A system and a method for slicing food objects
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-02-05
- Publication Date
- 2026-08-13
Smart Images
Figure EP2026053027_13082026_PF_FP_ABST
Abstract
Description
[0001] A SYSTEM AND A METHOD FOR SLICING FOOD OBJECTS
[0002] INTRODUCTION
[0003] The present invention relates to slicing of food objects, particularly slicing of fish, poultry, meat, fruit, or vegetables. Particularly, the invention relates to slicing fish such as fish fillets or loin of fish, such as slicing raw, smoked, salted, or graved fish.
[0004] BACKGROUND
[0005] Even though the present invention relates to slicing of food objects in general, the invention is particularly suitable for slicing fish, e.g. for slicing thin slices of fish.
[0006] Fillets of salmon, trout, and other species are often salted, smoked, or graved and distributed in sliced form. Slicing fish is a complicated art requiring either skilled workers or advanced machines. Particularly slicing of non-frozen food objects may be complicated.
[0007] In a similar manner, meat from poultry, cattle, and pork etc. is often distributed in portions, e.g. sliced in thin slices. This may apply both to raw meat and prepared meat such as sausage, smoked, dried, or salted ham, and other kinds of meat.
[0008] Automated slicing equipment is suitable for fast and identical processing of large amounts of food objects. For some food products, slicing is based on a pattern specifying the angle of the path of the knife relative to the food objects and sometimes that angle is specifically adapted to a particular size or shape of a specific food object. Such features may complicate handling of the food objects since the machine needs to be notified about transitions from one food object to a subsequent food object. This complicates the handling of the food objects and potentially reduces the processing speed.
[0009] SUMMARY OF THE INVENTION
[0010] It is an object of embodiments of the present invention to provide an improved slicing system which potentially can increase capacity, and which facilitates a more flexible loading of the slicing system with the food objects into the system. It is a further object to provide a simple and robust system.For these and other objects, the present invention, in a first aspect, provides a system for slicing food objects and in a second aspect provides a method of slicing food objects according to the independent claims.
[0011] It is often desired to apply slicing parameters which depend on a specific shape or size of the food object in question. Particularly, the slicing angle may depend on the height of the food object, defined herein as the distance from the conveyor belt to the upper surface facing away from the conveyor belt. This is done such that the slicing pattern counteracts differences in height along a food object, and the length of the slices becomes more uniform.
[0012] A fish fillet, taken as an example, typically has a fillet height which either decreases from the head end towards the tail end, or which increases from a head end of the fish fillet towards a centre of the fish fillet and decreases from the centre of the fish fillet towards a tail end of the fish fillet. In such a case, the slicing parameters may, correspondingly, change the angle from a low angle (e.g. in the head end) towards a higher angle (in centre part) and back to a low angle (e.g. in the tail end) relative to the fish fillet such that the length of each slice becomes more uniform. In this way, the slicing parameters counteract the variation in thickness.
[0013] The CPU of the present system is programmed to define an array of strokes thereby defining cutting lines for each food object. In this way, each food object can be sliced individually, and particularly with an individual pattern of cutting angles along the food object.
[0014] The slicing parameters may comprise, or consist of, a slicing angle, i.e. an angle of the path of the knife relative to the conveyor belt surface. The strokes may particularly comprise strokes with same or different angles. Accordingly, each food object may be sliced either by one single, individual, slicing angle, or with a plurality of different angles obtained by adjusting the angle during slicing of the food object.
[0015] The slicing parameters may comprise, or consist of, a stroke length, i.e. a length of the path of the knife or knife set relative to the conveyor belt surface, i.e. from the proximal position near the surface of the belt to the distal position of the stroke away from the belt.
[0016] The strokes may particularly comprise strokes with the same or with different stroke length. Accordingly, each food object may be sliced either by one single, individual, stroke length, or with a plurality of different stroke length obtained by adjusting the stroke length during slicing of the food object.When referring to the first food object and the subsequent food object in the row of food objects, it could be any two subsequent food objects in the row of food objects, e.g. number 1 and 2, 2 and 3, 4 and 5, or 9 and 10 etc. I.e. simply two food objects following each other in the row. Accordingly, even though first and subsequent food object in a row of food objects is mentioned, it does not limit the disclosure to slicing of two food objects.
[0017] The term "food object" when used herein denotes any kind of food object, inter alia meat from any species of animal, e.g. from fish, poultry, cattle, and pork etc. Particularly, the term may cover fish, particularly in the form of fish fillets. Examples could include slicing of raw, smoked, graved, salted or in any other way processed fish, e.g. white fish or fish from the Salmonidae family including salmon. If the food object is fish, it may be a fish fillet, but possibly also a whole fish, a fish with or without tail and / or intestines, and / or other parts of a fish. The term "slicing" when used herein denotes cutting slices of a food object. This may include portioning in relatively large pieces or cutting very thin slices of the food object.
[0018] Particularly, the term slicing may be limited to the process of cutting thin slices of a fish fillet. This process is particularly useful in connection with raw, salted, smoked, or graved fish fillets such as fillets of trout, salmon and generally fish from the Salmonidae family. Slices in this respect may particularly be thin slices with a thickness below 10 mm, or even below 5 mm.
[0019] The sequence of strokes may particularly be determined based on a height of the food object above the conveyor belt and / or a shape of the food object.
