Residue collector

Abstract

The invention relates to a residue collector (200) operable to receive residue from a combine harvester (210) during a training harvesting operation, wherein the residue collector (200) comprises: a residue separator (205) for separating the processed residue into a first portion and a second portion based on a characteristic of the processed residue; one or more weight sensors (204, 207, 208) for directly or indirectly determining the weight of the first portion and the second portion; and a controller configured to determine a quality factor of the processed residue based on the determined weight of the first portion relative to the weight of the second portion.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a residue collector that receives crop residue directly from a combine harvester, and the subsequent use of information that can be determined from the received crop residue. SUMMARY

[0002] According to a first aspect of the application, there is provided a residue collector operable to receive residue from a combine harvester during a training harvest operation, wherein the residue collector comprises:

[0003] a residue separator for separating the processed residue into a first portion and a second portion based on a characteristic of the processed residue;

[0004] one or more weight sensors for determining, directly or indirectly, the weight of the first portion and the second portion; and

[0005] a controller configured to determine a quality factor for the processed residue based on the determined weight of the first portion relative to the weight of the second portion.

[0006] Advantageously, such a residue collector can provide a better and more efficient means and method for quantifying the quality of residue deposited in a field. It can also collect more data in different places and circumstances. Furthermore, it can correlate this data with settings of the residue handling system of the combine harvester (or other parameters of the combine harvester) in an efficient and effective manner, which is useful for subsequent harvesting operations.

[0007] The controller can be configured to determine the quality factor during the training harvest operation.

[0008] The controller can be further configured to provide an indication of the determined quality factor to an operator of the combine harvester during the training harvest operation.

[0009] The residue separator can be for separating the processed residue into three or more portions based on one or more characteristics of the processed residue. The one or more weight sensors can be for determining, directly or indirectly, the weight of each portion. The controller can be configured to determine a quality factor for the processed residue based on the relative weights of the three or more portions.

[0010] The characteristic of the residue for separating it into the first portion and the second portion can comprise one or more of:

[0011] the size of components in the residue;

[0012] the shape of components in the residue;

[0013] the density of components in the residue;

[0014] moisture content of the residue; and

[0015] color of the residue.

[0016] The residue collector can comprise a trailer that is, in use, towable by the combine harvester.

[0017] The residue collector can have a residue collection configuration and a residue bypass configuration. In the residue collection configuration, the residue collector can be configured to divert the residue to components of the residue collector to determine a quality factor of the residue. In the residue bypass configuration, the residue collector can be configured such that the residue bypasses or avoids components of the residue collector to determine a quality factor of the residue.

[0018] The residue collector can further comprise a residue selection component for selectively diverting only a portion of the received residue to the residue separator.

[0019] The controller can be further configured to:

[0020] receive one or more sensor values from sensors associated with the combine harvester; and

[0021] store the one or more sensor values and the associated determined quality factor as training data for a machine learning algorithm.

[0022] The controller can be further configured to train a machine learning algorithm based on the training data, wherein the trained machine learning algorithm is for subsequent use during a harvesting operation.

[0023] The controller can be further configured to:

[0024] receive one or more sensor values from sensors associated with the combine harvester;

[0025] receive one or more operating parameters of the combine harvester, the operating parameters corresponding to a time at which the residue was harvested; and

[0026] train a machine learning algorithm based on the one or more sensor values, the one or more operating parameters, and the determined quality factor, wherein the trained machine learning algorithm is for subsequent use during a harvesting operation.

[0027] A controller for a combine harvester is also disclosed, the controller being configured to:

[0028] receive one or more sensor values from sensors associated with the combine harvester during a harvesting operation;

[0029] determining a computed quality factor using a machine learning algorithm that has been trained by any of the residue collectors disclosed herein and the received one or more sensor values; and

[0030] presenting the computed quality factor to an operator of the combine harvester during a harvesting operation, or setting one or more operating parameters of the combine harvester during the harvesting operation based on the computed quality factor.

[0031] The controller can also be configured to set one or more operating parameters of the combine harvester during a harvesting operation based on the computed quality factor and also based on a target quality factor.

[0032] A controller for a combine harvester is also disclosed, the controller being configured to:

[0033] receive one or more sensor values from sensors associated with the combine harvester during a training harvesting operation;

[0034] receive a target quality factor; and

[0035] determine one or more computed operating parameters using a machine learning algorithm that has been trained by any of the residue collectors disclosed herein, the received one or more sensor values, and the target quality factor and apply them to the combine harvester during the training harvesting operation.

[0036] The one or more operating parameters can include one or more operating parameters of a residue handling component of the combine harvester.

[0037] A method is also disclosed, comprising:

[0038] receiving residue directly from a combine harvester during a harvesting operation;

[0039] separating the residue into a first portion and a second portion based on a characteristic of the residue;

[0040] directly or indirectly determining a weight of the first portion and the second portion; and

[0041] determining a quality factor of the residue based on the determined weight of the first portion relative to the weight of the second portion.

[0042] A method of operating a combine harvester is also disclosed, the method comprising:

[0043] receiving one or more sensor values from sensors associated with the combine harvester during a harvesting operation;

[0044] determining a computed quality factor using a machine learning algorithm that has been trained by any of the residue collectors disclosed herein and the received one or more sensor values; and

[0045] presenting the computed quality factor to an operator of the combine harvester during the harvesting operation, or setting one or more operating parameters of the combine harvester during the harvesting operation based on the computed quality factor.

[0046] A method of operating a combine harvester is also disclosed, the method comprising:

[0047] receiving one or more sensor values from sensors associated with the combine harvester during a harvesting operation;

[0048] receiving a target quality factor; and

[0049] determining one or more computed operating parameters using a machine learning algorithm that has been trained by any of the residue collectors disclosed herein, the received one or more sensor values, and the target quality factor and applying them to the combine harvester during the harvesting operation.

[0050] A computer program can be provided which, when running on a computer, causes the computer to configure any apparatus comprising the controller or device disclosed herein or to perform any of the methods disclosed herein. The computer program can be a software implementation and the computer can be considered to be any appropriate hardware, including by way of non-limiting example, a digital signal processor, a microcontroller, and a read only memory (ROM), an erasable programmable read only memory (EPROM) or an electrically erasable programmable read only memory (EEPROM) implementation. The software can be an assembler.

