Food portioning device

Abstract

A food portioning device 1 includes: a gripping part 20 that inserts a gripper 210, which has a pair of finger parts 211 arranged facing each other and capable of opening and closing, into a lump of a food 42 placed in a predetermined place, grips and lifts the food 42, and moves the food 42 gripped by the gripper 210 to a predetermined position; an input means 11 for inputting a target gripping amount; an imaging means 30 for imaging a lump of the food to be served; an identification means 12 for identifying a surface shape of the lump of the food on the basis of the image captured by the imaging means 30; and a determination means 13 for determining an insertion position for the finger parts 211 on the basis of the target gripping amount and the identified surface shape.

Description

Food serving device

[0001] The present invention relates to a food serving device that automatically serves food into containers.

[0002] In a cooking system, the task of portioning food into containers is a time-consuming task. Non-Patent Document 1 discloses a robot that automatically performs this portioning task. The robot disclosed in Non-Patent Document 1 grasps food with a grip having a pair of fingers and portions the food into containers by moving the grip.

[0003] Connected Robotics Inc., DELIBOT™, [online], [Retrieved May 11, 2023], Internet <URL: https: / / connected-robotics.com / products / delibot / >

[0004] In food serving tasks, it is necessary to serve a uniform amount of food each time. However, in general, the surface shape of the food in the area where the food is placed may be uneven due to irregularities, etc., and in this case, there is a problem that the amount of food served varies depending on where the food is grasped.

[0005] The present invention has been made in view of the above circumstances, and has as its object to accurately grasp a desired amount of food when serving placed food.

[0006] One aspect of the present invention is a food serving device that includes a gripping unit that inserts a gripper having a pair of opposing finger sections that can be opened and closed into a block of food placed in a predetermined location and then lifts it up, moving the food gripped by the gripper to a predetermined position; an input means for inputting a target gripping amount; an imaging means for photographing the block of food to be served; an identification means for identifying the surface shape of the block of food based on the image captured by the imaging means; and a determination means for determining the insertion position of the finger sections based on the target gripping amount and the identified surface shape.

[0007] According to this food serving device, even if the surface shape of the food varies in the area where the food is placed, the amount of food served can be made uniform.

[0008] In a preferred embodiment, the gripping portion has a first finger portion and a second finger portion arranged parallel to each other, and the first finger portion and the second finger portion are selectively operated according to the target gripping amount and the surface shape of the food block.

[0009] According to this aspect, the first finger portion and the second finger portion are selectively operated according to the target gripping amount and the surface shape of the food block, so that the amount of food dispensed can be made uniform even in situations where the target gripping amount and the surface shape of the food block change.

[0010] In this aspect, the first finger portion and the second finger portion may have different shapes.

[0011] According to this aspect, when the food to be served changes, the finger portion can be quickly switched to one suitable for the new food.

[0012] In this aspect, the first finger portion and the second finger portion may have different maximum distances between the pair of fingers.

[0013] According to this aspect, the finger spacing of the finger portion having the longer maximum spacing between the pair of fingers is fixed to the maximum spacing, making it easier to grip food with the other finger portion.

[0014] In another aspect, the present invention provides a program for causing a computer to execute the steps of inserting a gripper having a pair of opposing, openable and closable finger portions into a block of food placed in a predetermined location, then lifting it up and moving the food gripped by the gripper to a predetermined position, inputting a target gripping amount, photographing the block of food to be portioned, identifying the surface shape of the block of food based on the photographed image, and determining the insertion position of the finger portions based on the target gripping amount and the identified surface shape.

[0015] FIG. 1 is a front view showing the configuration of a food portioning device according to a first embodiment of the present invention; FIG. 2 is a diagram showing an example of operation of the same embodiment; FIG. 3 is a diagram for explaining a search for a potential insertion area according to the same embodiment; FIG. 4 is a diagram for explaining the relationship between surface shape and accuracy of gripping amount; FIG. 5 is a diagram for explaining estimation of gripping amount according to the same embodiment; FIG. 6 is a perspective view showing an example of the configuration of a gripping unit according to a second embodiment of the present invention; FIG. 7 is a perspective view showing an example of the configuration of a gripping unit according to the same embodiment; FIG. 8 is a diagram showing an example of the configuration of a gripping unit according to the same embodiment; FIG. 9 is a diagram showing an example of operation of the same embodiment;

[0016] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0017] 1 is a front view showing the configuration of a food serving apparatus 1 according to a first embodiment of the present invention. The food serving apparatus 1 serves food 42 placed in a food placement area 41 into containers 52 transported by a conveyor 51.