[0020] The system may use the conveyor belt simultaneously with the slicing tool such that the food objects are conveyed stepwise or continuously forward while slicing. In this way, the slicing tool provides the slicing by movement of the food object past the slicing position and movement of the knife exclusively in a single slicing plane transverse to the conveying direction.
[0021] Typically, the angle of the slicing plane will be adjustable relative to the conveying direction and relative to the conveyor belt surface from 90 degrees down to near 0 degrees, e.g. down to 20-30 degrees.
[0022] The movement of the knife in the slicing plane includes movement towards and away from the conveyor belt surface along a path referred to herein as the slicing path. The movement of the knife itself when performing a cut is herein referred to as a stroke.Additionally, the movement of the knife may include movement in the slicing plane and transverse to the slicing path. This transverse movement is referred to as reciprocation.
[0023] The knife structure may comprise one or two knives, with or without reciprocation. In one embodiment, the knife structure comprises two knives where at least one of the two knives reciprocate in the slicing plane such that the knives move relative to each other. At least one of the knives may particularly be a serrated knife.
[0024] The CPU is programmed to define the strokes based on the height profile signal. Particularly, the CPU is programmed to define at least a corresponding stroke length and a stroke angle for a plurality of strokes.
[0025] The height profile signal could e.g., be a plurality of discreate positions along an upper surface of the food object, i.e. facing away from the conveyor belt, or a continuous curve representing the upper surface.
[0026] An example could be that the height profile signal corresponds to a relatively thick food object, and that could provide a heigh stroke angle, e.g., in the order of 60-90 degrees between the slicing path and the conveyor belt surface. Correspondingly, in another example the height profile signal may correspond to a relatively thin food object, and that could provide a low stroke angle, e.g., in the order of 20-60 degrees between the slicing path and the conveyor belt surface. In that way, the length of each slice may become more uniform.
[0027] Another example could be that the height profile signal corresponds to a food object which is relatively heigh, and that could provide a large stroke length, e.g., in the order of 60-90 mm between the distal and the proximal positions of the knife. Correspondingly, in another example the height profile signal may correspond to a relatively thin food object, and that could provide a small stroke length, e.g., in the order of 20-60 mm. between the distal and the proximal positions of the knife.
[0028] The stroke angle and / or the stroke length could be defined from the height profile signal e.g., in the form of empirical data, e.g. defined in a table containing different values of the height profile signal with corresponding stroke angles and / or stroke lengths.
[0029] The stroke angle and / or the stroke length could also be defined from the height profile signal e.g., in the form of a transfer function defining the stroke angles and / or stroke lengths as a function of the height profile signal.The plurality of strokes may comprise what herein is referred to as crossing strokes. These strokes extend across a crossover point between two food objects. Typically, the slicing parameters of the crossing strokes could be defined by either the first of these two food objects or from the subsequent food object, or it may be defined from both the first and the subsequent food object, e.g. by interpolation between slicing parameters corresponding to the first food object and slicing parameters corresponding to the subsequent food object.
[0030] Each stroke may comprise a clearance step between the distal position and an entry point where the knife enters the food object and a penetration step extending between the entry point and an exit point where the knife exits the food object. The clearance step could be anything from zero and up to the full available length until reaching the distal position.
[0031] The CPU may be configured to define each stroke such that the clearance step of a stroke is determined based on a clearance step of a subsequent stroke. This will allow the knife to become sufficiently free above the food object such that it can pass the food object above the food object while the food object is transported by the conveyor belt.
[0032] The slicing parameters for each stroke may define a slicing speed defining a speed of the knife when moving through the food object. As an example, it may define a relatively high speed when the food object is thick and a lower speed, when the food object is thin.
[0033] The sensor could be a height sensor configured to determine a height of the food object above the conveyor belt surface. The height sensor could be optical, e.g., based on a linear array of light sensors, or a camera. Alternatively, the height sensor is mechanical, e.g. comprising a hinged element which follows the shape of the food object and registers a rotation of the element at the hinge and based on this rotation the height along the transported food product is determined.
[0034] Particularly, the height sensor could determine a consecutive array of height data each representing a height of the food object above the conveyor belt surface.
[0035] Sometimes, the strokes may vary periodically such that one sequence of strokes corresponds to one food object, and that sequence is repeated with or without minor adjustments for the subsequent food object in the row of food objects.
[0036] The CPU is programmed to determine the strokes in sequences of strokes, where each sequence of strokes corresponds to one of the food objects in the row of food objects.When slicing fish fillets, taken as an example, the sequence may include a pattern defining three parts, a head-end, a mid-section, and a tail-end, where the slicing angles and / or the speed deviates between these three parts, e.g., such that the slicing speed is lower in the tail end. This pattern can be replicated from one fish fillet to the next, e.g. with adjustments of the slicing speed and / or angle depending on the actual size of the fish fillet.
[0037] The CPU may further be programmed to determine a crossover point between the first food object and the subsequent food object based on the height profile signal.
[0038] The CPU may further be programmed to shift between different sequences of strokes based on the height profile signal. Accordingly, the system becomes independent on a specific location of one food object relative to the preceding or subsequent food object.
