AGRICULTURAL MACHINE WITH NIR SENSOR AND DATA PROCESSING SYSTEM

Description

[0001] The present invention relates to an agricultural machine, in particular a harvesting machine such as a combine harvester or a forage harvester, with a NIR sensor, and to a data processing system comprising one or more of these machines.

[0002] NIR sensors measure the amount of light transmitted or reflected by a sample in the near-infrared range. Organic substances generally have structurally rich absorption and reflection spectra in this spectral range, resulting from the excitation of vibrational oscillations of bonds between atoms in these substances. However, there are no clearly distinguishable lines in this spectral range that could be assigned to a specific chemical compound; rather, virtually every organic compound contains C–C bonds and C–H bonds, whose lines may be slightly shifted from one substance to another as a result of interactions with neighboring atoms, but at the same time, they are also so broadened that the shifts are difficult to identify.The number of degrees of freedom by which the NIR spectra of different compounds can differ from one another is generally much smaller than the number of compounds that can be present in a sample. In order to derive useful information about the composition of a sample from an NIR spectrum, extensive assumptions about the nature of the sample must be made in advance, and the correctness of these assumptions determines the significance of the information obtained from the spectrum. In practice, this means that in order to be able to provide information about value-determining components of a grain sample by analyzing an NIR spectrum, the type and variety of the grain, as well as possibly other influencing factors, must be known in order to select a calibration model suitable for the sample. The optimization of these calibration models is the subject of intensive development.Results obtained using different calibration models are not easily comparable.

[0003] An NIR sensor head for use in an agricultural machine is known from DE 10 2004 048 103 B4. The sensor head comprises an illumination source for illuminating the material to be examined, a spectrometer arrangement as an NIR sensor, a processor for acquiring and processing the measured values, and an interface to a bus system of the agricultural machine, which can optionally output measurement results or raw data for further processing to the machine's bus system. Parameters such as moisture, protein, starch, or oil content, cutting length, fiber condition, or temperature can be derived from the acquired data—each based on a suitable calibration. The document does not provide any information on how the processor further processes the raw data into measured values ​​or how the raw data is processed by a receiver connected to the bus.Therefore, the problem here is that measurement results are of dubious quality if they are obtained using an unsuitable or outdated calibration model.

[0004] A generic agricultural machine is known from EP 3 366 104 A1.

[0005] However, it is desirable to be able to compare evaluation results over several years in order to be able to assess the effectiveness of maintenance measures such as fertilization, pest control, etc. and to optimize their scope.

[0006] An object of the invention is to provide an agricultural machine with an NIR sensor that enables such comparability.

[0007] The problem is solved by an agricultural machine having an NIR sensor configured to record NIR spectra of plant material and output them as raw data, an evaluation unit for deriving at least one parameter of the plant material in real time from the raw data, and an interface for data communication with at least one data processing unit outside the machine, wherein the interface is configured to transmit the raw data comprising the NIR spectra to the data processing unit. In this way, data obtained from a given field area and crop can be collected over several years at the data processing unit and evaluated if necessary using the same calibration model, and the effectiveness of maintenance measures carried out in different years can be assessed by comparing the evaluation results.Furthermore, the raw data can be used at the data processing unit to optimize calibration models used on the machine that provided them or on other machines, or to create new calibration models if necessary.

[0008] To ensure that these calibration models also benefit the work of the machine that supplies the raw data required for their creation, the evaluation unit should be configured to use an interchangeable calibration model to derive at least one parameter. This makes it possible, for example, to create calibration models specifically for the machine's NIR sensor, taking its individual properties into account, in order to optimize its performance. On the other hand, based on large volumes of raw data originating from different machines, commonalities in the measurement characteristics of their various sensors can be reliably identified and used as a uniform basis for the evaluation for all sensors.

[0009] Conveniently, the same interface used to transfer the raw data from the machine to the data processing unit can also be used to transfer new calibration models from the data processing unit to the machine's evaluation unit in the opposite direction.

[0010] The interface can be configured to transmit, in addition to the raw data, values ​​of at least one parameter derived from the evaluation unit. On the one hand, this ensures compatibility with systems already widely used in agriculture, which only provide evaluation results from a machine-supported evaluation unit to a stationary data processing unit. On the other hand, by using the same transmission path for the evaluation results and the raw data, the workload for a machine operator can be minimized. This is particularly important if the machine operator himself does not derive any direct benefit from the provision of the raw data.