[0051] The computer program can be provided on a computer readable medium, which can be a physical computer readable medium such as a disc or a memory device, or can be implemented as a transitory signal. Such a transitory signal can be a network download, including an internet download. One or more non-transitory computer readable storage media can be provided storing computer executable instructions that, when executed by a computing system, cause the computing system to perform any of the methods disclosed herein. BRIEF DESCRIPTION OF DRAWINGS

[0052] One or more embodiments will now be described, by way of example only, with reference to the accompanying drawings in which:

[0053] Figure 1 A combine harvester is shown;

[0054] Figure 2 An example of a residue collector is shown;

[0055] Figure 3 An example embodiment of a rake system that can be provided as part of a residue separator is shown;

[0056] Figure 4 An example of a rotating shaft is shown for facilitating the transfer of crop residue in a residue collector from one conveyor to another;

[0057] Figure 5 Further details of a chopper are shown, which is an example of a component that can process crop residue before it is discharged from a combine harvester;

[0058] Figure 6 An example embodiment of a method of determining a quality factor of residue discharged by a combine harvester is shown;

[0059] Figure 7 A method of operating a combine harvester is described; and

[0060] Figure 8 Another method of operating a combine harvester is described. DETAILED DESCRIPTION

[0061] Figure 1 A combine harvester 10 is shown. The combine harvester 10 includes a feeder 12 upon which a header (not shown) can be mounted at a front end of the feeder. The header includes a cylindrical header reel (not shown) that rotates and directs crop material from a growing crop to fall on the header after / between being cut by a cutting bar of the header, allowing the crop material to be separated from the growing crop and directed toward a crop elevator 14 in the feeder 12 that feeds the cut crop material to other systems of the combine harvester 10.

[0062] The combine harvester 10 includes a threshing system 22 disposed downstream of the feeder 14. The threshing system 22 includes a rotor 24 that is rotatable for separating grain from a harvested crop from straw and other plant residue (hereinafter collectively referred to as straw 28 or straw residue).

[0063] The combine harvester 10 also includes a beater 26 that is rotatable for pushing a flow or stream of the straw 28 rearward along an aerial trajectory through a rear cavity 34 enclosed by structural pieces of the combine harvester 10. In addition, there can be a chopper to reduce an average straw length.

[0064] The combine harvester 10 includes a cleaning system 30 for receiving grain of the harvested crop from the threshing system 22 and removing grain chaff and any other remaining residue, including seed pods, chaff, etc., collectively referred to as grain chaff 32 or grain chaff residue, from the grain and directing a flow or stream of the grain chaff 32 rearward through a lower region of the rear cavity 34 toward a lower opening 38.

[0065] A horizontal residue spreader assembly 36 is located in the rear cavity 34. The spreader assembly 36 includes a crop residue dispensing system 40, for example including two side-by-side spreader disks or impellers, configured to rotate in opposite directions about a generally vertical axis of rotation. The crop residue dispensing system 40 can also include a pivotably supported deflector door at a rear end 60 of the rear cavity 34. The deflector door can be pivoted between a closed position and an open position to control the spread of the straw 28 and chaff 32, collectively referred to as crop residue or simply residue (also referred to as “material other than grain” or MOG), behind the combine 10.

[0066] The performance of a combine can be gauged by evaluating grain loss. For example, an operator can stop the combine and count the grain lost per unit of field area. A flat pan can also be placed on the ground to collect the lost grain and more accurately collect and count the lost grain.

[0067] However, it is not only the grain in the tank that is important, but also the residue left behind in the field by the combine. Residue is the portion or all of the harvest by the combine that does not go into the grain tank. As mentioned above, this includes straw and chaff. The chaff / straw can also include plant material other than the crop, such as weeds and weed seeds.

[0068] After passing through the threshing and / or separating (which can include rotor concaves, sieves, beater bars, etc.), an operator can choose to deposit a percentage of the residue on the field and can choose to process / treat it before depositing it on the field. For example, the operator can choose to deposit the straw directly in a swath or to treat the straw before depositing it. For example, the straw can be treated by chopping it before depositing and / or spreading it in the field. The operator can choose to deposit the chaff with the straw or separate from the straw (forming a tram lining) or can choose to treat the chaff before depositing it. Possible treatments include mechanical treatment (such as milling, grinding), chemical treatment (such as mixing with herbicide), radiation treatment, and / or heat treatment. One example of a mechanical treatment that will be discussed below is by a chaff pulverizer, which can also be referred to as a weed seed destroyer.

[0069] The operator can choose to treat the residual portion before depositing it on the field to improve the field for subsequent agricultural operations and to improve future harvests. Chopping the straw can increase the utilization of the straw as fertilizer by speeding up decomposition. Treating the chaff can destroy weed seeds in the chaff and reduce weed pressure.

[0070] The performance of the residue handling is very important for obtaining good results and reducing the requirements for subsequent operations (and thus their costs). Such subsequent operations can include the use of (additional) fertilizers, herbicides, irrigation, etc. Therefore, it is very important to assess the quality of the residue handling. This can be done in a similar way as for the assessment of grain loss, i.e. stopping the combine harvester and manually inspecting the residue deposited in the field, or taking samples from the field for laboratory assessment. However, this is very time-consuming and it is difficult to obtain reliable quantitative results.

[0071] It is also difficult to directly relate the residue quality measure to specific parameters of the residue handling system, because the interaction can be very complex. The harvesting situation is very diverse, like different crops, different crop conditions and different environmental conditions; which of these potential influences affects the measured quality of the crop residue can not be self-evident.

[0072] Furthermore, it is also not recommended to rely (only) on the expertise / experience of the combine harvester operator to determine and sufficiently control the quality of the residue. This is because many operators do not have much experience in this respect (yet) and they are under a lot of time pressure when harvesting.

[0073] The embodiments described below advantageously provide better and more efficient devices and methods for quantifying the quality of the residue deposited in the field. They can also collect more data in different places and situations. Furthermore, they can relate this data to the settings of the residue handling system of the combine harvester (or other parameters of the combine harvester) in an efficient and effective way, which is useful for subsequent harvesting operations.

[0074] Figure 2 An example of a residue collector 200 according to embodiments of the present disclosure is shown. The residue collector 200 can quantify the quality of a residue stream received from a combine harvester 210 of a field by determining a quality factor of the residue. In Figure 2 In the figure, only a part of the combine harvester 210 is shown to help describe the residue collector 200; the back of the combine harvester 210 is shown, with most components omitted (except for the chopper 201).

[0075] During a harvesting operation, the residue collector 200 receives residue directly from the combine harvester 210. The residue can be processed residue as it has been processed by the combine harvester 210 (e.g. by the chopper 201) before leaving the combine harvester 210. The residue stream can include one or more of: (un-chopped) straw, chopped straw, (unprocessed) grain chaff, and processed grain chaff. As will be discussed in detail below, the residue collector 200 can be used during a training harvesting operation in order to calibrate a machine learning algorithm for the combine harvester 210 in order to achieve improved performance during subsequent (non-training) harvesting operations without the residue collector 200.