[0039] The food objects may therefore be placed arbitrarily on the conveyor belt surface, e.g., directly one after the other with no distance between the food objects, or with an overlap between a food object and the subsequent food object, or with a distance between two subsequent food objects. This may potentially improve the performance of the system since the food objects can be placed closer to each other and thereby fill up the conveyor belt surface and increase the number of sliced food objects per time unit. Moreover, it simplifies feeding of food objects into the system since the food objects may be placed more randomly than in existing slicing systems where a certain distance between e.g. two fish fillets are important for the CPU to determine the end of one food object and the start of the next food object and to calculate a sequence of strokes defining a cutting pattern along each individual food object based on the start and end of the fillet together with the fillet heigh profile of the food object.
[0040] The term "overlap" herein means that a line extending perpendicular to the downstream conveying direction can cross through two food objects which thereby overlap on the conveyor belt.
[0041] The CPU may be programmed to determine a row of successive minimum height values, herein referred to as local minimum values and optionally also maximum height values from the height data and to select the crossover point based the local minimum values.
[0042] The minimum height values and the maximum height values could be determined by:a. Defining consecutive data windows, where each data window defines a consecutive subset of the height data. The data window may as an example comprise 5 consecutive height measurements in the array of height data.
[0043] b. Fitting each data window to line segments. This provides a number of consecutive line segments representing an approximated upper contour of the food object.
[0044] c. Determining a slope of each line segment.
[0045] And
[0046] d. Using the slope of each line segment to identify a minimum or a maximum height. This could be done simply by comparing the slope direction of each line segment. A negative slope followed by a positive slope identifies a local minimum value and a positive slope followed by a negative slope identifies a local maximum value.
[0047] The data window could be fitted to line segments by a least square method or by other mathematical principles known in the art.
[0048] The CPU may be programmed to select from the row of successive local minimum values and optionally also the maximum height values, one minimum height value being considered as an optimal crossover candidate. This could inter alia be done according to the below method steps which could all be programmed into the CPU.
[0049] a) A length range is defined. The length range represents a length which is considered reasonable for the type of food objects presently being sliced. The length range could be fixedly defined in the CPU or it could be a redefinable range. The length range may e.g. specify a minimum reasonable length of the food object and a maximum reasonable length of the food object. The CPU is thereby configured to identify a range from a value x to a value y within which the food object should be. The range could e.g., be 250 mm. to 750 mm.
[0050] b) What is referred to herein as "a passed length of the food object" is defined. The passed length represents a length of the food object which has passed the sensor since the previous crossover point. The passed length is compared with the minimum reasonable length, and when the passed length exceeds the minimum reasonable length, the CPU is programmed to search for the next local minimum value and tocharacterise that local minimum value as being the crossover point, or at least as being a candidate for a crossover point.
[0051] c) As an additional feature, the CPU may be programmed to define what is referred to herein as "a maximum reasonable height value". This represents a height which is considered as the maximum height which is likely to occur at the crossover point. In this case, the CPU may, in addition to the step b) also compare the candidates for a crossover point, i.e. the local minimum values detected after reaching the minimum reasonable length, with the maximum reasonable height. One control approach could be to select the first occurring local minimum after having reached the minimum reasonable length and not exceeding the maximum reasonable height value or a certain percentage thereof, e.g., not exceeding 70 percent of the maximum reasonable height value.
[0052] d) The maximum reasonable height may be a fixed value in the CPU. However, the CPU may additionally be programmed to define the maximum reasonable height value dynamically based on signals from the sensor. This may be done in different ways, inter alia:
[0053] a. By setting the maximum reasonable height value to a certain percentage of a sensed maximum height. The sensed maximum height could be defined either within the present length range, or for a specific period of time, e.g. within the latest hour of measurement.
[0054] b. By defining an average height of all sensor data in a specific period of time and defining the maximum reasonable heigh to be a certain percentage of that average value.
[0055] The optimal crossover candidate is one of the minimum height values which are considered particularly likely to represent the place where one food object ends, and a subsequent food object begins in the row of food objects.
[0056] The optimal crossover candidate could, alternatively, be selected by:
[0057] a. comparing a row of successive minimum height values to identify a lowest one of the row of minimum height values; and
[0058] b. selecting the lowest one of the height values to be the crossover candidate when it reaches a predetermined distance between the front of the food object and the slicing position. This predetermined distance could be a fixed distance,e.g. 5 cm, or a distance depending on the speed by which the food object is conveyed on the conveyor belt.
[0059] The CPU may comprise a user interface such as a touch screen or keyboard allowing input from an operator or from an automatic infeed system of a placement parameter which identifies a desired overlap of food objects in the row of food objects. As an example, the user may input manually said length range, said maximum reasonable height, and additionally, a preferred overlap between each food object on the conveyor belt.
[0060] The conveyor belt surface may comprise an indicator visualising a distance or overlap between food objects in the row of food objects. This may facilitate an easier and faster infeed. The indicator may be any kind of indication on the conveyor band which indicates a position on the band, inter alia dots or lines printed on the conveyor belt or projected by light onto the conveyor belt, e.g. lines extending transverse to the downstream conveying direction. The lines may indicate to the user which overlapping distance there is between two subsequent food objects on the conveyor belt. The lines may indicate a degree of overlap to the user and thereby characterize the row of food objects. This can be used as an input, i.e. the operator may input a degree of overlap by which the food objects are placed on the conveyor belt, and that input may be used by the CPU when defining one or more of the slicing parameters such as the slicing angle, slicing length, and / or slicing speed.
[0061] Alternatively, the degree of overlap may be provided by a vision system or other sensor means.
