Computer-vision-based automated machine for sorting tahiti limes

Abstract

The invention relates to a computer-vision-based automated machine for sorting Tahiti limes, formed by a hopper (1) connected to a receiving cover (2) for dividing and moving the limes into five lanes (3.1, 3.2, 3.3, 3.4). To select the limes in the corresponding lanes, arms having two degrees of freedom are coupled to the machine, the arms being connected to grippers (11.1, 11.2, 11.3, 11.4). These are moved by signals provided by a diffuse sensor (8), and the movements are executed using servomotors (10.1, 10.2, 10.3, 10.4), which are connected to servomotor supports (9.1, 9.2, 9.3, 9.4). In addition, the entire structure of the arm having two degrees of freedom is connected to a support (7) for the arm having two degrees of freedom. By means of the selecting grippers (11.1, 11.2, 11.3, 11.4), the selected limes fall into a structures (4.1, 4.2, 4.3, 4.4) comprising a ramp for temporarily storing the limes, and the limes subsequently roll down the ramps into a rectangular structure (5.1, 5.2, 5.3, 5.4), according to a sorting order. Lastly, the entire structure of the machine is secured to a fixed structure of square tubes (6).

Description

[0001] Automatic Tahitian Lemon Grading Machine Based on Computer Vision

[0002] TECHNICAL FIELD

[0003] The following invention relates to the agroindustry in general, but is specifically developed in the technical field of automated sorting machines for Tahitian lemons based on computer vision, which perform this action by processing grayscale images. The principles of Industry 4.0 were implemented in the development of this mechanism.

[0004] BACKGROUND OF THE INVENTION

[0005] Currently, devices that present classification procedures are known, such as patent document ES2168499T3, which discloses an apparatus and method for classifying articles. Said invention presents an inspection method for classifying fruits and vegetables, in which the articles are presented on a conveyor belt, under a viewing device. Furthermore, a separation position is determined from the classification; therefore, the article is separated by means of an independently actuated ejector mechanism.

[0006] Also patent document ES2136621 T3 presents the classification of objects or articles of an agricultural product according to predetermined classification criteria, where the objects or individual articles of the agricultural product are rotatably supported between adjacent pairs of a sequence of roller installations above selectively elevated weighing frames and inclined discharge plates. The sequence of roller installations supporting articles of the agricultural product is subsequently displaced beyond a roller rotation surface that can interact with roller installations to rotate the agricultural product supported by it. The agricultural product is scanned to detect the presence of predetermined classification criteria.The agricultural product is conveyed through the weighing station, where it is captured and supported by elevating the weighing frame so that it passes over a load cell, allowing a measurement of the agricultural product's weight to be taken. Finally, the sequence of roller installations with the agricultural product supported between them is advanced past an actuating means. The machine is sensitive to the weighing and / or scanning medium; the actuating means selectively tilts individual discharge plates upward to discharge the corresponding agricultural product between five adjacent pairs of roller installations.

[0007] The first device in the patent (ES2168499T3) presents the problem of not clarifying which inspection mechanisms it will use to classify fruits and vegetables; furthermore, it does not provide the exact specifications of the independently actuated ejector mechanism, given that food quality processes must be maintained when classifying food, and many machine elements can damage the food's protective layer if an appropriate architecture is not established.

[0008] The second invention (ES2136621 T3) presents architectural problems, given that the machine's robustness is not ideally suited to the agricultural products it classifies. This system is highly likely to cause damage to these inputs, resulting in commercial losses. Furthermore, this system is not automated, as its operation is purely mechanical. This means that there is no control system capable of accounting for the 20 inputs per day, which makes it still dependent on operator intervention.

[0009] DESCRIPTION OF THE INVENTION As a solution to the aforementioned problems, the present invention was developed, an automated machine for the classification of Tahitian lemons based on computer vision. The Tahitian lemon was selected as the protagonist for the classification of this machine because it is the product with the highest export income to the Peruvian market. Therefore, images with the characteristics of the Tahitian lemon were added to the control system, which is characterized by being deep green until it ripens and turns pale yellow after that process. For this reason, the machine has five conveyor lanes. In the first selection phase, lemons that are in poor condition will be removed; that is, those that present necrosis and / or greasy spots on the peel in order to prevent them from infecting production.In the second sorting process, lane two will sort lemons with pale yellow peels, as these are overripe and must be separated for immediate sale. In the third sorting process, lane three will sort lemons with deep green peels measuring 4 to 5 cm in diameter. Finally, in the fourth process, lane four will sort lemons with deep green peels measuring 5 cm and 6 cm in diameter.

