BR112018075438B1 - DEVICE AND METHOD FOR CLASSIFYING GRAINS - Google Patents

Abstract

The present invention relates to a device and method for classifying grains, for example coffee beans. The device is characterized by a grain feeder mechanism; a grain accommodation mechanism connected to the grain feeder mechanism; an electronic grain viewer system operatively disposed in the grain accommodation mechanism; a grain expulsion mechanism connected to the outlet of the grain accommodation mechanism; where the electronic grain viewer system has a central processing unit that implements methods to classify the grains. And the method is characterized by the following steps: (a) obtain the digital image of the grain; (b) store the RGB components of the image obtained in step (a); (c) create a histogram for each calorimetric and luminosity component of the histogram from step (b); (d) determine the limit point according to the Otsu method; (e) obtain a binary image; (f) remove areas; (g) get the edges; (h) obtain the vectors corresponding to the grains; (i) identify black grains; (j) identify grains with fermentation defects and immature grains; (k) identify grains with mechanical damage, and; (l) activate a grain expulsion mechanism by activating actuators for black grains, actuators for (...).

Description

Field of Invention

[0001] The present invention relates to devices for classifying beans, especially washed coffee beans.

Description of the Technique

[0002] The classification of agricultural products results in the provision of a higher quality, which are purchased and acquired at higher prices. The market for specialty coffee beans, those with higher value due to their unique characteristics, is a good example of the above. Although there are coffee processing machines that select excellent quality products, there is a type of coffee bean that cannot be removed through traditional processes, those beans that are slightly affected by the coffee borer (Coffee Berry Borer = CBB) .

[0003] The coffee borer (Hypothenemusampei) is the most aggressive pest on coffee plantations in Colombia, causing enormous economic damage to coffee planters and producers due to the decline in its quality. Considering that coffee beans that have been slightly affected by this pest are difficult to be removed from the main crop during coffee processing and before export, the present invention proposes an alternative to remove coffee beans affected by the coffee borer and beans with deficient endosperm, during the moistening process, with the aim of providing coffee growers with the opportunity to obtain a higher remuneration for producing higher quality coffee and providing a higher quality raw material for exporters.

[0004] There are patents for systems that use computer vision techniques to identify and separate objects that have different characteristics from others. Similarly, there are several examples in the agriculture industry of machines using these techniques to classify agricultural products. All of these technologies have a raw material conditioner, a feeder, an object analysis system, and a removal system that responds to commands sent through the analysis system.

[0005] In the prior art, there are systems in which digital image analysis is used for decision making to determine whether the agricultural product is acceptable or not, which include ejection systems to remove grains of varying quality from the main stream, such as US Patent Nos. US 4,203,522, US 4,513,868, US 4,630,736, US 4,699,273, US 5,538,142, US 6,191,859 and US 5,135,114. However, the method used in the products analyzed in order to determine whether or not they are acceptable varies from product to product.

[0006] Some brands of machines for selecting agricultural products available on the market are: Weco, Cimbria, Sortex, Spectrum, Hongshi, Angelon, Isort, ABD, Really, Aslanjixie, ZRWS, SKS, JietaiZhineng, WB, WY, S-Precision, Xeltron, Delta, among others. These machines handle dry and semi-dry raw materials, different from the machine according to the present invention, which classifies a completely wet product. Coffee in this state has an average humidity of 53% on a wet basis, and has a film of water on its surface.

Summary of the Invention

[0007] The present invention relates to a device and a method for classifying grains. The device contains a grain feeding mechanism, a grain positioning mechanism, which in one embodiment of the invention positions the grains flat side up, an electronic grain vision system and a pneumatic ejection mechanism.

[0008] In one embodiment of the invention, the beans to be classified are washed coffee beans.