[0062] The user interface may refer to the indicator and thereby communicate to the operator where the food objects should be placed on the conveyor belt surface. The indicator could be printed marks of the surface of the conveyor belt.
[0063] The CPU may be configured to define separate sections of the food objects wherein each section is associated with a scheme for defining the slicing parameters. If the food object, as an example, is a fish or fish fillet, the sections could define the tail end, a centre part, and a head end. The CPU could be configured to define the slicing parameters for a specific section of the food object based on the scheme for that section of the food object.
[0064] The CPU may be configured to determine the slicing parameters depending on a thickness of the food object. Herein, the thickness of the food object is defined as the height above the conveyor belt surface. Moreover, the slicing parameters may depend on a desired slicing length through the food object, and thereby a desired size of the slice.The slicing parameters may, in addition to the slicing angle comprise a variable selected from the group consisting of slicing speed, reciprocation speed, and reciprocation amplitude.
[0065] In a second aspect, the invention provides a method of slicing a food object, the method comprising
[0066] a. determining a height profile of one or more food objects arranged in a row of food objects;
[0067] b. determine a crossover point between a first food object and the subsequent food object based on the height profile;
[0068] c. determining a sequence of strokes for each food object in the row of food objects; and
[0069] d. shifting between two different sequences of strokes based on the determined crossover point.
[0070] The order of the steps may vary, e.g. such that the step c is before step b etc.
[0071] The method may particularly be carried out with respect to a fillet or loin of a fish, and it may particularly include the steps mentioned relative to the programming of the CPU of the system according to the first aspect, particularly the method of determining the crossover point based on local minimum values.
[0072] A plurality of food objects may be arranged on the conveyor belt with different overlapping distance, and the overlapping distance may be classified based on a visual indication on the conveyor belt.
[0073] The method may comprise continuously sampling the height of the profile. Each time a sample has the lowest height so far, this sample is selected as the crossover candidate.
[0074] The crossover candidate may further be evaluated in terms of the resulting length of the food object, e.g. a fillet length of a fish fillet.
[0075] This length must be within predefined limits.
[0076] While searching for a good crossover point, the method may include measuring how far the sample point is from the current best crossover candidate, and the step of stopping sampling when this length reaches the minimum length which is predefined. This is to avoid a smallfood object getting sampled along with a larger food object while still allowing time for a potentially better crossover candidate to be found.
[0077] An important feature of this, is that the resulting height profile also contains part of the height profile of a following food object. This is necessary, to be able to calculate a smooth transition of slices between the two food objects. To differentiate between the two food objects in the same height profile, the same crossover point is used.
[0078] LEGENDS TO THE FIGURE
[0079] Figs. 1 and 2 illustrate a device for slicing food objects,
[0080] Fig. 3 illustrates degrees of freedom for the slicing of food objects,
[0081] Fig. 4 illustrates a sensor for determining a height profile signal representing a shape of an upper surface of the food object,
[0082] Fig. 5 illustrates sequences of strokes with corresponding stroke length and stroke angle,
[0083] Fig. 6 illustrates differences between adjacent strokes,
[0084] Figs. 7-10 illustrates a conveyor belt with overlapping food objects,
[0085] Figs. 11-13 illustrate and
[0086] Fig. 14 illustrates method steps in a slicing method.
[0087] DETAILED DISCLOSURE OF EMBODIMENTS
[0088] The following disclosure is based on slicing of fish, especially smoked fish. However, the system could also be configured for slicing other food objects.
[0089] Figs. 1 and 2 illustrate a system 1 comprising a belt conveyor 2 moving the food objects, in the case fish fillets 3, forward in a downstream direction indicated by the arrow 4 towards the slicing tool 5.Belt conveyor 2 comprises a first section 6, a second section 7, and a third section 8. Each section includes separate conveyor belts and can move individually such that a fish fillet on the second section 7 can be brought stepwise forward during slicing while fish fillets on the first section and cut fish fillets on the third section can be moved continuously towards and away from the second section, respectively.
[0090] The slicing tool 5 is illustrated in further details in Fig. 2 and is configured for slicing the fish with a knife structure 15. The knife structure moves in a slicing path indicated by arrow 10 towards and away from a slit between the second and the third sections 7, 8 of the belt conveyor 2. In this process, the slicing tool moves through the fish fillet along the slicing path.
[0091] The system comprises a sensor 9 with an arm 10 which follows an upper surface of the food object and thereby allows the sensor to define a height profile signal. Other sensors could include optic elements such as cameras and vision recognition software to define the height profile signal.
[0092] The slicing parameter used by the slicing tool when slicing includes specifically the angle of the slicing path relative to a plane defined by the belt conveyor, herein referred to as stroke angle, and the length of the stroke, i.e., the distance between the distal and the proximal position of the knife.
[0093] The slicing parameters are defined based on the height profile signal of the fish by the CPU 11. Additionally, the CPU may control other features of the slicing tool including controlling the angle of the knife structure relative to the surface of the second section 7 of the belt conveyor 2. Additionally, the CPU may control other slicing parameters such as the speed of the knife when it moves through the food object etc.