[0010] At the end of each lane, there is a structure and a ramp, with a total of four. Each of these is responsible for receiving the previously sorted lemons; due to the ramp, the lemons eventually fall into a rectangular storage structure.

[0011] Four two-degree-of-freedom (DOF) robotic arms were used for the sorting system, which will be responsible for selecting and sorting the lemons according to the information established in the control system. These robotic arms will be attached to the rails by a robotic arm support each. Servomotors were incorporated for the movement of the robotic arms, which, for protection and safety reasons, have a support structure. Additionally, the robotic arm's gripper is responsible for moving the lemons for sorting. Finally, the machine is fixed to a structure of 20 square stainless steel tubes.

[0012] BRIEF DESCRIPTION OF THE FIGURES

[0013] Figure 1: Signposted isometric view of each component used in the automated Tahitian lemon grading machine based on computer vision.

[0014] Figure 2: Exploded isometric view of each component used in the automated Tahitian lemon grading machine based on computer vision.

[0015] Figure 3: Isometric view of the Tahitian lemon grading system. Figure 4: Isometric view of the lemon receiving hopper and lid.

[0016] PREFERRED EMBODIMENTS OF THE INVENTION

[0017] An automated machine for the classification of Tahitian lemon based on computer vision comprises a hopper (1) through which the lemons will enter and then pass through the reception cover (2) which will be responsible for distributing the lemons to the five lanes as follows: (3.1) is the lane for lemons in poor condition, (3.2) for ripe lemons that have a pale yellow peel, (3.3) and (3.4) to select lemons in good condition and with characteristics previously mentioned in the description of the invention.

[0018] Each classified lemon will fall onto a structure with a ramp (4.1), (4.2), (4.3) and (4.4) which are responsible for temporarily storing the lemons so that they then fall through the ramp to a rectangular structure (5.1), (5.2), (5.3) and (5.4) according to the classification order mentioned above.

[0019] For the selection of lemons the machine uses two GDL arms that are assembled to selection grippers (11.1), (11.2), (11.3) and (11.4). These receive signals from the diffuser sensor (8) so that they properly select the lemons and distribute them to the corresponding lanes. On the other hand, the grippers (11.1), (11.2), (11.3) and (11.4) are subject to a support for two GDL arms (7) which keeps them fixed to the lanes (3.1), (3.2), (3.3) and (3.4). The movement of the two GDL arms is given by servomotors (10.1), (10.2), (10.3) and (10.4), which for their care and maintenance are assembled to servomotor supports (9.1), (9.2), (9.3) and (9.4). Finally, the machine is supported by a fixed structure of square tubes (6).

Claims

CLAIMS 1. An automated machine for the classification of Tahitian lemon based on computer vision is composed of a hopper (1), which is assembled to a receiving cover (2) that will be dedicated to dividing and moving the lemons to the five lanes (3.1), (3.2), (3.3) and (3.4). To select the lemons in their corresponding lane, arms of two GDL were coupled, which are assembled to grippers (1 1.1), (1 1.2), (1 1.3) and (1 1.4). These are moved by the signals provided by the diffuser sensor (8) and the movements will be executed by servomotors. (10.1 ), (10.2), (10.3) and (10.4), which are assembled to servo motor supports (9.1 ), (9.2), (9.3) and (9.4). Likewise, the entire structure of the two-degree-of-freedom arm is assembled to a support for two DOF arms through the selection grippers (1 1.1 ), (1 1 .2), (1 1 .3) and (1 1 .4) the selected lemons will fall to a structure with ramp (4.1 ), (4.2), (4.3) and (4.4) these are responsible for temporarily storing the lemons and then through the ramp they fall to a rectangular structure (5.1 ), (5.2), (5.3) and (5.4) according to the classification order. Finally, the entire structure of the machine is attached to a fixed structure of square tubes (6).

2. An automated machine for the classification of Tahitian lemon based on computer vision according to claim 1 is characterized by its control system, given that in each support for two GDL arms (7) a counting diffuser sensor is embedded, in total the machine has four sensors.

3. An automated machine for the classification of Tahitian lemon based on computer vision according to claim 1, characterized in that the rails (3.1), (3.2), (3.3) and (3.4) are U-shaped, which allows a less risk of lemons falling off the sides of the tank and wasting raw material.

4. An automated machine for the classification of Tahitian lemons based on computer vision according to claim 1 is characterized in that four robotic arms of two GDL were assembled to the fixed support structure, which are responsible for moving the lemons to each horizontal strip according to their characteristics.

Images