[0018] As shown in FIG. 1, the food serving device 1 has a gripping portion 20, a photographing means 30 consisting of a camera, and a controller 10.

[0019] The gripping unit 20 is a means for inserting a gripper 210 having a pair of opposing openable and closable fingers 211 into a block of food 42 placed in the food placement area 41, then lifting it up and moving the food 42 gripped by the gripper 210 to the position of the container 52. The photographing means 30 is a means for photographing the block of food 42 to be dispensed.

[0020] In a preferred embodiment, the controller 10 includes a memory and a processor that executes a program stored in the memory. The processor functions as the input unit 11, the specifying unit 12, and the determining unit 13 by executing the program stored in the memory.

[0021] Here, input means 11 is a means for inputting a target gripping amount, which is the weight of food to be gripped by gripping unit 20. This input means 11 may receive input of the target gripping amount via an operation unit externally attached to controller 10, or may receive input of the target gripping amount from a higher-level device via a wired or wireless communication channel.

[0022] The identification means 12 is a means for identifying the surface shape of the block of food 42 based on the image captured by the imaging means 30. Here, the surface shape of the block of food 42 refers to the surface shape of the block of food 42 that is a candidate for portioning (information about the unevenness, the shape and height / depth of peaks (convex portions) and valleys (concave portions), and other three-dimensional shape). Any well-known image analysis method can be used to identify the surface shape.

[0023] The determination means 13 determines control parameters, including the insertion position of the finger portion 211, based on the target gripping amount and the surface shape identified by the identification means 12. Here, determining the insertion position based on the surface shape means searching for the most ideal "protrusion" for separating the food 42 within the surface area of ​​the food 42. In a preferred embodiment, the determination means 13 selects multiple areas of a predetermined size surrounding the current insertion position as candidate insertion positions, scores (quantifies) each area for its suitability for separating the food 42, and determines the area with the highest score as the insertion position. Alternatively, instead of searching all areas and determining the area with the highest score, the area may be determined as the insertion position upon finding an area with a score above a preset threshold. This allows for efficient determination of a reasonably suitable insertion position. When determining the insertion position, for example, for a certain food, feature quantities of the surface shape of the insertion area, such as the uneven shape determined based on image data, and the error between the target grip amount and the actual insertion into that area may be input, and a learning model may be constructed that indicates the appropriateness of performing insertion for that area when image data showing a certain area is input, and the appropriateness of the position may be determined by inputting the input image data into this learning model.In a preferred aspect, the determination means 13 may first determine (narrow down) multiple insertion positions or position ranges (areas) for the finger unit 211 based on the target grip amount and surface shape, determine one or more control parameters related to the operation of the grip unit 20 at each of the determined positions or position ranges, and determine the combination of positions that is most suitable for gripping from multiple combinations of the positions and control parameters.Here, the control parameters may include at least one of the following information: the detailed position in the horizontal plane of insertion within the determined area; the opening and closing width of the finger pair (the width (maximum width) when performing insertion and the width (minimum width) when pulling up the finger portion 211); the insertion depth (the Z-direction position of the finger portion 211 when starting to close the finger portion 211); the speed at which the finger portion is inserted; the speed at which the inserted finger portion 211 is pulled out of the block of food; and the speed at which the finger pair is opened or closed.

[0024] An example of operation is shown in Figure 2. First, the user sets a target gripping amount via the input unit 11 (S101). Next, the identification unit 12 identifies the surface shape of the placed block of food from an image obtained by capturing the image, and determines multiple candidate areas for inserting the finger units 211 based on the identified surface shape and the target gripping amount (S102). Next, the determination unit 13 calculates an estimated gripping amount, which is the weight of the food expected to be gripped when inserting the food under multiple conditions related to control parameters for operating the finger units 211, for each of the multiple extracted candidate insertion areas (S103). Next, the determination unit 13 compares the calculated multiple estimated gripping amounts with the target gripping amount and determines the one condition that provides the smallest difference (S104). Finally, the controller 10 determines the control parameters corresponding to the determined conditions, and controls the finger units 211 based on the control parameters to start the gripping operation (S105).