[0076] In Figure 2 examples, the residue collector 200 is a trailer which, in use, is towed by the combine harvester 210. In other examples, the residue collector 200 can be self-propelled such that it is not mechanically coupled to the combine harvester 210 but can be operated such that it remains close enough to the combine harvester in order to directly receive residue. In Figure 2 examples, the residue collector 200 is mechanically connected to the combine harvester 210 by a tow hitch, and further comprises an engine 202. The engine 202 can be used to power the residue collector 200, which can or can not be used to power the residue collector 200. In this way, the residue collector 200 can be self-propelled and can not require power from the combine harvester 210. Drawing too much (or any) power from the combine harvester 210 can undesirably affect the measurements used to determine the quality factor.

[0077] The residue stream from the combine harvester 210 is conveyed from the combine harvester 210 to the residue collector 200 by a conveying system. In Figure 2 examples, the conveying system is a conveyor belt 203. The chopper 201 of the combine harvester 210 deposits the residue stream onto the conveyor belt 203, which conveys the residue to downstream components of the residue collector 200. It will be appreciated that other conveying systems can be used, such as a hopper / accumulator, a mobile platform, etc. In this example, the conveyor belt 203 is provided as part of the residue collector 200. The residue collector 200 further comprises a total weight sensor 204 which receives the residue from the conveyor belt 203. The total weight sensor 204 comprises its own conveying mechanism for conveying the residue to downstream components for processing.

[0078] In some examples, the residue collector 200 can be placed in a residue collection configuration or a residue bypass configuration. This can be achieved by using a conveying system which is movable between a residue collection position and a residue deposition position. In Figure 2This can be achieved using the conveyor belt 203 and / or the total weight sensor 204. Alternatively, the hook between the combine harvester 210 and the residue collector 200 can be extended to place the residue collector 200 in a residue bypass configuration so that residue does not reach the residue collector 200. Figure 2 In this configuration, a total weight sensor 204 is displayed at the residue collection location because it receives and conveys the residue to downstream components, allowing the determination of the residue's quality factor. At the residue deposition location, the residue collector 200 may deposit the residue in the field without determining its quality factor. That is, the conveying system can guide the residue around or avoid at least some components of the residue collector 200 that determine the residue's quality factor. As an example, this could involve moving the total weight sensor 204 so that it delivers the residue to the spreader 209 instead of the downstream component used to determine the quality factor. Figure 2 This involves moving the total weight sensor 204 to the right to receive residue from the conveyor belt 203 and depositing the residue vertically downwards onto the spreader 209. Specifically, the leftmost end of the total weight sensor 204 is spaced apart from the residue separator 205 (longitudinally) so that residue is not supplied from the total weight sensor 204 to the residue separator 205. Alternatively, the conveyor belt 203 may rotate about its vertical axis, causing residue to deposit on the side of the residue collector 200, or it may involve moving the conveying system so that it does not receive residue.

[0079] Advantageously, in some examples, the conveying system can be adapted to convey residual materials from different types and models of combine harvesters, so that the residue collector 200 can be used with a variety of combine harvesters.

[0080] exist Figure 2 In this example, the total weight sensor 204 determines the weight of the received residue. The total weight sensor 204 receives residue from the conveyor belt 203 and is implemented in this example as a weight measuring belt / conveyor. The total weight sensor 204 can periodically provide a total weight signal to a controller (not shown), where the total weight signal represents the weight of the residue received from the combine harvester 210 and will determine its quality factor. During training harvest operations, the combine harvester can supply residue to the residue collector 200, where the residue collector 200 is in a residue bypass configuration. As described above, this may involve placing the total weight sensor 204 at a residue deposition location such that it supplies residue to the spreader 209.

[0081] Once the training harvest operation is deemed to represent a stable operation, and the total weight sensor 204 receives a sustained amount of residue from the combine harvester, the combine harvester and residue collector can be stopped, and the transfer mechanism of the total weight sensor 204 can be paused so that it is loaded with residue.

[0082] When the combine harvester and residue collector 200 is stationary, the residue collector 200 can be placed in a residue collection configuration. As described above, in the residue collection configuration, the total weight sensor 204 can be positioned to provide all residue on the total weight sensor 204 to the residue separator 205. Figure 2 In this embodiment, this can involve positioning the total weight sensor 204 in a residue collection position so that it provides all residue on the total weight sensor 204 to the residue separator 205.

[0083] The residue separator 205 separates the residue being processed into a first portion and a second portion based on a characteristic of the residue being processed. In this way, the residue separator 205 divides the residue stream into at least two portions having different characteristics.

[0084] In this embodiment, the residue separator 205 is provided as a sieve 206 so that relatively smaller components of the residue can pass through the sieve, but relatively larger components cannot pass through the sieve. That is, the characteristic of the residue used to separate the received residue into the first portion and the second portion comprises the size of the residue. For example, short straws can be separated from long straws. The residue separator 205 can include various adjustable settings, such as: an adjustable shaking rpm (revolutions per minute) to shake the sieve 206 and assist smaller components of the residue to pass through the sieve 206; an adjustable shaking stroke also to shake the sieve 206 and assist smaller components of the residue to pass through the sieve 206; and an adjustable sieve opening to set the size of the residue components that can pass through the sieve 206 and the size that cannot pass through.

[0085] In this example, the components of the residue that pass through the sieve 206 (and thus are relatively smaller) can be considered the first portion of the received residue. The residue components that do not pass through the sieve 206 (and thus are relatively larger) can be considered the second portion of the received residue.

[0086] The residue collector 200 also includes a first portion weight sensor 207 that determines the weight of a first portion of the received residue (the portion that passes through the sieve 206). The residue collector 200 also includes a second portion weight sensor 208 that determines the weight of a second portion of the received residue (the portion that does not pass through the sieve 206). The first portion weight sensor 207 can provide a first portion weight signal to a controller (not shown), where the first portion weight represents the weight of the first portion of residue present on the total weight sensor 204 when the machine is stopped. The second portion weight sensor 208 can provide a second portion weight signal to the controller (not shown), where the second portion weight signal represents the weight of the second portion of residue present on the total weight sensor 204 when the machine is stopped.

