflective layer (8), transparent flexible layer (9), LED lamp (10) Detailed Embodiments

[0046] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

[0047] The present invention provides a device for rapidly grading the hardness of spherical fruits and a detection method thereof. The device for rapidly grading the hardness of spherical fruits is as Figure 1 shown in Figure 2 Figure [not shown]. Its main structure includes a fixed frame 1, a light-shielding and sealing plate 2, an elastomer layer 3, a position-fixing layer 4, an optical system layer 5 and an image acquisition device 6 within the fixed frame. By means of 3D printing technology, the fixed frame 1 and the black light-shielding and sealing plate 2 are prepared using resin as the material. The fixed frame and the light-shielding and sealing plate form a closed hexahedron structure to prevent external light from interfering with the acquired images. The centers of the elastomer layer 3, the position-fixing layer 4, the optical system layer 5 and the image acquisition device 6 are coaxial in the vertical direction. Therefore, the image acquisition device 6 can acquire the image information of the elastomer layer 3 through the optical system layer 5. The fixation between various parts is carried out by bolts and nuts.

[0048] As Figure 3 shown, the elastomer layer 3 from bottom to top includes a bottom transparent flexible layer 9, a middle reflective layer 8 and a top light-shielding layer 7, which is prepared through the following steps:

[0049] When preparing the bottom transparent flexible layer 9, silicone and a silicone curing agent are mixed and stirred evenly, and then placed in a vacuum environment at a low temperature of 24 °C for half an hour. The low-temperature environment prevents the silicone from complete curing, and the vacuum environment ensures that the silicone will not be poisoned by air before curing. Then the mixture is poured into a metal mold, taken out after heating and curing, and a cylindrical bottom transparent flexible layer 9 with a thickness of about 2 mm to 5 mm is obtained. If the thickness is too small, the elastomer cannot fully reflect the deformation process of the spherical fruit; if the thickness is too large, it will waste materials.

[0050] When preparing the middle reflective layer 8, copper powder is screened three times through a sieve with a mesh size of 100 meshes and a diameter of 0.01 microns. The screened copper powder is mixed with silicone, a silicone curing agent and a silicone diluent, stirred thoroughly, and then evenly sprayed on the bottom transparent flexible layer 9 using an air gun, and then heated and cured.

[0051] When preparing the top light-shielding layer 7, black silicone dye is mixed with silicone, a silicone curing agent and chloroform, stirred evenly, poured onto the middle reflective layer 8, transferred to a spin coater for spin coating, and then heated and cured. The rotation speed during spin coating should not be too high, otherwise the top light-shielding layer 7 will be too thin, resulting in a decline in the light-shielding effect and being prone to breakage during use. The recommended thickness of the formed top light-shielding layer 7 is about 0.4 mm.

[0052] The preparation process of the position fixing layer 4 is as follows: Step 1: Mix silica gel and silica gel curing agent and stir evenly, then place it in a vacuum environment at 24°C for half an hour. Then pour the mixture into a metal mold, heat and cure it, and take it out to obtain the position fixing layer 4, whose shape is a cylinder, and the cross-section is the superposition of a circle and an ellipse with the same center. The ellipse is the part that is dug out; Step 2: Fix the position fixing layer 4 on the elastomer layer 3 using a silica gel adhesive. Since the middle part of most spherical fruits has the largest change in hardness during the ripening process, the position fixing layer can help the middle part of the spherical fruit to contact the elastomer during the pressing process and help fix the contact position between the fruit and the elastomer, so as to reduce the influence of different light distribution conditions in different regions on the elastomer.