[0025] A more detailed explanation follows. Assume that the target grip weight is set to 30 g (grams) in S101. In S102, the surface shape is determined based on the captured image, and multiple convex portions (candidate insertion areas) of an appropriate size for the target grip weight are extracted according to a predetermined algorithm (S102). FIG. 2A is an example of a side view of the placed food and finger portion 211. As shown in the figure, the placed block of food is generally not uniform in the height direction (Z direction) (in other words, the surface is not flat), and the surface has a complex shape with unevenness (ridges and valleys).

[0026] In the example shown in FIG. 2A , the surface area can be broadly divided into a nearly horizontal area 41a, convex (mountain) areas 41c, 41e, and 41f, a sloped area 41b, and a valley (concave) area 41d. Generally, the order of areas that can be grasped stably (i.e., with less variation in accuracy) is convex > horizontal > sloped. This is explained using FIG. 2B . As shown in FIG. 2A , in a substantially flat surface, food in area B1 between the finger pairs is almost certain to be grasped, whereas the extent to which food will be grasped in the surrounding areas C1, C2, and C3 is uncertain. As shown in FIG. 2B , in a convex area, the ratio of the volume of area C4 where grasping is uncertain to the volume of area B2 where grasping is certain is small. From this perspective, it is preferable that the above algorithm, for example, first divides the area into convex parts > horizontal parts > inclined parts, and then for each divided area, defines an index indicating the suitability as a gripping area from the perspective of the area's attributes, with the index increasing in the order of convex parts, horizontal parts, and inclined parts.

[0027] Next, for each of the divided areas, the index is corrected based on a predetermined algorithm, taking into account the target gripping amount. Such correction is necessary because the surface shape is related to the volume of the food portion to be gripped, and the impact of the degree of surface irregularity on the accuracy of the gripping amount varies in relation to the target gripping amount. For example, when the target gripping amount is extremely small, the surface shape does not significantly affect the accuracy of the gripping amount. However, when the target gripping amount is large, differences in surface shape may significantly affect the accuracy of the gripping amount. Therefore, even for the same convex portion, the optimal curvature (sharpness) of the convex portion may differ in relation to the target gripping amount. Taking this into consideration, the determination unit 13 may, for example, perform machine learning or simulation in advance to determine the relationship between the accuracy of the target gripping amount and the three-dimensional shape pattern, including the curvature of the convex portion, and store the optimal parameters calculated from the learning model obtained by machine learning or the simulation results, and use these as the algorithm.

[0028] The determination means 13 compares the finally calculated index with a predetermined threshold and extracts area candidates whose difference is less than a predetermined value. Note that it is preferable to adjust the threshold so that multiple area candidates are extracted (in other words, so as not to narrow down the area to one in this step). This is because the error is not only related to the area, but also to other conditions. Here, in Figure 2A, it is assumed that two areas, area 41c and area 41f, have been extracted as insertion area candidates.

[0029] 2 , for each of the extracted candidate areas, the control parameters for the adjustable gripping operation within that area are changed using a predetermined algorithm, and an estimated gripping amount, which is the weight of the food expected to be gripped when the insertion operation is performed, is calculated (S103). Here, the control parameters may include, for example, the detailed insertion position within the horizontal plane within the area (XY coordinates; information on where in the convex region the finger portion 211 should be set, such as whether the center of the finger portion 211 should be at the highest point of the mountain or a position slightly offset from the highest point), the opening and closing width of the finger (the width (maximum width) when inserting the finger portion 211 and the width (minimum width) when lifting the finger portion 211), and the insertion depth (the position of the finger in the Z direction when starting the operation of closing the finger portion 211).

[0030] In the example shown in Figure 2C, the estimated gripping amounts calculated for areas 41c and 41f corresponding to conditions relating to four different detailed positions in the horizontal plane as examples of control parameters are {25g (condition 1), 36g (condition 2), 37g (condition 3), 24g (condition 4)} (area 41c) and {23g (condition 1), 58g (condition 2), 32g (condition 3), 38g (condition 4)} (area 41f), respectively.

[0031] 2, the condition under which the difference between the calculated estimated grip amounts and the target grip amount is smallest is determined (S104). In the example of FIG. 2B, the estimated grip amount of 32 g is determined as the condition under which the difference from the target grip amount is smallest (condition 3).

[0032] 2, the controller 10 then sets a control parameter corresponding to the determined condition (condition 3) and operates the finger unit 211 in accordance with the control parameter. As a result, the gripping unit 20 moves to the detailed position of the convex portion of area 41f corresponding to condition 3, and inserts the finger unit there to grip the food (S105).