[0094] The CPU 11 may use standard hardware circuits, using software programs and data in conjunction with a suitably programmed digital microprocessor or a general-purpose computer e.g. a PLC programmed with suitable computer code for enabling determining of the slicing parameter and the crossover point based on the height profile signal. The CPU may be an application specific integrated circuitry, or it may be a standard CPU e.g. using standard servo systems or step motors controllers for controlling servo drives or step motors of the slicing tool to thereby apply the slicing parameters in accordance with the defined sequence, i.e., including the slicing angle and / or other slicing parameters such as stroke length, and reciprocation which are adequate for the specific food object being sliced.The CPU 11 may comprise memory and computer executable code for enabling various functions including reading the contour data from the sensor.
[0095] Software program instructions and data may be stored on a non-transitory, computer-readable storage medium, and when the instructions are executed by the CPU these functions are carried out.
[0096] Fig. 3 illustrates schematically the primary controlled degrees of freedom. The arrow 4 illustrates the freedom of the belt conveyor to move the food object forward, the arrow 12 illustrates the freedom of the knife structure to move along the slicing path, the arrow 13 illustrates the ability to change the angle of the knife structure relative to the surface of the conveyor belt, and arrow 14 illustrates an optional reciprocation of the knife structure or at least of one or more knifes of the knife structure. Each of these movements could be controlled by the CPU 11.
[0097] Fig. 4 illustrates schematically the sensor 9 for providing the height profile signal. In this case, the sensor is configured to determine the width of the fish corresponding to the height of the fish fillet lying on the conveyor belt. The sensor comprises a hinged plate element 40 under which the fish is conveyed. The fish lifts the hinged plate element while the plate element pivots about a pivot point 41. The angular displacement of the hinged plate determines the height of the fish.
[0098] Fig. 5 illustrates two sequences of strokes, wherein each stroke is defined by the stroke length and the stroke angle. In Fig. 5, the abscissa illustrates a length along the conveyor belt measured in mm. and the ordinate illustrates a height over the conveyor belt measured in mm.
[0099] In Fig. 5, continuous line 50 defines the contour of the food objects. The contour is determined by the sensor while the food objects pass the sensor. The contour corresponds to the highest points of the product when being transported on the conveyor belt i.e. to the height of the product.
[0100] The strokes are illustrated by the inclined lines below the continuous line 50. Each of the inclined lines has a length indicating the stroke length and an angle indicating the slicing angle of that stroke. Each intersection between the continuous line 50 and each of the strokes defines an entry point, exemplified at the entry point 55, where the knife enters the food object.The dotted line 51 defines a crossover point determined to distinguish a first food object from a subsequent food object in an area of overlap between the two food objects. The crossover point is found by the CPU by identifying a plurality of local minimum values. Some of these are identified by numeral 56, 57, and 58. The dotted lines 59, and 60 indicate a length range defined in the CPU and consider as a reasonable range of length of the food objects presently being sliced. The dotted line 59 indicates a minimum reasonable length and the dotted line 60 indicates a maximum reasonable length. The first local minimum value 58 occurring after the minimum reasonable length is selected as the most promising candidate as a crossover point, and assigned to be the crossover point 51
[0101] The first and second sequences of strokes are indicated by brackets 52 and 53, and they contain 4 crossing strokes which extend across the crossover point through both the first food object and the subsequent food object. The crossing strokes are indicated by bracket 54 and they belong to both the first and subsequent sequence 52, 53 of strokes.
[0102] Fig. 5 illustrates each sequence of strokes comprises strokes having different stroke length. The stroke length is determined by the CPU based on the contour in order to move the knife least possible along the slicing path since that saves not only time, but it also reduces wear and power consumption.
[0103] Fig. 5 further illustrates that each sequence of strokes comprises strokes having different stroke angle. In the illustrated example, the angle relative to the surface of the conveyor belt is generally height when the food object is higher. That allows the penetrated length through each slice to become more uniform. The penetrated length may, after the slicing, be seen as the length of the slice of the food object, and the variation of the stroke angle may therefore be controlled to obtain a more uniform length of the slices.
[0104] Each stroke may comprise a clearance step between the distal position and an entry point where the knife enters the food object and a penetration step extending between the entry point and an exit point where the knife exits the food object.
[0105] In the example illustrated in Fig. 5, most of the strokes have only a very short clearance step meaning that the knife stops almost immediately when it gets out of the food object.
[0106] However, strokes 61 and 62, taken as an example, have a larger clearance step indicated by the bracket 63. The clearance steps of these two strokes are defined based on the subsequent stroke 64 which are longer due to the increasing contour of the food object.By programming the CPU to define the clearance step based on the subsequent stroke, the knife is allowed to pass above the food object, when the food object is conveyed by the conveyor belt.
[0107] EXAMPLE 1
[0108] The following exemplifies one way of determining local minimum height values and selecting a crossover point.
[0109] The example uses a function to find the point where one food object ends, and the next begins, i.e. the crossover point. This point is called the crossover point. A function is created to select candidates for crossover points, while the food object is being measured by the height sensor. This is considered to work for various methods of height sensing.
[0110] Finding crossover candidates is done by statistics for a line by using a least square method to calculate a straight line that best fits your data and then returns an array that describes the line.
[0111] Example of data for a fish fillet is shown in the table below in which the window size is 5 and the middle of the window is 3. In the table below, x is the moved distance provided by the conveyor belt and y is the height of the food object.