[0011] The agricultural machine typically has a processing unit for processing the plant material, such as a threshing or chopping unit. The data obtained in real time by the evaluation unit can be used to adapt at least one operating parameter of the processing unit to the nature of the crop—which may vary within a contiguous field depending on soil quality or other circumstances.

[0012] Such a parameter - e.g. influencing the energy expenditure during threshing or chopping - can be, for example: the water content of the plant material or a parameter representative of the degree of maturity of the plant material.

[0013] The machine preferably comprises a timer and / or a position-determining device, such as a satellite navigation device. The evaluation unit can then be configured to add the times and / or locations of their generation to the NIR spectra. Based on these times, the data processing unit outside the machine, if the machine's trajectory is known to it, e.g., from previous route planning, can determine the location corresponding to a received spectrum and store it with the spectrum. Knowing the location where a spectrum was generated enables location-dependent optimization of maintenance measures.

[0014] The interface can be configured to transmit the raw data in an integrity-protected format, e.g., encrypted or linked to verification information calculated from the raw data according to a confidential rule, which yields a different result in the event of a change to the raw data. This way, the data processing unit outside the machine can be sure at all times that the data has not been altered. This is particularly important if the data is used to determine a selling price for the plant material.

[0015] The interface is preferably configured to transmit the raw data via a wireless connection. This saves the personnel operating the machine from having to establish a physical transmission connection or handle a data storage device. However, it is also conceivable for the interface to include a removable data storage device on which the raw data collected during machine use is stored. This removable data storage device is removed from the machine after use to be connected to the data processing unit outside the machine and read out.

[0016] The interface may be designed to transmit the raw data in real time; in the event that the machine moves out of range of the wireless connection, a buffer may be provided to collect the raw data until the connection is available again.

[0017] The invention further relates to a data processing system in which a data processing unit interacts with at least one agricultural machine as described above, the interface of which is configured to transmit the raw data to the data processing unit. The data processing unit is configured to generate a calibration model for deriving at least one parameter of the plant material using the raw data. The greater the number of machines involved in the data processing system, the larger the amounts of raw data that can be collected in the data processing unit in a short time. This enables the creation of calibration models in the data processing unit based on large amounts of data, so that statistical errors during model generation can be largely suppressed by averaging.the influence of a large number of variables on the parameters of the plant material to be determined can be investigated and taken into account in order to create specific models for particular values ​​of these variables.

[0018] The models created in this way can be downloaded from the data processing unit to the machines for use there. They can replace existing calibration models as part of an update to enable more accurate, faster, or otherwise improved determination of parameters already determined by previously existing models; they can also expand the machine's functional range by determining new parameters.

[0019] Further features and advantages of the invention will become apparent from the following description of exemplary embodiments with reference to the accompanying figures. They show: Fig. 1 shows an agricultural machine with an NIR sensor; Fig. 2 shows a block diagram of the NIR sensor and an evaluation unit; and Fig. 3 shows an application environment of the machine.

[0020] Fig. 1 shows a forage harvester 1 in a schematic, partially sectioned side view during the harvesting of a crop 6, in particular corn plants, in a field. The forage harvester 1 has, in a conventional manner, a harvesting attachment 2 that can be exchanged to adapt to the plant material to be harvested, which cuts and collects the plant material S to feed it to a chopping unit 4.

[0021] The chopping unit 4 comprises a conveyor device with several pairs of rollers 3, which removes the plant material from the harvesting header 2, a rotationally driven chopping drum 5, which interacts with a stationary cutting edge to chop the plant material S in a first processing step, and a corn cracker 7. After passing through the corn cracker 7, the plant material S, driven by a post-accelerator, reaches an ejection spout 8. Adjustable operating parameters here are a gap width between the cutting edge and the chopping drum, conveying speed of the pairs of rollers 3 and rotational speed of the chopping drum 5 or the ratio between these speeds, the gap width of the corn cracker 7 and the blower power of the post-accelerator.

[0022] An IR light source and an NIR sensor 8 for monitoring the condition of the plant material S can be provided at various locations along the path of the plant material S through the forage harvester 1; preferably at locations where the plant material to be monitored is shielded from visible external light and IR radiation. Fig. 1 shows, as an example, two NIR sensors 8 which are arranged on the discharge spout 9 in order to monitor the condition of the chopped plant material passing through.

[0023] The NIR sensors 8 are connected to an evaluation unit 15 on board the forage harvester 1, which receives raw data 19 output by the NIR sensors 8 in the form of spectra. Fig. 2is a block diagram of an NIR sensor 8, the evaluation unit 15 and units interacting with them inside and outside the forage harvester 1. The evaluation unit 15 can be part of an on-board computer 10 of the forage harvester 1 or can be combined with the NIR sensor in one structural unit.