[0009] The method of the present invention is a method for classifying grains characterized by the following steps:

[0010] a. obtain a digital image of the grain;

[0011] b. store the RGB components of the image obtained in step (a);

[0012] c. create a histogram for each of the calorimetric and luminosity components of the histogram from step (b);

[0013] d. determine the limit point, according to the Otsu method;

[0014] e. obtain a binary image;

[0015] f. remove areas of the binary image;

[0016] g. determine the limits of the binary image;

[0017] h. obtain the vectors corresponding to the grains;

[0018] i. identify black grains;

[0019]j. identify grains with fermentation defects and immature grains;

[0020] k. identify grains with mechanical damage; It is

[0021] l. activate a grain expulsion mechanism by activating actuators for black grains, actuators for grains with fermentation defects and immature grains and actuators for grains with mechanical damage

[0022] An embodiment of the invention corresponds to a device for classifying coffee beans characterized by:

[0023] (i) a feeding mechanism (see Figures 1 and 5) consisting of a hopper (1) with longitudinal rods that prevent the formation of structures that could block a continuous flow of coffee beans; a rotating drum having a smooth surface (2) that adds energy to continuously feed the washed coffee beans and a scraper (3) (see Figures 2 and 4) of stretched nylon to loosen the beans that may be stuck to the surface of the drum rotary. In one embodiment of the invention, the feeding mechanism provides a continuous flow of washed coffee beans into the system.

[0024] (ii) a grain positioning mechanism (see Figures 6 and 7), which positions the washed coffee beans with their flat surface upwards, consisting of another rotating drum (4) of larger diameter having a surface smooth surface on which the washed coffee beans are fed to a housing (5) having a mixer (6). The grains are positioned with the flat side up (10) or down (9), the latter being maintained on the surfaces (9) through the adhesive and cohesive forces of the water. The drum also has a stretched nylon scraper (7) which loosens grains that may have become stuck by gently discharging them onto a horizontal conveyor belt (8) that moves at a slightly faster speed, with the flat side up. The positioning mechanism that positions the coffee beans flat side up performs this function efficiently.

[0025] (iii) with reference to Figures 3 and 9, an electronic grain vision system consisting of a CCD camera (12), an inspection dome (11), an illumination unit and a central processing unit, the which can be a computer or an embedded (autonomous) mechanism in which analyzes are carried out and decisions are made regarding the quality of the washed coffee beans.

[0026] In one embodiment of the invention, the lighting mechanism of the electronic grain vision system consists of high-intensity LEDs. The central processing unit processes a method to identify grains that are selected from the group consisting of CBB-affected grains, black-black grains, mechanically damaged grains, unripe grains and fermented grains.

[0027] With reference to Figures 8 and 9, in one embodiment of the invention the bean positioning mechanism is a pneumatic ejection mechanism consisting of 10 pneumatic electrovalves that generate jets of air capable of ejecting defective coffee beans from the flow main.

[0028] The device according to the present invention contains a mechanism that feeds the washed coffee beans, a mechanism that positions the coffee beans with the flat side up, an electronic mechanism that identifies defective beans and an ejection mechanism pneumatic.

[0029] Typically, coffee beans have a flat side and a convex side and must pass through the identification mechanism with the flat side up, because this is the side through which the coffee bur penetrates the bean and on which the damage is visible.

[0030] One of the technical features of the present invention consists in taking advantage of the fact that the parchment (endocarp) of the grain is translucent and attached to the seed (endosperm) when the grain is fully moist, which means that defects such as small holes made by coffee borer, or seeds with an unusual color, show greater contrast and apparent area at a glance, at a glance (visible). Once the coffee enters the drying process, the endocarp becomes opaque and stops being translucent, which is why it is very difficult to identify and extract coffee beans that have been slightly affected by the coffee drill.

[0031] These mechanisms are described in detail below.