[0053] As Figure 4 shown, the optical system layer is composed of a transparent light guide plate and four-color light sources arranged evenly in a surrounding manner. Among them, the light source colors in the four side directions of the square are different, namely white light, yellow light, magenta light, and cyan light. The light sources come from commercial LED light strips. The number of LED lights and the light intensity on each side are adjusted according to the finally formed image. The basic requirement is that there is no large-area overexposure or underexposure in the finally captured picture. If there is large-area overexposure or underexposure without pressing, then the overexposed or underexposed area will remain unchanged during the fruit pressing process, which means information loss. In addition, it is also necessary to ensure that the final image is easy to extract the fruit contour, so that in the data processing step, the part outside the fruit pressing area can be removed to reduce the interference of useless information. As Figure 5 shown, therefore, there should be no or only a small amount of overexposure or underexposure in the fruit contour area during the pressing process, otherwise as Figure 6 shown. The combination of the light guide plate and the surrounding light sources, with the light source irradiation direction in the plane of the light guide plate and perpendicular to the edge of the light guide plate and irradiating towards the center of the light guide plate, rather than directly towards the elastomer layer, can improve the uniformity of light distribution in the elastomer area.

[0054] The present invention proposes a rapid hardness grading device for spherical fruits and its detection method. The detection method based on the above device is as follows:

[0055] Step 1: Define the side of the spherical fruit close to the fruit stalk as the top and the side far from the fruit stalk as the bottom. The line connecting the top and bottom of the spherical fruit is called the axis. When pressing the spherical fruit on the position fixing layer 4, keep the axis of the spherical fruit coinciding with the long axis of the ellipse to ensure that the middle part of the spherical fruit can fully contact the middle reflective layer 8. During the pressing process of the spherical fruit, use the graphic acquisition device 6 to record five consecutive pictures of the middle reflective layer 8 under the irradiation of the four-color light source at equal time intervals.

[0056] Step 2: Perform image processing on the collected images, remove the environmental parts that do not change during the pressing process, and uniformly set them to black.

[0057] Step 3: Calibrate the hardness of the spheroidal fruits in Step 1 using a texture analyzer, and classify them into three hardness grades according to the hardness.

[0058] Step 4: Use the images in Step 2 and the corresponding hardness grade labels as a data set, and divide it into a training set, a validation set, and a test set. Construct a feature extraction layer through a convolutional neural network, then construct a classifier, feed the data in the training set into the neural network model for training, adjust the network parameters through the validation set to optimize the model, and finally evaluate the model performance through the test set.

[0059] Step 5: For spheroidal fruit samples with unknown hardness grades, collect image data through Step 1 and Step 2, and input it into the model in Step 4 to obtain the hardness grade.