[0033] As described above, according to this embodiment, the insertion position of the finger portion 211 is determined based on the target gripping amount and the surface shape, so that the food 42 can be gripped with high precision relative to the set target gripping amount. In other words, when gripping operations are performed multiple times, variation in the gripping amount can be reduced.

[0034] 3 to 5 are perspective views showing examples of the configuration of a gripping portion in a food serving device according to a second embodiment of the present invention. Gripper 20a shown in Fig. 3, gripper 20b shown in Fig. 4, and gripper 20c shown in Fig. 5 all have first and second finger portions arranged parallel to each other. The food serving device in this embodiment serves food by selectively operating the first and second finger portions.

[0035] 3 has two rails 201a and 201b arranged parallel to each other fixed to the underside of a support 21 supported by a robot arm (not shown). In the gripper 20a, a first finger portion 211a and a second finger portion 211b each consisting of a pair of opposing fingers are supported by the rails 201a and 201b, respectively.

[0036] In the gripper 20a, the fingers constituting the first finger portion 211a and the second finger portion 211b are fork-shaped and suitable for gripping foods such as stir-fried green peppers and pork. Meanwhile, in the gripper 20b shown in FIG. 4, the fingers constituting the first finger portion 211c and the second finger portion 211d are scoop-shaped and suitable for gripping foods such as sweet and sour pork. In the gripper 20c shown in FIG. 5, the fingers constituting the first finger portion 211e and the second finger portion 211f are elongated, curved, rod-shaped and suitable for gripping bulk foods such as fried chicken. The only difference between the grippers 20a, 20b, and 20c is the shape of the first finger portion and the second finger portion. Therefore, the detailed configuration of only the gripper 20a shown in FIG. 3 will be described below.

[0037] In FIG. 3 , a pair of fingers constituting the first finger portion 211a are aligned below the rail 201a and face each other in the longitudinal direction of the rail 201a. Each finger constituting the first finger portion 211a is bent at a right angle near its upper end, and this bent portion is fixed to a rectangular plate-shaped sliding portion 202a by a screw 203a. A protrusion (not shown) is formed on the upper surface of the sliding portion 202a. This protrusion is received in a rail groove (not shown) formed on the underside of the rail 201a. Therefore, each finger constituting the first finger portion 211a can slide in the longitudinal direction of the rail 201a. Note that a detachable fastening member may be used instead of the screw 203a. This allows for easy replacement of fingers with various shapes depending on the purpose and application.

[0038] Similarly, a pair of fingers constituting the second finger portion 211b are aligned below the rail 201b in the longitudinal direction of the rail 201b and face each other. Each finger constituting the second finger portion 211b is bent at a right angle near its upper end, and this bent portion is fixed to a rectangular plate-shaped sliding portion 202b by a screw 203b. A protrusion (not shown) is formed on the upper surface of the sliding portion 202b, and this protrusion is received in a rail groove (not shown) formed on the underside of the rail 201b. Therefore, each finger constituting the second finger portion 211b can slide in the longitudinal direction of the rail 201b.

[0039] In this embodiment, the food serving device includes a means for independently driving each of the fingers 211 constituting the first finger portion 211a and each of the fingers 211 constituting the second finger portion 211b in the longitudinal direction of the rail 201a or 201b. Therefore, the food serving device according to this embodiment selectively operates the first finger portion 211a and the second finger portion 211b depending on the target gripping amount and the surface shape of the block of food. Here, selective operation refers to, for example, switching between using both the first finger portion 211a and the second finger portion 211b to grip food or using only one of the first finger portion 211a and the second finger portion 211b to grip food.

[0040] Figure 6 shows an example of operation of this embodiment. In Figure 6(a), the target gripping amount is large, and the surface shape of the block of food 42 at the insertion position of first finger portion 211a and second finger portion 211b is also suitable for gripping a large amount of food. Therefore, the food serving device grips the block of food 42 using both first finger portion 211a and second finger portion 211b. In contrast, in Figure 6(b), the target gripping amount is small, and there is no need to use both first finger portion 211a and second finger portion 211b. Therefore, the food serving device grips the block of food 42 using only second finger portion 211b.

[0041] If the target gripping amount is large but the surface shape of the block of food 42 at the insertion position of the first finger portion 211a and the second finger portion 211b is not suitable for gripping a large amount of food, the following two solutions can be considered.