[0032] Feeding Mechanism:

[0033] The dosage of washed coffee beans is carried out by a rotating cylinder located below the machine's main receiving hopper, whose rotation moves the beans towards one side, thus creating a downward flow. This mechanism has a scraper that releases grains that may be stuck to the surface of the rotating cylinder due to the adhesive and cohesive forces of the water. In one embodiment of the invention, the scraper developed for this purpose consists of a 0.5 mm diameter nylon thread, stretched longitudinally between two lateral structures, which touch the cylinder without placing significant resistance on its movement. In the case of the dosing mechanism, the flow depends on the angular velocity of the cylinder and the distance between the exit flow of the hopper and the cylinder. Greater angular velocity results in greater feed, and greater heights result in greater feed.

[0034] Figures 1 and 2 show in detail the dosing mechanism and its parts, the main hopper (I), the rotating cylinder (2) and the nylon thread scraper (3). The walls of the machine's main hopper are formed with longitudinal rods that prevent the washed coffee beans from adhering to surfaces and forming bridges that can later block the continuous flow of coffee beans.

[0035] Coffee Positioning Mechanism:

[0036] Once the mass of coffee beans is constantly being fed, it passes through a mechanism that positions the beans with the flat side up. As coffee beans contain water on their surface, this coffee positioning mechanism uses the principles of adhesion forces and fluid cohesion in relation to various surfaces. Therefore, when a washed coffee bean is positioned with its flat side on a smooth surface, there is a high probability that it will stick to the surface due to the adhesive and cohesive forces of the water between the surface and the outer layer of the bean. If the washed coffee bean is positioned with its convex face on the same smooth surface, the adhesive and cohesive force of the water would not be strong enough to keep it there, because these forces depend on the contact area.

[0037] With reference to Figures 2, 6 and 7, the bean positioning mechanism, for example, for positioning coffee beans facing upwards, (Figure 2) consists of a rotating drum (4), 40 cm in diameter. diameter, with a smooth surface, which has a tangential speed of 0.35 m/s. The washed coffee flows from the feeding mechanism to a housing (5), similar to a hopper, which comprises a mixing mechanism (6) to break up any grain structure that may block the flow.

[0038] With reference to Figures 2 and 7, the grains that remain adhered continue to rotate with the drum (9) while the grains with their convex surface facing inwards do not stay fixed and fall freely (10 ) when they reach the 180° orthogonal point on the drum from Figure 2.

[0039] In one embodiment of the invention, the drum has a nylon thread scraper, 0.5 mm in diameter (7), positioned at 270° on the cylinder of Figure 2, which breaks the water film on the grains and detaches them. . Just under the orthogonal point of 270° there is a conveyor belt (8), with a linear speed (0.38 m/s) slightly faster than the tangential speed of the drum, with the aim of receiving the grains positioned with the flat side up and transport them.

[0040] With reference to Figures 6 and 7, the distance between the conveyor belt and the cylinder is slightly greater than the thickness of a grain (between 8 and 10 mm).

[0041] The positioning mechanism that positions the coffee beans with the flat side up performs this function highly efficiently. Coffee beans that do not settle and fall freely at the 180° point of the cylinder are mainly coffee beans that do not have a planar - convex shape, such as triangle or pea beans. The small amount of acceptable grain that falls at this point is reprocessed because a lifting bucket returns it to the main hopper.

[0042] The mechanism has a variety of flexible attachments that organize and guide the grains into rows in such a way that the grains are not stacked on top of each other.

[0043] Electronic Grain Vision System:

[0044] With reference to Figures 3 and 4, once the grains are positioned on the conveyor belt in an organized manner and with the flat side facing upwards, they pass under an inspection dome (11) comprising a mechanism for lighting with high intensity LEDs and a CCD digital camera (12) as shown here in Figure 3. A central processing unit analyzes the images captured through the camera with the aim of determining which grains are acceptable, which are defective and where they are located on the conveyor belt, with the aim of ejecting them later.

[0045] The central processing unit has a series of methods that segment and classify grains into categories, for example: acceptable grains, grains affected by CBB, unripe grains, fermented grains, black grains and damaged grains mechanics.