[0112] II 1 2 3 4 5 6 7 8 9 10
[0113] X 0 0,4 0,8 1,2 1,6 2 2,4 2,8 3,2 3,6
[0114] Y 2 3 3 3,2 4 4,2 4 5 5,3 5,5
[0115] Window[i]=Window[l]
[0116]
[0117] In this case the size of the data window is 5. The value of the data window is fitted to the data from the application.A first step is the calculation of slope for Window[i], where i = l. This is done by iterating over all data in the given data window, c.f. the below formula:
[0118] sczelVmdow sizeWindow x(n) n=i X
[0119]
[0120] _avr = Y_avr =
[0121] sizeWindow sizeWindow
[0122] where sizeWindow is a number of samples in this window.
[0123] Then, the least square method is used to calculate the slope(i) for a given data window, i:
[0124] [formula - see imgf000016_0002]
[0125] ii=l
[0126] slope(i) =
[0127] sizeWindow
[0128] Σ (x_avr - x(ii))²
[0129]
[0130] This is repeated for all data windows, as illustrated below:
[0131] 1 2 3 4 5 6 7 8 9 10 x 0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 y 2 3 3 3.2 4 4.2 4 5 5.3 5.5
[0132] Wind
[0133] Window [21
[0134] Window[3]
[0135] Window[4]
[0136]
[0137] The values of slopes are now used for finding the minimum and maximum height, i.e. a negative slope followed by a positive slope indicates a minimum value (= minimum heightabove conveyor belt) and a positive slope followed by a negative slope indicates a maximum value (= maximum height above conveyor belt).
[0138] Pseudo Code for finding minimum height:
[0139] # Finding the change from negative to positive slope: minimum Height for i=l: sizeSlope-1
[0140] if (slope(i)<0)&& (slope (i+l)>=0)
[0141] abs(slope(ii)) < abs (slope(i+l))
[0142] minSampleNew = i;
[0143] minHeightNew = y(i+midWindoW“l)
[0144] else if
[0145] minSampleNew = i+1
[0146] minHeightNew = y(i+midWindow)
[0147] endif
[0148] if(minHeightNew < minHeight)
[0149] minHeight = minHeightNew
[0150] minSample = minSampleNew
[0151] endif
[0152] / / Find xMinIntercept
[0153] xMinIntercept = x(minSample+midWindow-1)
[0154] Referring to the pseudo code, xMinIntercept can be used for identifying new food objects, for example it can be a hardcoded offset from the MinIntercept that defines where the new food object starts. The hardcoded value could be chosen to give the best quality of slices. The chosen value of the offset could be based on empirical experiments and always related to a distance by which the food objects overlap on the conveyor belt.
[0155] When a number of minimum values are found, the next step is to select one of these minimum values as a candidate for being a crossover.
[0156] When a local minimum is found, it is compared to the highest point of the current food object, a highest point of food objects registered within a certain duration or a highest theoretical point keyed in by an operator etc.If the local minimum is under a certain percentage, e.g. 70%, of the highest point, it is considered to be a valid crossover candidate. If not, the point is deemed to be a bump in the surface of the food object, and not a potential crossover.
[0157] Only one crossover candidate can exist at a time. When a second crossover candidate is calculated, the two are compared and the best is kept while the other is discarded. The selection of the best crossover candidate is, currently, based on the lowest height of the two.
[0158] A minimum reasonable length of the food object and a maximum reasonable length for the food object, e.g. 150mm to 800mm, or 250-750mm may be defined.
[0159] To improve the selection of the best crossover point, the CPU may be programmed to wait until the current food object measures 800mm, and then select the lowest minimum. The problem with this approach, is that multiple minimum length food objects can fit inside one maximum length of the food object. This means that the number of food objects and thereby the number of crossover points is unknown.
[0160] Therefore, as soon as a local minimum height value is considered as a crossover candidate, and that local minimum results in a food object being longer than the minimum reasonable length, that local minimum is selected as the crossover point.
[0161] If no crossover candidate is found within the maximum fillet length, the crossover point is selected as being where the fillet matches the maximum length. This will happen if the fillets are placed as overlapping too much, or the crossover candidate selection software is not tuned correctly.
[0162] I.e., a maximum reasonable length of the food object may be implemented as a "forced" crossover - i.e. if no crossover is determined, a fixed value of an expected maximum length, i.e. the aforementioned maximum reasonable length of the food object is used to determine when there necessarily must have been a crossover, and that is considered by the CPU as the beginning of a new food object. Following this point, the CPU will await for the minimum reasonable length to have passed the sensor and looks for a new candidate amongst the local minimum values to select as the next crossover point, and the procedure repeats itself, i.e., if no candidate is selected, the next crossover point will be at the distance corresponding to the maximum reasonable length of the food object away from the previous crossover point.
[0163] Referring to Fig. 5, the first food object contains a number of local minimum height values including those marked with 56, 57. When a local minimum height value is determined, e.g.,by the above explained procedure, a flag may be generated indicating whether the local minimum height value is a candidate for a crossover point or not. Particularly that flag can be set when the minimum height value is lower than a threshold. The threshold may e.g., be based on the height of the food object, an average height of many food objects or a height of the highest food object in a row of food objects. Herein, reference for this height is made as "a maximum reasonable height". In one example, the threshold could be a specific percentage such as 50, 60, 70, or 80 percent of the maximum reasonable height. Only when the local minimum height value is below that threshold, it is flagged as a candidate for a crossover point.