[0042] The first solution is to change the candidate insertion positions for the first finger portion 211a and the second finger portion 211b, and use both the first finger portion 211a and the second finger portion 211b to grasp the target amount of food.

[0043] The second solution is to change the target gripping amount to half, and perform the action of gripping the food 42 twice using only the second finger portion 211b.

[0044] According to this embodiment, the first finger portion 211a and the second finger portion 211b are selectively operated according to the target gripping amount and the surface shape of the food block, so that food can be efficiently portioned, particularly when the target gripping amount is large.

[0045] Although the first and second embodiments of the present invention have been described above, other embodiments of the present invention are also possible. For example, the following embodiments are possible.

[0046] (1) In the second embodiment, the first finger portion and the second finger portion may have different shapes. For example, in the second embodiment, the finger portion 211 of the first finger portion may be fork-shaped, suitable for holding things like stir-fried green peppers and pork, and the finger portion 211 of the second finger portion may be scoop-shaped, suitable for holding things like sweet and sour pork. According to this aspect, when the food to be served changes, it is possible to quickly switch to finger portions suitable for the new food.

[0047] (2) In the second embodiment, the maximum distance between the paired fingers of the first and second finger sections may be different. For example, the maximum distance between the fingers of the second finger section may be longer than the maximum distance between the fingers of the first finger section. Furthermore, the target gripping amount may be small, making it possible to grip food with only one finger section. In this case, the finger distance of the second finger section is fixed at the maximum distance, and the food is gripped by the first finger section. Since the finger distance of the second finger section is widened to the maximum distance, it does not interfere with the gripping action of the first finger section.

[0048] (3) In the above embodiments, the area into which the finger portion 211 is inserted and the control parameters related to the gripping operation are determined, but the determination of the control parameters related to the gripping operation may be omitted. In short, it is sufficient if at least the position or range (area) of the position into which the finger portion 211 is inserted can be determined based on the surface shape and the target gripping amount.

[0049] (4) In each of the above embodiments, the type, ingredients, cooking method, and other physical and chemical properties of the food to be processed are not particularly limited. Generally, any food may be processed as long as the surface of the block of food is uneven (the surface condition is not uniform) when placed on the food. In short, the method according to the present invention may include the steps of inserting a gripper having a pair of opposing, openable and closable fingers into a block of food placed in a predetermined location, then lifting it up and moving the food gripped by the gripper to a predetermined position, inputting a target gripping amount, photographing the block of food to be dispensed, identifying the surface shape of the block of food based on the image captured by the photographing means, and determining the insertion position of the fingers based on the target gripping amount and the identified surface shape.

[0050] 1...food portioning device, 10...controller, 11...input means, 12...identification means, 13...determination means, 30...photography means, 20, 20a, 20b, 20c...gripping portion, 210...gripper, 211...finger portion, 211a, 211c, 211e...first finger portion, 211b, 211d, 211f...second finger portion, 41...food placement area, 42...food, 51...conveyor, 52...container, 41a to 41e...area, 21...support portion, 201a, 201b...rail, 202a, 202b...sliding portion, 203a, 203b...screws.

Claims

1. A food serving device having a gripping unit which inserts a gripper having a pair of opposing openable and closable finger parts into a block of food placed in a designated location and then lifts it up, moving the food gripped by the gripper to a designated position; input means for inputting a target gripping amount; photographing means for photographing the block of food to be served; identification means for identifying the surface shape of the block of food based on an image taken by the photographing means; and determination means for determining the insertion position of the finger parts based on the target gripping amount and the identified surface shape.

2. The food serving device of claim 1, wherein the gripping portion has a first finger portion and a second finger portion arranged parallel to each other, and the first finger portion and the second finger portion are selectively operated according to the target gripping amount and the surface shape of the food block.

3. The food serving device according to claim 2, wherein the first finger portion and the second finger portion have different shapes.

4. The food serving device according to claim 2, wherein the first finger portion and the second finger portion have different maximum distances between the pair of fingers.

5. A program for causing a computer to execute the steps of: inserting a gripper having a pair of opposing openable and closable finger portions into a block of food placed in a predetermined location, then lifting it up and moving the food gripped by the gripper to a predetermined position; inputting a target gripping amount; photographing the block of food to be separated; identifying the surface shape of the block of food based on the photographed image; and determining the insertion position of the finger portions based on the target gripping amount and the identified surface shape.

Images