[0024] The evaluation unit 15 is connected to a satellite navigation system 11 on board the forage harvester 1, which receives signals from navigation satellites 12, e.g. of the GPS or Galileo system, in a manner known per se in order to determine the geographical position of the forage harvester 1 and to use this, e.g. to support autonomous navigation of the forage harvester 1 in the field.

[0025] Sensor software 16 runs on the evaluation unit 15. Part of this sensor software 16 is formed by a calibration model 17, which is a subprogram of the sensor software 16 that is selected application-specifically from various models stored in a mass storage of the evaluation unit 15 or loaded onto the evaluation unit 15 as needed. It is configured to determine the quantitative proportions of individual substances such as protein, fat, and sugar in the crop for a specific type and / or variety of crop, possibly also specifically for certain environmental conditions during harvesting, and to transmit the results to a control unit 18 of the chopping unit 4. By applying suitable calibration models, additional ingredients can be detected.

[0026] A parameter relevant for the operation of the chopping unit 4 can be, for example, the dry matter content in the harvested material, as this influences the chopping ability and thus the required drive power and speed of the chopping drum 5, the gap width and drive power of the corn cracker 7 via the hardness of the grains contained in the plant material, and the required power of the post-accelerator via the weight of the chopped particles. It is also conceivable to infer the grain content from an assessment of the starch content in the chopped material and to determine the operating parameters of the corn cracker 7 depending on this. In order to keep the computing power requirements of the evaluation unit 15 low, the calibration model 17 calculates only those parameters of the plant material S that are actually used by the control unit 18 to control the chopping unit 4.

[0027] An independent task of the sensor software 16 is to assign to each spectra received over time by the NIR sensor 8 a reference to the location at which the spectrum was recorded. The spectra themselves are not altered by this, so the designation "raw data" 19 is justified for both the spectra output directly by the NIR sensor 8 and the spectra with added location information. The location information is provided by the satellite navigation system 11.

[0028] The sensor software 16 can also encrypt the spectra or calculate a test date from them according to a secret protocol and add it to the spectra to obtain integrity-assured spectra. Since no information is lost through encryption, but rather the original spectra can be restored through decryption, the integrity-assured spectra can also be considered raw data 19.

[0029] A mass storage device 20 is provided to temporarily store the spectra – possibly with integrity-assured data – with the associated location information until they can be transmitted to a stationary data processing unit 21. For this purpose, complementary radio interfaces 22, 23 are provided on the side of the forage harvester 1 and on the side of the data processing unit 21, respectively. The radio interfaces 22, 23 can be compatible with a public mobile radio network in order to enable the transmission of the spectra whenever the forage harvester 1 is within the coverage area of ​​the mobile radio network; however, they can also be, for example, WLAN interfaces, which only enable data transmission when the radio interfaces 22, 23 are within mutual range, e.g., in a barn or garage belonging to the operator of the forage harvester 1.The storage capacity of the mass storage device 20 is dimensioned in each case according to the amount of data to be expected during an interruption of the communication between the radio interfaces 22, 23 and can therefore be smaller when using mobile radio interfaces than when using WLAN interfaces.

[0030] In extreme cases, the mass storage device 20 may have a storage capacity sufficient for the raw data 19 of a harvesting operation and be replaceable so that it can be removed after use and connected to the data processing unit 21; in this case, the mass storage device 20 itself forms the interface for data communication between the machine and the data processing unit 21, and the radio interfaces 22, 23 are not required.

[0031] The data processing unit 21 can be as in Fig. 1shown may be a computer in an agricultural operation to which the forage harvester 1 and / or a field on which it is used belongs; it may also be a computer of a service provider for a plurality of operations, a cloud system, or a combination of the above. The data processing unit 21 stores various calibration models 24 for different types of crop and, if applicable, utility programs for decrypting or checking the integrity of the raw data 19. Among the calibration models 24, there may be some for calculating the parameters taken into account by the chopping unit 4 for different types of crop, from which one is selected before the start of a harvest to match the variety to be harvested and downloaded via the radio interfaces 22, 23 to the evaluation unit 15 to serve there as calibration model 17.Additional calibration models 24 can be provided to calculate the proportions of other compounds in the crop. Thus, the data processing unit 21 is capable of determining the amounts of any value-determining or otherwise interesting substances in the crop based on the raw data 19 transmitted to it, without requiring any computing power on board the forage harvester 1.