[0164] In a specific example, the maximum reasonable height is selected as the highest point detected since last crossover point and it is 40 mm. The threshold is defined as 70 percent of the maximum reasonable height, i.e. 28 mm. The CPU is configured to compare different local minimum height values with the 28 mm and only add a flag to those being lower than 28 mm. Finally, the CPU selects amongst the flagged values, a best candidate. That could be the first arriving after the minimum reasonable length of the food product has passed since the previous crossover was determined.
[0165] In another specific example, the maximum reasonable height is selected as an average of all hight measurements during a certain period of time. Again, this value may e.g. be 40 mm. making the rest of the example comparable with the previous example.
[0166] If the food object is followed directly by another food object on the conveyor belt, it may be desired to let the last slices of the first food object to extend beyond the height profile and into the subsequent food object. That is illustrated by the four strokes marked with the bracket 54 in Fig. 5.
[0167] Fig. 6 illustrates variation of stroke angle between two subsequent strokes. The angle variation leads to thickness variations and seeks to maintain the length of each slice within a certain range. Fig. 6 illustrates a first slice 65, and a subsequent slice 66. The angle of the first slice is indicated at 67, and the angle of the subsequent slice is indicated at 68 thereby making a thickness difference, indicated at the slice length d, between the two slices. The thickness difference is indicated by the line 69, i.e. where the length indication d intersects the line 69.
[0168] Figs. 7-10 illustrate different ways of placing food objects on a conveyor belt.In fig. 7, two conveyor belts 7, 7' extend in parallel. The system comprises a sensor with a sensor arm 10 which extends over both belts and a knife 15 which extends over both belts. The food objects are arranged arbitrarily on the conveyor belts. On one belt, the two food objects overlap not only in downstream direction of the conveying, but they are also arranged partly over each other. On the other conveyor belt, the two food objects overlap in the downstream direction, but they do not touch each other Due to the overlap, the two food objects 3 are sliced simultaneously by use of a knife 15 extending the full width of the conveyor belts.
[0169] The visual indication lines 70, 71, 72, 73 are printed on the surface of the conveyor belt, or they are made as projected illumination on the belt. The lines support the operator when placing the food objects by indicating a level of overlap.
[0170] In Fig. 8 it is illustrated that both food objects end in a space between the same indication lines 71-72 thereby indicating a level 1 overlap. In Fig. 9, it is indicated that the food objects end in adjacent spaces between different indication lines, 70-71 and 71-72, respectively. This indicates a level 2 overlap.
[0171] In Fig. 10, there is no overlap between the food objects, but they are arranged end-to-end. Due to the CPU which is programmed to identify the crossover point, the food objects can be placed arbitrarily, and the sequences of strokes with sets of slicing parameters are selected based on the identified crossover point.
[0172] Figs. 11-13 illustrate flowcharts of the process of determining crossover points. Each process could be implemented in the CPU of the system.
[0173] In Fig. 11:
[0174] 110 is the process header for the process of providing profile measurement.
[0175] 111 is a process for determining crossover candidate. This is further explained in Fig. 12.
[0176] 112 is a process of selecting crossover point. This is further explained in Fig. 13.
[0177] In Fig. 12:
[0178] 120 is the process header for the function of determining a crossover candidate.121 is a process reading heigh input in the form of an array of height measurements.
[0179] 122 is a process of finding local minimum height values, e.g. by a least square method.
[0180] 123 is a process of differentiating between whether the local minimum height value is below a threshold. Subsequent process 124 represents a function to be selected if false, and subsequent process 125 represents a process to be selected if true.
[0181] 124 is a process that disregards the local minimum height value as a crossover candidate.
[0182] 125 is a process that saves the local minimum height value as a new crossover candidate.
[0183] 126 is a process of differentiating between whether the new crossover candidate is better than a current optimal crossover candidate. Subsequent process 127 represents a process to be selected if false, and subsequent process 128 represents a process to be selected if true.
[0184] 127 is a process that disregards the crossover candidate.
[0185] 128 is a function that replaces the current optimal crossover candidate with the new crossover candidate as being the crossover point.
[0186] 129 ends the process.
[0187] In Fig. 13:
[0188] 130 is the process header for the function of selecting crossover point.
[0189] 131 is a process of differentiating between whether a crossover candidate exists. Subsequent function 132 represents a function to be selected if true, and subsequent function 135 represents a function to be selected if false.
[0190] 132 is a process of differentiating between whether the crossover candidate is located beyond the minimum reasonable length. Subsequent function 133 represents a function to be selected if true, and subsequent function 135 represents a function to be selected if false.
[0191] 133 is the process of selecting the crossover candidate as the crossover point.
[0192] 134 ends the process of selecting crossover point.135 is a process of differentiating between whether the current food object is longer than a maximum reasonable length. Subsequent process 136 represents a process to be selected if true, and subsequent process 137 represents a function to be selected if false.
[0193] 136 is a process of selecting the current position of the food object as a crossover point.
[0194] 137 ends the process of selecting crossover point.
[0195] Fig. 14 illustrates method steps of a slicing method.
[0196] In step A, the height profile of food objects arranged in a row of food objects is determined by a sensor.
[0197] In step B, a crossover point between a first food object and the subsequent food is determined object based on the height profile, particularly by use of a least square method in which a plurality of local minimum is determined and one of these local minimum values is selected as the most promising point for a crossover.
[0198] In step C, a sequence of strokes for each food object in the row of food objects is determined based on a height profile signal from the sensor.