[0046] For grain segmentation, a blue contrasting backdrop is used. A calibration is required before machine operation, during which the values to be extracted from the background are determined via dynamic limit setting.

[0047] Method 1 - Description of Steps

[0048] 1. Obtain digital image

[0049] 2. Store RGB components

[0050] 3. Represent the image in other color spaces

[0051] 4. Create a histogram for each of the colorimetric and brightness components

[0052] 5. Determine the limit point according to the Otsu Method (1979) of the selected component.

[0053] 6. Get binary image

[0054] 7. Remove small areas

[0055] 8. Obtain the threshold using, for example, the LoG filter (Computer and Machine Vision, Davies, 2012; Digital Image Processing, Gonzales and Woods, 2008)

[0056] 9. Obtain the vectors that correspond to the grains

[0057] To obtain the class “black-black grains”, the following method is used:

[0058] Method 2 - Description of Steps

[0059] 1. Average of the RGB components for each of the image pixels, that is, calculate the grayscale image.

[0060] 2. Extract the average intensity for the information of each of the segmented grains.

[0061] 3. If the average intensity is less than 80, the corresponding grain is black/black.

[0062] 4. Engage the grain ejection mechanism.

[0063] For example, to identify an "unripe bean" and a "fermented bean", a scatterplot of green and blue average intensity information on the GB Cartesian plane (green-blue of the RGB color space) which is used for a central segment that includes the divided part of each of the grains. Figure 10 shows the scatterplot. A straight line with a slope of 0.71 separates healthy grains from defective grains and fermented grains. The method developed for this purpose is as follows:

[0064] Method 3. Identification of defective, fermented and unripe grains

[0065] 1. Adjust the segmented grain vector to an ellipse and discover/find the minor axis and the major axis.

[0066] 2. Use the long axis as the base and extract a longitudinal strip that is 1/3 the width of the short axis

[0067] 3. Obtain the average G and B intensity of the strip from each of the grains

[0068] 4. If the B/G slope is less than 0.71, the corresponding grain is defective due to fermentation or lacks ripening.

[0069] 5. Engage the grain ejection mechanism.

[0070] Since the coffee borer beetle drills a hole with an average diameter of 1.23 mm, the following algorithm is used to identify beans that have been damaged by insects:

[0071] Method 4. Identification of affected beans using the coffee borer beetle

[0072] 1. Take the matrix of each of the grains

[0073] 2. Move a 5x5 pixel mask in an orderly manner

[0074] 3. If the mask matches an area with a gray intensity lower than 50 (dark) it is considered to be a washed coffee bean affected by an insect.

[0075] 4. Engage the grain ejection mechanism

[0076] Grains with mechanical damage usually have an incomplete shape. To determine which grains have these defects, an ellipse is fitted and the integrity of the shape is determined as shown here in the following algorithm:

[0077] Method 5. Identification of grains with mechanical damage

[0078] 1. Adjust the segmented grain vector to an ellipse.

[0079] 2. Determine the area of the ellipse

[0080] 3. Determine the grain area

[0081] 4. Obtain the area ratio

[0082] 5. If the area ratio is less than 80%, the corresponding grain is considered to have mechanical damage.

[0083] 6. Engage the grain ejection mechanism

[0084] The methods have an identification efficiency of 90%.

[0085] Grain ejection mechanism:

[0086] With reference to Figures 8 and 9, as a result of the aforementioned process herein, the central processing unit sends digital activation signals (on/off) to the ejector solenoid valves. The air currents produced eject the defective grains from the main flow

[0087] In one embodiment of the invention, washed coffee beans are classified into five types of defective beans: CBB-affected coffee beans, black-black coffee beans, unripe coffee beans, fermented coffee beans, and mechanically damaged coffee cups. It should be understood that the present invention is not limited to the embodiments described and illustrated herein, as it would be obvious to a person skilled in the art that there are possible variations and modifications which do not deviate from the spirit of the invention, which are only defined. for the claims here below.