[0087] In this example, it can be advantageous to stop the machine to weigh the first and second portions, as more accurate measurements can be made. In other examples, measurements can be made accurately enough while the machine is in motion. That is, once the training harvesting operation is deemed to represent a steady operation, it is not necessary to stop the combine harvester 210 and the residue collector 200. Rather, the residue collector 200 can be placed in the residue collection configuration while it is in motion, and residue can be separated and weighed without stopping the residue collector 200. In this way, the residue collector 200 can also work in a continuous mode, where it performs measurements on a continuous flow of residue from the combine harvester 210 without requiring the combine harvester 210 / residue collector 200 to stop in the field. In some examples, this can involve processing only a portion of the residue flow, particularly if the residue collector 200 has a capacity limit, which can require only a portion of the residue flow to be monitored. This can be achieved by alternating the residue collector 200 between the residue collection configuration and the residue bypass configuration, such that the residue collector 200 processes an appropriate amount of residue. Alternatively, the residue collector 200 can process only a portion of the residue while the residue collector 200 is in the residue collection configuration.

[0088] One or both of the first portion weight sensor 207 and the second portion weight sensor 208 can be implemented as a weight measuring belt / conveyor in the same way as the total weight sensor 204. Alternatively, any of the weight sensors described herein can be implemented by accumulating residue in a weigh bin for a predetermined period of time or in any other way known in the art.

[0089] In Figure 2 examples, once the first portion is weighed, it is provided to the spreader 209 for deposition on the field. Optionally, the second portion can also be provided to the spreader 209 after weighing.

[0090] The residue collector 200 also includes a controller (not shown) for processing one or more of the total weight signal, the first portion weight signal, and the second portion weight signal. The controller may be provided locally to the residue collector 200 or remotely to the residue collector 200. For example, the functionality of the controller may be provided in the cloud, and the weight signals measured by the weight sensors 204, 207, 208 may be transmitted to the remote controller for processing.

[0091] The controller determines the quality factor of the treated residue based on the weight of the determined first portion relative to the weight of the second portion. In this way, the controller can determine the quality factor during the harvesting operation because the signal required to determine the quality factor is available almost instantaneously during the harvesting operation. It should be understood that calculating the quality factor in this way can be achieved by processing any two of the following three signals: the total weight signal, the first portion weight signal, and the second portion weight signal. Therefore, in some examples, only two of the following three weight sensors may be needed: the total weight sensor 204, the first portion weight sensor 207, and the second portion weight sensor 208. That is, the residue collector 200 may include one or more weight sensors (because weight sensors can potentially be reused, thus providing more than one weight signal) for directly or indirectly determining the weights of the first and second portions. An example of indirect determination of the weight of the first portion can be achieved by subtracting the weight signal of the second portion (as provided by the second portion weight sensor) from the total weight signal (provided by the total weight sensor).

[0092] exist Figure 2 In the example, the quality factor is calculated by dividing the first partial weight signal by the total weight signal. In this way, it represents the proportion of residue received through sieve 206, and is therefore classified as relatively small. It will be appreciated that the same general information can be determined by performing appropriate calculations on any two weight signals.

[0093] Furthermore, the controller can provide the combine harvester operator with an indication of the determined quality factor during harvesting operations. For example, this can be done by displaying the determined quality factor on a screen in the combine harvester cab. Additionally, the controller can periodically update the quality factor when receiving and processing updated weight signals.

[0094] In some examples, the residue collector 200 can include a residue selection component that selectively transfers only a portion of the received residue stream to the residue separator 205. The selection can be based on a lateral position of the received residue (e.g., selecting residue only on a portion of the width of the received residue, such as only the left portion of the residue coming out of the combine harvester). Additionally or alternatively, the selection can be based on time (e.g., transferring residue to the residue separator 205 every half of the 5 minutes). Such selection can be beneficial as it can enable the combine harvester 210 to produce residue at a rate that matches the capacity of the residue collector 200. The portion of the residue that is not selected for processing by the residue separator 205 can be directly deposited on the field.

[0095] In yet another example, the residue collector 200 can include a residue scanner (not shown) that scans the residue received from the combine harvester 210 to determine the volume of residue provided per unit of time. In this way, the flow rate of residue provided by the combine harvester 210 can be determined. As non-limiting examples, such a residue scanner can be implemented with an ultrasonic sensing system, a lidar sensing system, or a radar sensing system. The controller can then be used to determine the residue volume that is provided to automatically control one or more operating parameters of the combine harvester 210 and / or the residue collector 200. For example, the operating parameters of the combine harvester 210 can be set so that the volume of residue provided by the combine harvester 210 better matches the volume of residue that the residue collector 200 can process without overloading or underloading it. In addition, the operating parameters of the residue collector 200 can be set so as to selectively switch between the residue collection configuration and the residue bypass configuration so that the volume of residue provided to the residue separator 205 better matches the volume of residue that the residue separator 205 can process. Further, the operating parameters of the residue collector 200 can be set so as to control the residue selection component that selectively transfers only a portion of the received residue stream to the residue separator 205.

[0096] Figure 2The residue collector 200 can provide very significant advantages. By combining the collection of residue with the determination of the quality factor of the residue on a single device (the residue collector 200), the training harvesting operation can be completed quickly and efficiently. This can greatly reduce the time and cost of testing. Furthermore, as the testing can be performed more quickly, the variation in the test results is also less, as there are no significant changes in the crop and field conditions over the time required to perform the training harvesting operation. This is particularly beneficial when performing the training harvesting operation to establish a data set that will be used to improve the performance of subsequent (non-training) harvesting operations. This will be discussed in more detail below in relation to obtaining training data for a machine learning algorithm that will be implemented as part of the subsequent (non-training) harvesting operations.

[0097] In examples where the controller is used to train a machine learning algorithm, it can also receive one or more sensor values from sensors associated with the combine harvester 210 during the training harvesting operation. In some examples, the one or more sensor values are from sensors associated with residue handling components of the combine harvester 210 that are used to process / treat the residue before it is discharged from the combine harvester 210. As the sensor values are recorded at a time corresponding to substantially the same components of the residue that will have the quality factor determined, the sensor values can be associated with the determined quality factor. For example, the controller can apply a time offset to the one or more sensor values to account for the time required for the residue stream to travel from i) the point at which the sensor values are recorded in / at the combine harvester to ii) the point or points at which the measurement is made in the residue collector 200 to determine the quality factor. In alternative examples, the controller can not apply a time offset. This is based on the changes in the residue being considered relatively slow compared to the time required for the residue to pass through the combine harvester 210 and the residue collector 200.