[0199] In step D, the slicing shifts between two different sequences of strokes based on the determined crossover point, i.e. when one food object is replaced by a subsequent food object at the slicing position.
Claims
CLAIMS1. A system (1) for slicing food objects, the system comprising:a conveyor belt (2) configured for conveying a row of the food objects (3) in a downstream direction (4) across a slicing position;a sensor (9) configured to provide a height profile signal defining a height profile of the food objects (3);a slicing tool (5) arranged for slicing the food objects by sequentially:• moving the food object on the conveyor belt in the downstream direction to define a slice thickness, and• providing strokes of a knife towards and away from a surface of the conveyor belt, each stroke extending between a proximal position near the surface of the conveyor belt and a distal position away from the surface of the conveyor belt thereby defining a stroke length being a distance between the proximal position and the distal position,a CPU (11) programmed based on the height profile signal, to define strokes each having a corresponding set of slicing parameters defining at least a corresponding stroke length and a stroke angle.
2. The system according to claim 1, wherein the strokes comprise strokes having different stroke length.
3. The system according to any of the preceding claims, wherein the strokes comprise strokes having different stroke angle.
4. The system according to any of the preceding claims, wherein each stroke comprises a clearance step between the distal position and an entry point where the knife enters the food object on a distal side of the food object facing away from the surface of the conveyor and a penetration step extending between the entry point and an exit point where the knife exits the food object on a proximal side of the food object facing towards the surface of the conveyor.
5. The system according to claim 4, wherein the CPU is configured to define each set of slicing parameters such that the clearance step of a stroke is determined based on a subsequent stroke.
6. The system according to any of the preceding claims, wherein the set of slicing parameters for each stroke comprises slicing speed defining speed of the knife when moving through the food object.
7. The system according to any of the preceding claims, wherein the sensor is a height sensor configured to determine a consecutive array of height data each representing a height of the food object above the conveyor belt surface.
8. The system according to claim 7, wherein the height sensor extends over several conveyor belts.
9. The system according to any of the preceding claims, wherein the CPU is programmed to determine the strokes in sequences of strokes, where each sequence of strokes corresponds to one of the food objects in the row of food objects, and wherein the CPU is further programmed to determine a crossover point between a first food object and a subsequent food object based on the height profile signal, to assign a first sequence of strokes to the first food object in the row of food objects and a subsequent sequence of strokes to the subsequent food object in the row of food objects, and to control the slicing tool in accordance with the sequences of strokes such that the first sequence of strokes is replaced by the subsequent sequence of strokes based on the crossover point during slicing of the food object.
10. The system according to claim 9, wherein the CPU is programmed to determine a row of successive local minimum height values.
11. The system according to claim 10, wherein the local minimum height values are determined by:a. defining consecutive data windows, where each data window defines a consecutive subset of the height data;b. converting each data window to line segments;c. determining a slope of each line segment; andd. using the slope of each line segment to identify the local minimum height.
12. The system according to claim 11, wherein the data window is converted to line segments by a least square method.
13. The system according to any of claims 9-12, wherein the CPU is configured to define a length range representing a length which is considered reasonable for the type of food objects presently being sliced and specifying a minimum reasonable length of the food object and a maximum reasonable length of the food object.
14. The system according to any of claims 9-13, wherein the CPU is configured to define a passed length of the food object representing a length of the food object which has passed the sensor since the previous crossover point.
15. The system according to any of claims 9-14, wherein the CPU is configured to define a maximum reasonable height representing a height which is likely to occur at the crossover point.
16. The system according to any of claims 10-15, wherein the CPU is programmed to select from the row of successive local minimum height values, a selected one being considered as an optimal crossover candidate.
17. The system according to one of claims 13-15 and claim 16, wherein the optimal crossover candidate is selected based on at least one of the length range, the passed length, and the maximum reasonable height.
18. The system according to claims 9-17, wherein the strokes comprise crossing strokes extending across the crossover point through both the first food object and the subsequent food object.
19. The system according to any of the preceding claims, wherein the CPU defines a user interface allowing input of a placement parameter which identifies a desired overlap of food objects in the row of food objects and wherein the CPU is configured to define the slicing parameters and / or the crossover point based on the desired overlap.
20. The system according to any of the preceding claims, wherein the conveyor belt surface comprises an indicator visualising a distance or overlap between food objects in the row of food objects.
21. The system according to claim 19 and 20, wherein the user interface refers to the indicator.
22. The system according to any of the preceding claims, wherein the CPU is configured to define separate sections of the food objects wherein each section is associated with a scheme for defining the slicing parameters, and wherein the CPU is configured to define the slicing parameters for a specific section of the food object based on the scheme for that section of the food object.
23. The system according to any of the preceding claims, wherein the CPU is configured to determine the slicing parameters depending on a thickness of the food object, and a desired slicing length through the food object.
24. A method of slicing a food object, the method comprisingdetermining a height profile of a plurality of food objects arranged in a row of food objects on a conveyor belt;determine a crossover point between a first food object and the subsequent food object based on the height profile;determining a sequence of strokes for each food object in the row of food objects; andshifting between two different sequences of stokes based on the determined crossover point.
25. The method according to claim 24, wherein the food object is a fillet or loin of fish.
26. The method according to claim 24 or 25, wherein a plurality of food objects is arranged on the conveyor belt with different overlapping distance.
27. The method according to claim 26, wherein the overlapping distance is classified based on a visual indication on the conveyor belt.