Method for operating machine for harvesting and / or separating root crops, associated machine and associated computer program product

By using image detection units in the excavator to evaluate soil characteristics and adjust the operating parameters of the transport elements in real time, the difficulty in adjusting excavation depth caused by changes in soil characteristics is solved, and the harvesting efficiency and equipment efficiency of root crops are improved.

CN119073078AActive Publication Date: 2024-12-06GRIMME LANDMASCHINENFABRIK SE & CO KG
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411474924.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2020-02-14
Filing Date
2021-02-09
Publication Date
2024-12-06
Estimated Expiration
2041-02-09

AI Technical Summary

Technical Problem

The prior art is difficult to effectively adapt to soil characteristics when excavating root crops, resulting in harvest losses and inefficient equipment, especially when transitioning from soft soil to hard soil, digging depth adjustment is difficult.

Method used

Soil information is recorded using electromagnetic, optical or acoustic image detection units, and the machine's transport element operating parameters are set through evaluation equipment, especially the screening capability characteristics of the screening machine with screening machine, such as the size, shape and color of soil aggregates, and adjust in real time. Setting up the screening section to optimize soil screening.

Benefits of technology

By adjusting the operating parameters of transport components in real time, the damage of root crops is reduced and the harvest efficiency is improved, and the economy and yield of the equipment is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119073078A_ABST
    Figure CN119073078A_ABST
Patent Text Reader

Abstract

The invention relates to a method for operating a machine (2) for harvesting tuberous root crops and / or for separating tuberous root crops (24) from other materials conveyed together, said other materials comprising at least soil in the form of loose soil and / or soil aggregates (28) and possibly weeds and / or stones, wherein at least one test image of at least one part of the material moved forward relative to a machine frame (6) of the machine (2) by means of at least one transport element, in particular a belt screening machine (10), is recorded by at least one electromagnetic, in particular optical or acoustic image detection unit (12), the evaluation device generates a setting signal for setting at least one operating parameter of a transport element and / or another transport element of the machine (2) on the basis of at least one test data set generated on the basis of the test image and / or formed by the test image. Wherein at least one characteristic for describing the screening capacity of the soil transported together is determined by an evaluation device and used for setting operating parameters, and a machine for harvesting root crops and a computer program are proposed.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application is a divisional application of international application PCT / EP2021 / 053070, which entered the Chinese national phase on August 15, 2022, entitled "Operation method of a machine for harvesting and / or separating root crops, the corresponding machine and the corresponding computer program product", with an international application date of February 9, 2021 and Chinese application number 202180014742.8. Technical Field

[0002] The invention relates to a method for operating a machine for harvesting root crops and / or for separating root crops from other materials transported together, the other materials comprising at least soil in the form of loose earth and / or soil aggregates and possibly weeds and / or stones. The method also relates to recording at least one test image of at least a portion of the material which is moved forward relative to a machine frame of the machine by means of at least one transport element, in particular a belt screen, by means of at least one electromagnetic, in particular optical or acoustic image capture unit. Furthermore, operating parameters of the machine are set by means of an evaluation device. Background Art

[0003] In US2017 / 013773 A1, the texture and / or friability of the soil is detected by means of, for example, optical or acoustic sensors with the purpose of adapting the depth of the digging element according to the soil information. This is based on the knowledge that, given the same settings, the working depth of the digging element decreases when transitioning from soft to hard soil, which can result in root crop losses. Thus, by setting the digging depth of the harvester according to the soil properties before digging the harvest, harvest losses are reduced and a better harvest is obtained. Summary of the invention

[0004] The object of the invention is to use soil information for processing material in a machine.

[0005] This object is achieved by operating a method for a machine for harvesting root crops and / or for separating root crops from other materials transported together. In addition, this object is achieved by a machine according to the invention and by a computer program product according to the invention.