[0098] The controller can then train the machine learning algorithm based on the one or more sensor values and the determined quality factor. For example, the machine learning algorithm can be a classification algorithm that uses the one or more sensor values as input and uses the determined quality factor as ground truth data for training. Then, during subsequent harvesting operations, the trained machine learning algorithm can be used to process received sensor values and determine a quality factor without the need to use the residue collector 200. In this way, the residue collector 200 can be used to obtain training data for calibrating the machine learning algorithm.

[0099] The types of machine learning algorithms that are suitable for providing this functionality are well known in the art and can include, as non-limiting examples, neural networks (NNs), convolutional neural networks (CNNs), and support vector machines (SVMs).

[0100] Examples of sensors that can be used to provide sensor values include optical sensors, cameras, acoustic sensors, temperature sensors, speed or rpm sensors, pressure sensors, humidity sensors, loss sensors, knock sensors, and mechanical sensors. In fact, any sensor can be used to sense data that has an impact on the quality of the residue, as defined by the quality factor.

[0101] In more detail, the one or more sensors can include a video camera (or other optical sensor) that records images of the crop or residue and / or the crop flow path or residue flow path in the combine harvester 210. The sensor values can thus include one or a series of pictures taken of the residue flow within the combine harvester, for example pictures from a so-called threshing video camera that records images of the straw being threshed in the combine harvester 210.

[0102] If the one or more sensors include one or more acoustic sensors, these sensors can be positioned to record the sound as the crop / residue passes along the flow path in the combine harvester 210. One or more of the size, shape, and material of the crop / residue will affect the sound produced as the crop / residue passes through the combine harvester 210, and thus the sound recorded can be indicative of the quality factor of the residue.

[0103] If the one or more sensors include one or more mechanical sensors, these sensors can be positioned to record the impact / vibration caused by the crop / residue as it travels along the flow path in the combine harvester 210. One or more of the size, shape, and material of the crop / residue will cause different mechanical sensor values as the crop / residue passes through the combine harvester 210, and thus the mechanical sensor values recorded can also be indicative of the quality factor of the residue.

[0104] Loss sensors are known in the art as being able to sense the amount of grain that is not successfully recovered by the combine harvester 210 and thus lost. Such loss sensors can include optical sensors, acoustic sensors, and knock sensors. Knock sensors can be excited by the grain hitting them.

[0105] Optionally, known signal processing techniques can be applied to the sensor values before the controller uses the sensor values to train the machine learning algorithm. Such techniques include calibration, noise reduction, low pass filtering, etc. In the case of one or more images, known image processing techniques can be applied, for example calibration, noise reduction, thresholding, edge and shape recognition, classification, and counting.

[0106] Advantageously, the calculation (and optional display) of the residue quality factor during the training harvesting operation enables the operation of the combine harvester 210 to be adjusted during the training harvesting operation in order to create a varied data set for training the machine learning algorithm. For example, one or more operating parameters of the combine harvester 210 can be adjusted, thereby obtaining different measurements of sensor values and / or determining different quality factors. In this way, a good set of training data for the machine learning algorithm can be obtained quickly and efficiently. This can be much better than the example where the quality factor of the residue is calculated after the training harvesting operation is complete, in which case it can be necessary to initiate another training harvesting operation in an attempt to and complete the data set. Even then, the subsequent training harvesting operation can not provide all of the required information as it is difficult for the operator to know the quality of the residue being produced. Furthermore, when determining the quality factor from residue that is separately collected and processed, it is more difficult to accurately associate any sensor values with the determined quality factor.

[0107] Returning to Figure 1 we will now describe how the combine harvester 10 uses the machine learning algorithm trained by the above method to provide residue with a desired quality factor. The skilled person will appreciate that the desired value of a particular quality factor of the residue will depend on the details of the harvesting operation. For example, for a harvesting operation in a tropical country (such as Brazil), it can be desirable to leave relatively large amounts of residue on the ground after the harvesting operation to protect the topsoil from the effects of tropical rainfall. Whereas in a relatively cold country (such as Canada), it can be desirable to leave relatively little residue on the soil so that it is more easily decomposed. As mentioned above, this potential range of values for the quality factor that is considered to be desirable is one reason why it can be advantageous to use a varied data set when training the machine learning algorithm.

[0108] The combine harvester 10 for the (non-training) harvesting operation comprises a controller (not shown). The controller receives one or more sensor values from sensors associated with the combine harvester during the harvesting operation. The sensors and the sensor values can be any corresponding sensors and sensor values discussed above in relation to training the machine learning algorithm.

[0109] The controller can then use the machine learning algorithm that has been trained by the above method to determine a calculated quality factor using the received one or more sensor values as input. The quality factor can be calculated almost immediately as the combine harvester 10 produces residue.

[0110] In this example, the controller also causes the computed quality factor to be presented to an operator of the combine harvester 10 during the harvesting operation. In this way, a real-time feed of the instantaneous value of the computed quality factor can be presented to the operator, so that the operator can continue to use the combine harvester 10 in an improved manner. For example, if the operator can see that the computed quality factor is too low or too high (when compared to the quality factor they expect), then they can adjust the operating parameters of the combine harvester 10 to bring the computed quality factor closer to the desired value. (Examples of how the operating parameters of the combine harvester 10 can be adjusted by changing the performance of the chopper are described below.) The controller can cause the computed quality factor to be presented to the operator in a visual (e.g. by using a display in the cab), audible (e.g. by a speaker), or any other manner known in the art.

[0111] Additionally or alternatively, the controller can set one or more operating parameters of the combine harvester 10 during the harvesting operation based on the computed quality factor. This can be achieved by applying a control loop that adjusts the one or more operating parameters to bring the computed quality factor closer to a target quality factor that has been provided to the controller by an operator of the combine harvester 10. Such control loops are known in the art and can involve adjusting the one or more operating parameters until the computed quality factor is considered close enough to the target quality factor or until further iterations of the control loop no longer result in a sufficiently large improvement in the computed quality factor. In this way, the controller can set one or more operating parameters of the combine harvester during the harvesting operation based on the computed quality factor and also based on a target quality factor (or a range of target quality factors if a range has been provided).

[0112] In yet another example, the controller associated with the residue collector can train the machine learning algorithm differently during the training harvesting operation. In such an example, in addition to the determined quality factor, the controller receives one or more sensor values and one or more operating parameters of the combine harvester during the training harvesting operation. In the same way as described above, the one or more sensor values are from one or more of the various sensors associated with the combine harvester. The one or more operating parameters of the combine harvester can be associated with the determined quality factor as they are recorded at a time corresponding to substantially the same component of the residue from which the quality factor has been determined. In the same way as the sensor values described above, the controller can or can not apply a time offset to the one or more operating parameters to account for the time taken for the residue stream to travel from: i) the point in / at the combine harvester at which the operating parameter took effect; ii) the one or more points in the residue collector from which measurements were taken to determine the quality factor.