[0032] Based on the geographical position data associated with the NIR spectra in the transmitted raw data 19, the data processing unit 21 is able to calculate the aforementioned parameters and substance quantities with spatial resolution and generate maps of their geographical distribution. These can be used to decide on possible plant protection or soil care measures and, if necessary, to implement them with location-dependent intensity, e.g., applying a fertilizer in a location-adapted amount based on a requirement determined from the location-dependently determined substance quantities.

[0033] The data processing unit 21 may belong to an agricultural holding, such as that of the operator of the forage harvester or the owner of the field 13. It may be more expedient if the data processing unit 21 is operated by a service provider that works for a large number of agricultural holdings and is thus able to collect raw data 19 from all of these holdings and use it for the development and optimization of the calibration models.

[0034] It is also conceivable for the data processing unit 21 to comprise multiple computers, including those each belonging to an agricultural holding and storing data on the areas belonging to the holding and the crops grown there, as well as calibration models required for these crops and licensed by the holding, in order to download them to the machines 1 or to perform local evaluations. Furthermore, a central computer that collects the raw data 19 to be used for the development and optimization of the calibration models. The transmission of the raw data 19 from the forage harvester 1 to the data processing unit 21 can be easily ensured if the same transmission path is used for its transmission as for parameter values ​​that have already been determined by the evaluation unit 15 of the forage harvester 1.Since these data are important for the agricultural holding, employees of the holding will generally ensure that they can be transferred to the data processing unit 21, and the raw data 19 can then be transferred via the same route without the need for any further targeted action by the personnel of the agricultural holding.

[0035] Fig. 3shows a typical application situation of the forage harvester 1. The forage harvester 1 harvests a field 13, accompanied by a transport vehicle 14, to which the chopped plant material is continuously transferred via the discharge spout 9. Both navigate the field 13 autonomously or semi-autonomously by continuously checking and correcting their position relative to each other and to pre-planned routes 25 with the aid of the navigation satellites 12. The evaluation unit 15 of the forage harvester communicates with the data processing unit 21 via the radio interfaces 22, 23. Fig. 3shows the data processing unit 21 in a building of an agricultural holding to which the forage harvester 1 and / or the field 13 belong. In practice, the data processing unit 21 can also be set up remotely from the agricultural holding, e.g. at a service provider's premises, or can be a network comprising computers from several agricultural holdings, each of which supplies raw data 19. The radio interface 23 here is a base station of a public mobile radio network 28 that supports mobile data traffic, e.g. via the Internet. The use of a satellite-based mobile radio system is also conceivable. Additional radio interfaces 23 connected to the data processing unit 21, e.g. via LAN, can be provided in a barn 26, in which the forage harvester 1 and / or the transport vehicle 14 are parked when not in use, or at a silo 27 to which the transport vehicle 14 transports the chopped plant material from the field 13.

[0036] The data processing unit 21 maintains a database of the fields cultivated by the farm, which records, among other things, their geographical data, the plant varieties currently and previously grown there, fertilization and plant protection measures, etc., as well as a library of calibration models 24. Based on the geographical data, the data processing unit 21 plans the routes 29, 30 of the forage harvester 1 and the transport vehicle 14 and transmits them, before the start of a harvesting operation, typically via the radio interface 23 of the barn 26. If an update of the routes 29, 30 becomes necessary during use in the field 13, this can be done via the radio interface 23 of the mobile radio network 28.

[0037] Knowing the variety cultivated on the field to be worked, the data processing unit 21 selects the calibration model appropriate for the cultivated variety from the library of calibration models 24 and transmits it to the evaluation unit 15 of the forage harvester 1. Typically, this transmission takes place via the radio interface 23 of the barn 26; the radio interface 23 of the mobile network 28 can also be used, particularly if weather conditions or other unforeseen circumstances force rescheduling during an operation and the downloaded model is no longer appropriate.

[0038] NIR spectra obtained during harvesting, as well as any parameter values ​​determined therefrom by the evaluation unit 15, can be uploaded to the data processing unit 21 via the radio interface 23 of the mobile network 28; alternatively, the transmission can take place after the end of the operation via the radio interface 23 of the barn 26. It is also conceivable for the forage harvester 1 and the transport vehicle to maintain a WLAN while driving on the field 13, via which the NIR spectra collected by the forage harvester 1, together with the plant material to which they refer, are transmitted to a memory of the transport vehicle 14. As soon as the vehicle is within range of a radio interface 23 of the farm, the data are forwarded to the data processing unit 21 via this interface.In either case, upon arrival of a new load of plant material at the silo 27, the data processing unit 21 has the NIR spectra belonging to this load, which enable it to optimally control the compaction of the plant material at the silo 27.