[0006] According to the present invention, in an operating method of a machine for harvesting root crops and / or for separating root crops from other materials transported together, wherein the other materials include at least soil in the form of loose dirt and / or soil aggregates and possible weeds and / or stones, firstly at least one test image of at least a part of the material moved forward relative to a machine frame of the machine by means of at least one transport element, in particular a belt screen, is recorded by means of at least one electromagnetic, in particular optical or acoustic image detection unit, and a setting signal for setting at least one operating parameter of a transport element and / or another transport element of the machine is generated by means of an evaluation device based on a test data set generated according to and / or formed by the test image, wherein at least one feature for describing the screening capacity of the soil transported together is determined by the evaluation device and used for setting the operating parameter.

[0007] The action of one or more regulating or driving mechanisms of the transport element is changed by a setting signal to vary the screening of soil. The invention therefore relates to a transport element in the form of a separation device for screening soil, which is in particular a belt screen, a hedgehog belt and / or a drum device.

[0008] In particular, the image capture unit is designed with its sensor aligned as a transport element of a belt screen, which transports the material through the machine as a first or second belt screen after picking up the material. Thus, in the described embodiment of the method according to the invention, the soil is analyzed very early after picking up, in particular when the soil is detected in the first half or the first third of the belt screen, so that a reaction to a change in the soil state can be made very quickly by adapting the operating parameters, and thus at least a part, in particular a major part, of the screening section comprising one or preferably several belt screens can be adapted immediately and even preventively. Preferably, within a time window of at most 1 minute, preferably at most 30 seconds after recording the test image, the operating parameters are set, i.e. the evaluation device is designed so that it completes the evaluation of the test image within the described time. Furthermore, in particular within the described time, at least the adaptation of the operating parameters is prompted or carried out.

[0009] Such aggregates consisting of soil or earth components with a minimum diameter of not less than 5 mm are considered as soil aggregates. For example, in the case of non-spherical aggregates, the equivalent diameter can be considered as the diameter. Soil aggregates are always composed of a plurality of interconnected particles, which in particular include sand, silt and / or clay. Loose soil is a particle size segment up to 2 mm and, according to the above definition, is a coherent particle with a diameter less than 5 mm.

[0010] Likewise, an optical or acoustic image detection unit or another such image detection unit can direct its sensor toward the area behind the transport element in order to be able to respond to any remaining soil aggregates. Thus, if too few smaller aggregates are detected and too many large aggregates are detected, the screen width can be reduced and, for example, the striking performance of the strikers acting on the belt screen can be increased to reduce the large aggregates.

[0011] Preferably, the components of the material present in the test image are determined, in particular by means of a classification method, which material comprises loose soil and / or soil aggregates, in particular clods, wherein clods are generally regarded as larger aggregates, i.e. soil fragments with a diameter of more than 5 cm. The term "clod" is sometimes used as a synonym for elongated clods. Based on this classification, for example by evaluating the color values ​​of the test data set, the individual components of the observed transported material can be identified. Therefore, in order to observe soil aggregates, test images or parts thereof can be considered in a simple manner, which test images or parts thereof only or at least have a predominant component (see below) of soil aggregates.

[0012] Preferably, the features used to describe the sieving capability include one or more values ​​describing the size, shape, moisture, hardness or color of one or more soil aggregates and / or one or more, in particular statistical, distributions of the size, shape, hardness or color of a plurality of soil aggregates. For example, via such values, all soil aggregates present in a test image or test data set can be assigned a size class so that for the purpose of better automated processing, no individual soil aggregates have to be considered.

[0013] Preferably, the size of the soil aggregate involved, in particular the soil clod, is taken into account. For example, this can be a diameter, an equivalent diameter, a projected area or a volume. In another alternative or supplementary embodiment of the present invention, the shape or color parameters of the aggregate can be extracted, because the color and shape change accordingly under the conditions of different soil types and fluctuating soil moisture, which are the main influencing factors of the screening capacity. The same applies to hardness, which can be derived from color, shape and size. For example, the breaking load can be used for the measurement of the hardness of the soil aggregate.

[0014] In a further alternative or supplementary embodiment of the method according to the invention, one or more, in particular statistically aggregated, characteristics or distributions formed by values ​​of size, color, hardness and / or shape of a plurality of, in particular different soil aggregates are used as screening capacity characteristics.