[0113] The operating parameters of the combine harvester can include operating parameters of residue handling components of the combine harvester, such as operating parameters of a chopper that chops the residue before it leaves the combine harvester. Reference is made below to Figure 5 Further details of such examples are provided.

[0114] In this way, a dataset can be created that combines data of the working state of the combine harvester and data measured in the combine harvester with data measured by the residue collector (optionally over time).

[0115] The controller associated with the residue collector can then train a machine learning algorithm based on the one or more sensor values, the one or more operating parameters, and the determined quality factor. For example, the machine learning algorithm can be a classification algorithm that uses the one or more sensor values and the determined quality factor as input, and uses the one or more operating parameters as ground truth data for training. As will be discussed below, a machine training algorithm trained in this way is one way of providing autonomous control of at least certain aspects of the harvesting operation.

[0116] Referring to Figure 1 We will now describe how the combine harvester 10 uses a machine learning algorithm trained by the above method to provide residue having a desired quality factor.

[0117] In this example, the controller of the combine harvester 10 for a (non-training) harvesting operation receives one or more sensor values from sensors associated with the combine harvester during the harvesting operation; and a target quality factor (e.g. provided by an operator of the combine harvester 10 as described above). Again, the sensors and the sensor values can be any respective sensors and sensor values discussed above in relation to training the machine learning algorithm.

[0118] The controller can then use the machine learning algorithm that has been trained by the above method to take the received one or more sensor values and the target quality factor as input to determine and apply one or more calculated operating parameters for the combine harvester during the harvesting operation. In this way, at least part of the combine harvester can be automatically controlled such that it provides residue having the desired quality factor.

[0119] From the above description, it will be appreciated that by using field tests with residue collectors having a variety of combine harvester and residue handling settings, a sufficiently broad and large dataset can be created to enable a known artificial intelligence (AI) algorithm to be trained such that for a subsequent harvesting operation the combine harvester and residue handling settings can be controlled to obtain a desired residue quality over a range of conditions.

[0120] In another example, any of the machine learning algorithms described herein can be trained with additional types of data, including crop data, field data, and / or environmental data, as input acquired during training harvesting operations. Similarly, when the trained machine learning algorithms are used in subsequent (non-training) harvesting operations, the same additional types of data can be acquired and used as input.

[0121] Crop data is an indicator of one or more characteristics of the field crop being harvested by the combine harvester. The crop data can be an indicator of one or more of:

[0122] Crop height;

[0123] Crop density;

[0124] Crop moisture;

[0125] Crop feed rate (i.e. how much crop is entering the machine); and

[0126] Crop type.

[0127] Field data is an indicator of field conditions of the field in which the combine harvester is operating. The field data can be an indicator of one or more of:

[0128] Moisture content of the soil;

[0129] Soil temperature; and

[0130] Soil type.

[0131] Environmental data is an indicator of environmental conditions in which the combine harvester is operating. The environmental data is an indicator of one or more of:

[0132] Air humidity; and

[0133] Air temperature.

[0134] Various types of sensors suitable for providing the above types of data are well known in the art.

[0135] As described above, the residue collector described herein includes a residue separator for separating the residue into a first portion and a second portion based on a characteristic of the residue. In Figure 2In one example, the residue separator is implemented as a sieve. In another example, the residue separator can include a fan / blower that provides a gas flow to the residue. Such a gas flow can cause a first portion of the residue to be transported along a first flow path (e.g. into a first bin / hopper), and can also cause a second portion of the residue to be transported along a second flow path (e.g. into a second bin / hopper). It will be appreciated that the weight and / or size of the various components of the residue will determine how much influence the gas flow will have on those components, and thus whether the gas flow will cause those components to be transported along the first path or the second path. Such a residue separator can separate the residue into the first portion and the second portion based on the weight and / or size of the residue.

[0136] It will also be appreciated that the residue separator can separate the received residue into more than two portions (e.g. by passing the residue through a plurality of sieves in turn, each sieve having smaller holes than the previous one), and can determine a respective quality factor accordingly. For such an example, the quality factor can be composed of a plurality of different sub-components, each sub-component representing a different portion. As one numerical example: if the first portion comprises 50% of the total weight of the received residue, the second portion comprises 30% of the total weight of the received residue, and the third portion comprises 20% of the total weight of the received residue, then the quality factor can be represented as 50.30.20. That is, the quality factor of the residue being processed can be determined based on the relative weights of the three or more portions.

[0137] In another example, the residue separator for separating the residue can include a camera that records an image of the residue. The residue separator can also include an image processing algorithm that extracts one or more features from the image. These extracted features are examples of characteristics of the residue that can be used to separate the residue into the first portion and the second portion. Non-limiting examples of features that can be extracted from such an image include:

[0138] • the size of the components in the residue, for example:

[0139] the distinction between short and long straw (e.g. in relation to one or more threshold values);

[0140] the distinction between large and small particles of processed grain chaff (again, for example, in relation to one or more threshold values). Such an example is particularly useful in examples that include a grain chaff mill for processing the residue before it is expelled by the combine harvester. The grain chaff mill can also be referred to as a weed seed destructor, as it is used to destroy weed seeds before they are returned to the field as residue, thereby reducing their germination ability. Thus, being able to distinguish between the size of the particles in the residue (and the extent to which the weed seeds are destroyed) can determine a useful quality factor;

[0141] • colour;

[0142] For example, the amount of one or more specific colors in the residue image can be indicative of one or more characteristics of the residue. For example: the amount of green color in the image can be indicative of the number of weeds; the amount of brown color in the image can be indicative of the amount of grain; the ratio of green pixels to brown pixels can be indicative of the ratio of weeds to grain in the residue;

[0143] • the number of predetermined shapes in the image, e.g. shapes representative of damaged seeds or grains, spliced straw, or more generally straw conditions: e.g. bending, buckling, indentation, crushing, etc.

[0144] In yet another example, the residue separator can separate the residue into a first portion and a second portion based on the moisture content of the residue. The residue separator can comprise a moisture sensor that determines the moisture of the residue, and can comprise a controller that compares the sensed moisture value to one or more threshold values to determine whether the relevant residue should be part of the first portion or the second portion.