[0039] The collected raw data 19 remain stored there even after evaluation by the data processing unit 21. This allows NIR spectra from a large number of farms, relating to a variety of plant varieties grown on different soils and cultivated or harvested under a variety of weather conditions, to be collected in order to create improved calibration models 31 on the basis of this comprehensive data basis or, if a new influencing factor not previously considered in the state of the art is identified, to develop various calibration models 31 for various possible values ​​of this influencing factor. These can then be transmitted to machines of the connected agricultural farms in order to be used there for the evaluation of the raw data 19; however, they can also be transmitted to third parties 32, e.g.to processors of the plant material, in order to enable them to efficiently monitor the quality of the plant material supplied to them. For the creation of these calibration models 31, in addition to the raw data supplied by the machines, results from analytical values ​​obtained by means other than an NIR sensor can also be used, e.g., by conventional analysis of the constituents of a material sample in a laboratory.

[0040] The transmitted raw data 19 can remain stored on the data processing unit 21 for as long as deemed appropriate. By collecting raw data sets obtained on a given field over the course of various harvests, these can be evaluated using a common calibration model 24, thus providing information about the development of absolute or relative amounts of substances of interest in the harvested crop over several years. Since calibration models can be used that are more recent than the data to which they refer, high-quality evaluations can be obtained even retrospectively.Among other things, this also makes it possible to determine quantities of substances in the harvested plant material that had not yet been searched for at the time the spectra underlying the evaluation were obtained, or for which no sufficiently accurate calibration models were available at the time the spectra were created. Reference symbol

[0041] 1Forage harvester 2Harvesting attachment 3Pair of rollers 4Chopping unit 5Chopper drum 6Plant 7Corn cracker 8NIR sensor 9Discharge spout 10On-board computer 11Navigation system 12Navigation satellite 13Field 14Transport vehicle 15Evaluation unit 16Sensor software 17Calibration model 18Control unit 19Raw data 20Mass storage 21Data processing unit 22Radio interface 23Radio interface 24Calibration model 25Route 26Barn 27Silo 28Mobile network 29Route 30Route 31Calibration model 32Third party

Claims

1. An agricultural machine (1) with a NIR sensor (8) which is configured to detect NIR spectra of plant material and to output as raw data (19), an evaluation unit (15) for deriving at least one parameter of the plant material in real time from the raw data (19), and an interface (22) for data communication with at least one data processing unit (21) externally of the machine (1), characterized in that the interface (22) is configured to transmit the raw data (19) comprising the NIR spectra to the data processing unit (21).

2. The agricultural machine according to claim 1, characterized in that it has a processing assembly (5-10) for processing the plant material and in that a control unit (18) is configured to control at least one operating parameter of the processing assembly (5-10) with the aid of the at least one parameter output by the evaluation unit (15).

3. The agricultural machine according to one of the preceding claims, characterized in that the evaluation unit (15) is configured to use an interchangeable calibration model in order to derive the at least one parameter.

4. The agricultural machine according to claim 3, characterized in that the interface (22) is configured to receive and install the calibration model onto the evaluation unit (15).

5. The agricultural machine according to one of the preceding claims, characterized in that in addition to the raw data (19), the interface (22) is also configured to transmit values for the at least one parameter derived by the evaluation unit (15).

6. The agricultural machine according to one of the preceding claims, characterized in that the at least one parameter is the water content of the plant material or a parameter which is representative of the level of maturity of the plant material.

7. The agricultural machine according to one of the preceding claims, characterized in that it further comprises a timing unit and / or a device for determining position (11) and the evaluation unit (15) is configured to add times and / or locations of their generation to the respective NIR spectra.

8. The agricultural machine according to one of the preceding claims, characterized in that the interface (22) is configured to transmit the raw data (19) in an integrity-protected form.

9. The agricultural machine according to one of the preceding claims, characterized in that the interface (22) is configured to transmit the raw data (19) via a wireless link.

10. The agricultural machine according to claim 9, characterized in that it has a temporary memory (20) for storing the raw data (20) when the wireless link is unavailable.

11. A data processing system with a data processing unit (21) and at least one agricultural machine (1) according to one of the preceding claims, the interface (22) of which is configured to transmit the raw data (19) to the data processing unit (21), characterized in that the data processing unit (21) is configured to generate a calibration model for deriving at least one parameter of the plant material using the raw data (19).

12. The data processing system according to claim 11, characterized in that the data processing system (21) is configured to download the calibration model to one of the agricultural machines (1).