[0015] Such statistical features may be, for example, the mean, standard deviation / variance, median, percentile or first, second or k-th order moment, wherein said feature aggregates the distribution of a plurality of individual features.

[0016] According to a development of the invention, the size, shape or color is classified into a predetermined number of classes, preferably five classes, wherein the values ​​increase or decrease monotonically from the first class to the last class. The average membership or the most likely class of a plurality of considered soil aggregates of the test data set is then used as a feature to describe the screening capacity. For example, if small aggregate size classes are present on average, for example due to predominantly excavating in dry sandy soils, the screening performance can be reduced, for example by setting a small screen bar or roller spacing.

[0017] Advantageously, in a further embodiment of the method according to the invention, the screening capacity characteristics obtained in a plurality of temporally successive evaluation cycles are temporally filtered out or offset against each other before one or more operating parameters or their settings are determined on the basis of the latter. In particular, the use of a low-pass filter or a smoothed mean value is suitable for smoothing possible outliers.

[0018] The test data set is formed by a test image or a portion thereof. Likewise, it may be a data set obtained by processing and / or analyzing a test image or a portion thereof. Furthermore, the test data set may have the test image itself, a processed test image and / or a data set derived from an analysis of the test image. The same applies to corresponding portions of the test image, the processed test image and a data set created based on an analysis of at least a portion of the test image.

[0019] At least one characteristic obtained on the basis of a test data set for determining the screening capacity is used to set the operating parameters of the transport element, in particular the transport element or one of the transport elements in a belt screen, which is used to screen out soil. After the qualitative determination of the components present in the harvested crop, for example according to the color values ​​as described in DE 102018127844 A1, one or more screening capacity characteristics are then quantitatively determined. The transport element can be set in an improved manner by means of the method according to the invention in the following manner: that is, at the area of ​​the transport section that is present after picking up, there is a desired amount of soil aggregates, such as clods or clod sizes, so that the harvest to be used in the form of root crops is optimally separated according to the desired soil aggregates, in particular the soil aggregate size, or is also protected during transportation through the device. Therefore, by means of the method according to the invention, the yield of the operating machine is increased while reducing the risk of damaging the root crops, which leads to better economy in the use of the machine.

[0020] Such an image acquisition unit for producing an image of the material transported on the transport element is referred to as an electromagnetic image acquisition unit, which has a sensor that works for detecting electromagnetic waves, in particular light waves. Here, it can be a sensor that can be used for detecting one-dimensional or multi-dimensional images. For example, it is one or more radar sensors that can receive waves reflected from the material, the frequency of which is in the range between 10 and 150 GHz. The optical image acquisition unit is designed to detect light, in particular light in the visible, ultraviolet and / or infrared range.

[0021] In particular, cameras in the form of 2D, 2.5D or 3D cameras, i.e. for example RGB cameras, time-of-flight cameras, black-and-white or grayscale cameras or stereo cameras, can be considered as optical image detection units. Likewise, methods using light gap methods or structured light, plenoptic cameras, etc. can be used to detect the harvested crop. The acoustic image detection unit can be present in particular in the form of an array of distance-measuring sound sensors, in particular ultrasonic sensors, which are also suitable for generating imaging input data, i.e. test images, for the evaluation device due to the continuous movement of the harvested crop flow along the sensor array mounted on the machine frame. The same applies to an array arrangement of optical distance sensors or an array arrangement of mechanical scanning probes. RGB cameras usually have a two-dimensional image sensor in the form of a CCD or CMOS sensor.

[0022] It goes without saying that in the method according to the invention and in the device according to the invention, an electromagnetic image capture unit and an acoustic image capture unit can be combined and that the image capture unit can have one or more electromagnetic and / or acoustic sensors.

[0023] The feature is preferably determined based on an input data set generated by or formed by a test data set by means of a neural network analysis, a histogram analysis and / or a structure-from-motion analysis. The analysis is particularly suitable for large data volumes, which are formed when observing crop flows during a harvesting trip, in particular of a potato or beet harvester.

[0024] In particular, in a development of the method according to the invention, the neural network is designed as a convolutional neural network, which classifies each input data set into one of a plurality of classes, which represent values ​​of different sieving capacity characteristics. This type of neural network has proven to be particularly good for identifying soil aggregates and their assigned sieving capacity characteristics during the harvesting process.