[0145] For at least some of the above characteristics used to determine whether the residue should be in the first portion or the second portion (or any other portion, if any), the residue separator can comprise a separation mechanism for separating the residue into different portions. In one example embodiment, the residue can be conveyed on a conveyor belt towards a movable panel that can divert the residue to a first bin / hopper or a second bin / hopper. The position of the movable panel is set based on the determined characteristics of the residue, such that the residue can be selectively directed to one of the bins / hoppers and thus separated into different portions.

[0146] Reference Figure 1 As mentioned above, in some examples, the combine harvester 10 can produce a grain chaff residue 32. Such grain chaff residue 32 can be treated within the combine harvester 10 by mechanical processing, e.g. by milling, to destroy weed seeds. For such examples, the residue collector can separate the treated grain chaff residue into two portions, e.g. based on the grain size / particle size. For example, separate the pulverized treated grain chaff into a portion having larger particles, such as weed seeds and / or broken weed seed particles, and a portion having smaller particles, such as pulverized grain chaff and dust. By quantifying the weed seeds (particles) having a size above a certain threshold, a quality indicator of the grain chaff treatment, in particular the mechanical treatment, e.g. milling or grinding, can be assessed. The residue collector can have a residue separator comprising a rake system, typically adapted for handling straw-like residue.

[0147] Figure 3 An exemplary embodiment of a rake system 310, which can be provided as part of a residue separator according to the present disclosure, is shown. The rake system 310 can be located at the top of a sieveFigure 3 The rake system 310 helps to spread the residue / stalks across the screen and helps to prevent clogging in the screen.

[0148] In Figure 3 Examples, the rake system 310 includes a rake 311 having a plurality of tines. The rake 311 is operable to move parallel to the screen in superimposed up / down motion (e.g., in a wave-like motion). This is achieved in Figure 3 Examples by a wavy track 312 on which the rake 311 is mounted. As the rake 311 moves along the screen in a longitudinal direction, the wavy track 312 causes the rake 311 to also move up and down in a direction perpendicular to the plane of the screen. The wave-like motion can be used for forward and backward motion (such that the rake 311 follows the wavy track as it moves on the screen in both longitudinal directions). Alternatively, the rake 311 can move in a wave-like motion in only one longitudinal direction, while in the other longitudinal direction the rake 311 returns without raking the stalks, e.g., the rake 311 returns above the stalks. This can be achieved by the rake 311 following the wavy track 312 in one longitudinal direction, and following a different track (further spaced apart from the screen) in the other longitudinal direction. As another example, the rake system 310 can change the wave shape (in terms of phase and / or amplitude) followed by the rake 311 by any suitable mechanism. This can further help to separate the residue into different parts.

[0149] As mentioned above, as Figure 2 shown, the residue collector 200 can have an additional on-board power source, e.g., an internal combustion engine 202, an electric motor with a battery and / or a fuel cell. Such a power source can be used to drive the residue separator. This can be in addition to or instead of providing propulsion to the residue collector 200. This can be advantageous because the power of the combine harvester is not required to drive the residue separator, which can undesirably affect the measurements.

[0150] Referring again to Figure 2 , to facilitate the transition from one belt to another (e.g., from the conveyor belt 203 to the weight measurement belt of the total weight sensor 204), the residue collector 200 can include a rotating shaft with tines that engage the residue as it moves between the two belts. Figure 4 An example of such a rotating shaft 413 is shown. This can improve the uniformity of the residue feed and help to ensure a continuous flow of residue.

[0151] Figure 5Further details of a chopper 520, which is an example of a component that can process residue as it is being discharged from a combine harvester as part of a harvesting operation, are shown. The chopper 520 includes a plurality of rotating knives / blades 521 that rotate about an axis and periodically come into close contact with stationary / opposed knives 522. As the rotating knives 521 and stationary knives 522 pass by each other, any residue between the rotating knives 521 and stationary knives 522 is cut.

[0152] As described above, one or more operating parameters of a combine harvester can be adjusted to change a quality factor of residue discharged by the combine harvester. One or more operating parameters of the chopper 520 are examples of operating parameters of a combine harvester that can be set to adjust the quality factor. In particular, if the quality factor is or represents the size of components in the residue. Such operating parameters can include the speed of the chopper (i.e., the speed at which the rotating knives 521 rotate) and the position of the stationary knives 522. As shown, the stationary knives 522 can be inserted or retracted to adjust the amount that they overlap the rotating knives 521 as the rotating knives 521 turn. In addition, the angle of the stationary knives 522 can be adjusted to change the performance of the chopper 520. Figure 5

[0153] Other examples of operating parameters of a combine harvester that can be adjusted to change the quality factor of residue include:

[0154] • the speed of the residue through the chopper 520. In one example, this can be adjusted by setting the position of a chopper bar. The chopper bar can be inserted into or retracted from the flow path of the residue through the chopper 520 to selectively impede the flow of residue. By slowing the flow of residue through the chopper 520, it can be cut more times and thus can be smaller than it would be if it were less impeded as it passed through the chopper 520.

[0155] • the aggressiveness of the pre-chopper threshing system. For example, the rotor speed and / or the operating characteristics of the concaves can be adjusted. As another example, the angle of the rotor blades can be adjusted to change the aggressiveness of the threshing system. Generally, the longer the residue stays in the threshing system, the more likely it is to be damaged.

[0156] • the feed rate of the residue into the chopper 520, which in one example can be adjusted by changing the speed of the combine harvester.

[0157] Figure 6 An example embodiment of a method of determining a quality factor of residue discharged by a combine harvester is shown. The method can be performed by any of the residue collectors described herein.

[0158] ​At step 650, the method receives the residue directly from the combine harvester during the training harvest operation. The residue is received directly such that subsequent processing steps can also be performed during the training harvest operation. That is, without the combine harvester, the residue collector or the residue itself would have to be moved away from the location where the residue is collected.

[0159] At step 651, the method separates the residue into a first portion and a second portion based on a characteristic of the residue. At step 652, the method determines the weight of the first portion and the second portion directly or indirectly. Numerous examples of how these steps can be performed are discussed above.

[0160] At step 653, the method determines a quality factor for the residue based on the determined weight of the first portion relative to the weight of the second portion. In this way, the quality factor can be determined and updated in real-time during the training harvest operation

[0161] Figure 7 A method of operating a combine harvester is illustrated. The method is computer implemented and can be performed by any of the controllers described herein.

[0162] At step 755, the method receives one or more sensor values from a sensor associated with the combine harvester during the harvest operation. Various examples of such sensors are described above and can include any sensor that measures a value that can impact the quality factor of the residue being discharged by the combine harvester.