Claims

1. A rapid grading device for the hardness of spherical fruits, characterized in that it includes a fixed frame (1), an elastomer layer (3), and an optical system layer (5) and a graphic acquisition device (6) installed in the fixed frame (1); The fixed frame (1) has a hexahedron structure, and a light-shielding and closed plate (2) is installed on each surface of the fixed frame (1); a through hole is opened in the middle of the light-shielding and closed plate (2) on the top surface of the fixed frame (1), and an elastomer layer (3) is embedded in the through hole. A position fixing layer (4) that is convex and arranged along the circumference of the elastomer layer (3) is provided on the upper surface of the light-shielding and closed plate (2) on the top surface; The graphic acquisition device (6), the optical system layer (5), and the elastomer layer (3) are coaxially arranged from bottom to top. The elastomer layer (3) is located on the upper surface of the optical system layer (5), and the graphic acquisition device (6) acquires the image information of the elastomer layer (3) through the optical system layer (5); The elastomer layer (3) includes a transparent flexible layer (9), an intermediate reflective layer (8), and a top light-shielding layer (7) from bottom to top. The preparation process is as follows: 1.1) Preparation of bottom transparent flexible layer (9): The transparent silicone and the silicone curing agent are mixed and stirred evenly, and placed in a low-temperature vacuum environment and left to stand for half an hour; the mixture is then poured into a metal mold, heated and cured, and then taken out to obtain a cylindrical bottom transparent flexible layer (9); 1.2) preparing an intermediate reflective layer (8), The metal powder is screened three times through a sieve, the screened metal powder is mixed with silica gel, silica gel curing agent and silica gel diluent, and after sufficient stirring, the metal powder is evenly sprayed on the bottom transparent flexible layer (9) using an air gun, and then heated and cured to obtain an intermediate reflective layer (8); 1.3) preparing a top light shielding layer (7), Mix black silica gel dye with silica gel, silica gel curing agent and chloroform, stir thoroughly and evenly, pour onto the intermediate reflective layer (8), transfer to a spin coater for spin coating, and then heat and cure to obtain a top light shielding layer (7) with a circular arc top surface on the surface of the intermediate reflective layer (8); In the step 1.1), the thickness of the transparent flexible layer (9) is 2 mm to 5 mm; In the step 1.2), the metal powder is screened three times through a 100-mesh sieve with a mesh size of 0.01 μm; In the step 1.3), the thickness of the top light shielding layer (7) is 0.4 mm; The position fixing layer (4) is a quasi-annular columnar structure, which is obtained by digging out an elliptical column from a cylinder, and the center of the outer ring circle and the inner ring ellipse are the same; the elastic layer (3) is located in the area surrounded by the inner ring of the position fixing layer (4); The position fixing layer (4) is used to fix the contact position between the fruit and the elastic layer (3), and the preparation method is as follows: Mixing the silica gel and the silica gel curing agent and stirring them evenly, placing them in a low-temperature vacuum environment and letting them stand for half an hour, then pouring the mixture into a metal mold, heating and curing them, taking them out to obtain a position fixing layer (4), and fixing the position fixing layer (4) on the light shielding sealing plate (2) using a silica gel adhesive; The optical system layer (5) is composed of a transparent light guide plate and LED lamps (10); the cross section of the transparent light guide plate is a square; a plurality of LED lamps (10) are arranged around the circumference of the transparent light guide plate at equal intervals; the LED lamps on four circumferential surfaces have different colors, namely white light, yellow light, magenta light, and cyan light; The LED light (10) is irradiated in a direction perpendicular to the edge of the light guide plate and irradiated towards the center of the light guide plate; By adjusting the number and light intensity of each LED light around the transparent light guide plate, the fruit pressing area on the transparent light guide plate is prevented from being overexposed or underexposed; The hardness of the position fixing layer (4) is close to the hardness of the spherical fruit to be tested; The hardness of the elastic layer (3) is close to the hardness of the spherical fruit to be tested.

2. The rapid grading device for the hardness of spherical fruits according to claim 1, characterized in that, The device adopts a rapid hierarchical detection method, which specifically includes the following steps: Step 1: When pressing the spherical fruit onto the position fixing layer (4), the axis of the spherical fruit is kept coincident with the major axis of the inner ellipse of the position fixing layer (4), so as to ensure that the middle of the spherical fruit can fully contact the top light shielding layer (7); Step 2: During the pressing process of the spherical fruit, a plurality of images of the intermediate reflective layer (8) illuminated by four-color light sources are collected at equal time intervals by means of an image collection device (6), and the collected images are processed to remove noise; Step 3: Collect images of the middle reflective layer (8) of multiple spherical fruits when they are pressed through Step 1 and Step 2. Then, calibrate the hardness of all spherical fruits using a texture analyzer and divide them into three hardness grades according to national standards, namely, relatively soft, relatively hard, and the optimal grade; Step 4: Use all the collected pictures marked with fruit hardness grade labels as a data set, and divide the data set into a training set, a validation set, and a test set; input the training set data into a convolutional neural network model for training, adjust the network parameters through the validation set to optimize the model, and finally evaluate the model performance through the test set to obtain a convolutional neural network model that can classify the fruit hardness after training; Step 5: For spherical fruit samples with unknown hardness grades, collect picture data through Step 1 and Step 2 and input it into the model in Step 4 to obtain the hardness grade of the spherical fruit samples.

3. The spherical fruit hardness rapid grading device according to claim 2, characterized in that, In Step 1: Define the side of the spherical fruit close to the fruit stalk as the top, and the side away from the fruit stalk as the bottom. The line connecting the top and the bottom is called the axis of the spherical fruit.

Images