[0025] In particular, the evaluation device comprises one or more CPU units and / or one or more graphics processor units, in particular in the form of a GPU (graphics processing unit) or GPGPU (general purpose graphics processing unit) and / or a processor unit based on an FPGA (field programmable gate array). This form of evaluation device allows the test data sets to be evaluated in a particularly resource-saving manner and in particular locally. It goes without saying that an evaluation device constructed as an EDV device or constructed in this way has other conventional means, such as means for power supply, interfaces, working memory and non-volatile program and data memory.

[0026] According to the present invention, in particular in the case of difficult harvesting conditions in which the picked-up soil or the picked-up earth is broken into only a few large aggregates, a knocker in the form of a rotor knocker or a swing knocker can be set to improve the screening performance of the transport element, in particular the screening performance of a hedgehog belt or a belt screen. Thus, an additional oscillating movement of the transport element is caused and thus the screening performance is improved, because, for example, soil aggregates in the form of mud blocks are broken by impact. Other or supplementary operating parameters of the conveying element configured as a belt screen are in particular the belt screen speed, the pick-up belt screen speed, the set height of at least one triangular roller, the set height of the possible descending step, the frequency of one or more knockers, for example the amplitude of a vibrating knocker, the position of the knocker or the regulating mechanism acting on the belt screen relative to the belt screen and / or the net width of the belt screen. Therefore, the operating parameters of the belt screen are operating variables, such as speed, frequency, amplitude or position, which can be set by a possible unit acting on the belt screen. Thus, the setting signal is correspondingly a signal that causes the setting of the variable, which is output or initiated by the evaluation device. When the transport element is configured as a hedgehog belt, the aforementioned setting mechanism is sometimes also present, in particular for setting the belt speed. When the transport element is a roller arrangement, its rotation speed or the spacing between them can be set.

[0027] An analysis, in particular based on a camera, of the material flow on a transport element, in particular a belt screen, results in the detection of features from which the average aggregate size can be estimated, so that on the basis of this information, knockers or other means for influencing the screening performance of the belt screen can be automatically activated or deactivated in a demand-oriented manner and their function can be adjusted. Thus, for example, a plurality of setting means arranged one behind the other relative to and in particular parallel to the transport direction of the screening section can be variably set.

[0028] The image capture unit preferably captures in particular at least the front third of the belt screen. Depending on the position of the viewing angle of the camera, the optically distorted image of the section of interest is advantageously converted into a rectangular display.

[0029] In order to detect the crop particularly well and precisely, in particular independently of the ambient conditions, the considered belt screen section or the crop transported thereon can be illuminated by means of a lighting unit.

[0030] Preferably, for determining the aggregate size of soil or soil aggregates, for example the size of clods, a region of the test image or test data set which has at least 75%, preferably 90%, more preferably 95% and even more preferably only soil or clods and which is particularly continuous and preferably rectangular is selected by the evaluation device. Even if the region does not appear in every image at high travel speeds during harvesting operation, it has been shown that a correspondingly limited region for determining the clod size provides particularly good results and that even small regions, particularly at least 15 cm×15 cm, are sufficient to represent possible settings or changes in operating parameters.

[0031] In particular, such regions are selected automatically. This is done in particular by examining the color information contained in the test data set, possibly the edges or gradients of the 2D or 3D image and possibly using statistical analysis. For this purpose, in particular, a pixel-by-pixel classification based on the color information is used.

[0032] A representation of a test data set, which represents at least a coherent and preferably rectangular portion of the area having primarily soil, is directly input or processed as an input data set into a neural network analysis, a histogram analysis and / or a kinematic reconstruction structure analysis to determine at least one clod size, in which at least one clod size for setting operating parameters is associated with the image area.