[0163] At step 756, the method uses a machine learning algorithm that has been trained by a dataset collected at least in part by the residue collector as described above. More specifically, a machine learning algorithm trained using one or more sensor values as input and the determined quality factor as ground truth output data. At step 756, the method applies the received one or more sensor values as input to the trained machine learning algorithm to determine a calculated quality factor.

[0164] After step 756, the method can perform step 757 and / or step 758. At step 757, the method presents the calculated quality factor to an operator of the combine harvester during the harvest operation. This can enable the operator to manually adjust the operation of the combine harvester based on the calculated quality factor that would otherwise be difficult or impossible for the operator to identify during the harvest operation. At step 758, the method sets one or more operational parameters of the combine harvester during the harvest operation based on the calculated quality factor, thereby providing at least one autonomous control element.

[0165] Furthermore, the calculated quality factor can be stored, preferably in combination with location and / or date and time. A map of the calculated quality factor can be created for use as input for subsequent agricultural operations on that location.

[0166] Figure 8 Another method of operating a combine harvester is illustrated. As with the method of Figure 7 the method is computer-implemented, and can be executed by any of the controllers described herein. Figure 8

[0167] At step 860, the method receives one or more sensor values from a sensor associated with the combine harvester during a harvesting operation. At step 861, the method receives a target quality factor. The target quality factor represents a desired value for a residue quality factor, and can be set by an operator of the combine harvester. In some examples, the target quality factor can be implemented as a range of quality factor values.

[0168] At step 862, the method uses a machine learning algorithm trained by the residue collector as described above. More specifically, a machine learning algorithm that has been trained with one or more sensor values and determined quality factors as input and one or more operating parameters as ground truth output data. At step 862, the method applies the received one or more sensor values and the received target quality factor as input to the trained machine learning algorithm to determine one or more calculated operating parameters during the harvesting operation, which are applied to the combine harvester. In this way, at least one autonomous control element can be provided.

[0169] Examples disclosed herein can relate to a method of quantifying the quality of a field residue stream of a combine harvester, the method comprising:

[0170] - during a harvesting trial, a residue collector following closely behind a combine harvester over a field;

[0171] - the residue stream of the (at least part of) the combine harvester is diverted to the residue collector;

[0172] - the weight of the residue stream is determined;

[0173] - the residue stream is divided into at least two parts having different characteristics;

[0174] - the weight of at least one of the separated residue stream parts is measured;

[0175] - optionally further characteristics of the residue stream are measured; and

[0176] - the residue stream parts are deposited onto the field.​

Claims

1. A residue collector (200) capable of receiving residue from a combine harvester (210) during training harvest operations, wherein the residue collector (200) comprises: A residue separator (205) is used to separate the residue to be treated into a first part and a second part based on the characteristics of the residue being treated; One or more weight sensors (204, 207, 208) are used to directly or indirectly determine the weight of the first part and the second part; and The controller is configured to determine the quality factor of the treated residue based on the weight of the determined first portion relative to the weight of the second portion.

2. The residue collector of claim 1, wherein the controller is configured to determine the quality factor during the training harvest operation.

3. The residue collector according to claim 2, wherein, The controller is also configured to provide the operator of the combine harvester (210) with an indication of the determined quality factor during the training harvest operation.

4. The residue collector according to any one of the preceding claims, wherein: The residue separator is used to separate the residue to be treated into three or more parts based on one or more characteristics of the residue being treated; The one or more weight sensors are used to determine the weight of each part directly or indirectly; and The controller is configured to determine the quality factor of the treated residue based on the relative weight of the three or more portions.

5. The residue collector according to any one of claims 1 to 3, wherein the characteristics of the residue for separating the residue into a first portion and a second portion include one or more of the following: The size of the components in the residue; The shape of the components in the residue; The density of the components in the residue; The moisture content of the residue; and The color of the residue.

6. The residue collector according to any one of claims 1 to 3, wherein the residue collector (200) comprises a trailer towed by the combine harvester (210) in use.

7. The residue collector according to any one of claims 1 to 3, wherein the residue collector has a residue collection configuration and a residue bypass configuration, wherein: In the residue collection configuration, the residue collector is configured to transfer the residue to a component of the residue collector for determining the quality factor of the residue; and In a residue bypass configuration, the residue collector is configured to allow residue to bypass or avoid components of the residue collector used to determine the quality factor of the residue.

8. The residue collector according to any one of claims 1 to 3, further comprising a residue selection component for selectively transferring only a portion of the received residue to a residue separator.

9. The residue collector according to any one of claims 1 to 3, wherein, The controller is also configured to: Receive one or more sensor values ​​from sensors associated with the combine harvester; and The one or more sensor values ​​and associated determined quality factors are stored as training data for machine learning algorithms.

10. The residue collector according to claim 9, wherein, The controller is also configured to: A machine learning algorithm is trained based on the training data, and the trained machine learning algorithm is used for subsequent use during the harvest operation.

11. A controller for a combine harvester (210), the controller being configured to: During harvesting operations, one or more sensor values ​​are received from sensors associated with the combine harvester; The calculated quality factor is determined using a machine learning algorithm trained by the residue collector (200) of claim 10 and the received values ​​from one or more sensors; and The calculated quality factor is presented to the operator of the combine harvester (210) during harvesting operations, or one or more operating parameters of the combine harvester (210) are set during harvesting operations based on the calculated quality factor.

12. The controller according to claim 11, wherein, The controller is also configured to: One or more operating parameters of the combine harvester (210) are set during harvesting operations based on the calculated quality factor and also based on the target quality factor.

13. A controller for a combine harvester (210), the controller being configured to: During training harvest operations, one or more sensor values ​​are received from sensors associated with the combine harvester; Receive target quality factor; and One or more calculated operating parameters are determined using a machine learning algorithm trained by the residue collector (200) of claim 10, one or more received sensor values, and the target quality factor, and applied to the combine harvester (210) during training harvest operations.

14. The controller according to any one of claims 11 to 13, wherein, The one or more operating parameters include one or more operating parameters of the residue treatment component of the combine harvester (210).

15. A method of operating a combine harvester (210), the method comprising: During harvesting operations, one or more sensor values ​​are received from sensors associated with the combine harvester; The calculated quality factor is determined using a machine learning algorithm trained by the residue collector (200) of claim 10 and the received values ​​of one or more sensors; and The calculated quality factor is presented to the operator of the combine harvester (210) during harvesting operations, or one or more operating parameters of the combine harvester (210) are set during harvesting operations based on the calculated quality factor.

Images