Claims (8)

1. A device for classifying grain comprising: a grain feeding mechanism; and CHARACTERIZED by the fact that it comprises a bean arrangement mechanism connected to the bean feeding mechanism, which arranges the washed coffee beans with the flat side facing upwards; an electronic bean vision system, placed in operating mode within an inspection dome oriented towards the flat side of the washed coffee beans; a grain ejection mechanism, located after the inspection dome (11); wherein the grain arrangement mechanism comprises a rotating drum (4) with a smooth surface, and the flat side of the grains is fixed to the smooth surface of the drum (4) and after a rotation of 180° a scraper (7) detaches the grains with the flat side up, and wherein the electronic grain vision system has a central processing unit that implements grain sorting methods. 2. Device for classifying grains, according to claim 1, CHARACTERIZED by the fact that the grain feeding mechanism comprises a rotating cylinder (2). 3. Device for classifying grains according to claim 1, CHARACTERIZED by the fact that the grain arrangement mechanism comprises a container (5) with a stirring mechanism (6). 4. Device for classifying beans, according to claim 1, CHARACTERIZED by the fact that the grain arrangement mechanism organizes the washed coffee beans with the flat side up on a conveyor belt (8), in which the vision system grain electronics is located above the conveyor belt (8) and wherein the ejection mechanism comprises at least one ejector comprising a pneumatic electrovalve. 5. Device for classifying grains, according to claim 1, CHARACTERIZED by the fact that the central processing unit is configured to: identify black grains by calculating an average intensity of an image of each grain on a gray scale and comparing the average with a threshold value; identify grains with fermentation defects or green grains, calculating the average green and blue color intensities on a GB Cartesian plane and comparing the average with a limit value; identify grains with mechanical damage by calculating a grain area and an ellipse area adjusted to the grain; and activating the grain ejection mechanism, activating one or more ejectors to eject identified black grains, grains with fermentation defects, green grains and grains with mechanical damage from the main flow. 6. Method for classifying beans, CHARACTERIZED by the fact that it comprises the steps of: positioning washed coffee beans with the flat side up in the device, as claimed in claim 1; capture a digital image of the grains; store RGB components of the image, creating a first histogram; create a second histogram for each colorimetric and brightness component of the first histogram; determine the threshold point according to the Otsu method; obtain a binary image; remove small areas of the binary image; get the edges of the binary image; obtain the vectors that correspond to the grains; identify black grains by averaging an image intensity for each grain on a gray scale and comparing the obtained value with a first threshold value; identify defective fermented and green beans by calculating an average RBG of green and blue intensity from the RGB components and comparing the average RBG with a second threshold value; identify beans affected by the coffee borer by calculating a matrix from an image of beans and moving a 5 x 5 pixel mask over the matrix; identify grains with mechanical damage by calculating an area of the grain and an area of an ellipse fitted to the grain; and activating a bean ejection mechanism by activating triggers for black beans, for fermented or green defective beans and beans affected by the coffee borer and beans with mechanical damage. 7. Method according to claim 6, CHARACTERIZED by the fact that the stage of identifying defective fermented and green grains also comprises: from the image obtained by adjusting a vector of a segmented grain to an ellipse and finding the minor and major axes; using the minor and major axes and extracting a longitudinal strip of 1/3 of the width of the minor axis; obtain the average intensities G and B of the strip for each grain; calculate a B/G slope; when the B/G inclination is less than 0.71, activate the grain ejection mechanism. 8. Method according to claim 6, CHARACTERIZED by the fact that the step of identifying grains with mechanical damage also comprises: from the image obtained, adjusting a vector of a segmented grain to an ellipse; calculate an area of the ellipse; calculate a grain area; obtain the ratio between the area of the ellipse and the area of the grain; When the ratio between the ellipse area and the grain area is less than 80%, activate the grain ejection mechanism.