[0033] In particular, for the use of neural networks, preferably convolutional neural networks (CNNs), comprising convolutional layers, the input data set can be converted, for example, into vector form. Topologically, CNNs can be sequential or regressive networks. Convolutional layers are processing planes in the network that apply convolutional filters to input matrices. As in other layers, convolutional filters also have degrees of freedom in the form of weights. Here, convolutional filters are used to extract image features, which are then classified, for example, based on the image features. However, in order to make the application of CNNs practical for the present invention, some improvements must be performed in their development. In order to reduce the storage requirements of the network, separable weights are used for the neurons of the convolutional filters, assuming that image features are independent of their position in the image and are equally interesting. In addition, separable weights are also stable with respect to translation variance, rotation variance, scale variance and brightness variance. In addition, in particular, pooling layers are inserted between convolutional layers, which discard values ​​by region. Preferably, maximum pooling is used, thereby discarding all values ​​in a small range (e.g., 2×2 or 4×4) except for the largest and most important values.

[0034] In order to make the calculation of CNN as simple as possible, ReLU (Rectified Linear Unit) is preferably used as the activation function. The activation function is always 0 in the negative range and increases linearly in the positive range. With this improvement, CNN can be used particularly well for real-time operation of the machine according to the invention. In addition, computer architectures that can be used for parallel calculations are advantageous. Neurons in a layer can be calculated independently of each other, and thus in particular in parallel. Therefore, architectures with multiple parallel computing units, for example in the form of GPUs, are well suited for CNN.

[0035] The image part is preferably selected so that it has the maximum value of the soil density distribution, that is, the image part is automatically selected so that it has the maximum value of the corresponding density distribution that mainly includes clods of soil or soil. This distribution is derived in particular based on a pixel-by-pixel classification, in which all pixels of the image area under consideration are classified in terms of their realized probability with possible object groups, namely root crops, soil / clods of soil and possible stones and / or weeds. Other classification variables can also represent "empty belt screen", that is, there is no possible harvested crop. Therefore, preferably, the pixels that show the area with the largest amount of soil are used to select the image part. Here, the maximum value of "soil density" is the best area, which is provided as an input variable to an evaluation method that preferably works based on a neural network, in particular a CNN, to detect one or more screening capacity characteristics. Preferably, image parts with a soil or soil density of less than 75% are not used to determine the characteristics describing the screening capacity, so that the determination of at least one screening capacity characteristic is not impaired by too many possible foreign objects, such as stones, weeds or root crops. Therefore, for example, the image part can also be automatically defined according to the lower limit threshold of the minimum soil density.

[0036] It goes without saying that the neural network is trained in advance with the aid of a plurality of test images or test data sets of the same size corresponding to the adapted image. In the case of images of similar size, the edge regions of the smaller image can be filled with zeros by so-called zero padding. In the case of image size fluctuations within a range of maximum 10-15%, this does not lead to a significant deterioration of the image analysis results. For example, for this purpose, one of the five size classes is manually assigned to each of the plurality of test data sets.

[0037] In particular, when using CNN, one or more levels of the size of soil aggregates can be associated with a test image or an image region of a test data set. Thus, for example, a corresponding value can be associated with each image on a scale of 1 to 5. During the learning phase of the CNN, the screening capacity value is determined by a human expert for a representative number of example images, based on which the CNN can learn its internal weights. During the online prediction phase, the values ​​estimated by the CNN are then automatically used to adjust or set one or more operating parameters.

[0038] In the context of the prior art, extensive tests have shown that for a number of harvesting conditions it has proven to be advantageous to define a particularly rectangular input data set for the neural network with edge lengths of between 100 and 400 pixels, preferably between 150 and 250 pixels. These sizes represent a very good compromise between the required computing power and the duration in the case of prevailing variances in order to implement the setting of the operating parameters during a continuous operation of less than 30 seconds, preferably less than 3 seconds, after recording an image.

[0039] Furthermore, when using the input image size, a sufficiently large image region can be found in a sufficiently large proportion (approximately more than 75%) of the camera image during machine operation, which image region is substantially occupied only by soil or soil aggregates. The image size can also be linked to the resolution of the image sensor device used, for example a camera, which is in particular in the range of 0.5 to 10 megapixels, preferably in the range of 1.0 to 1.5 megapixels.

[0040] The evaluation device preferably evaluates the test data set at least partially locally on the machine or directly connected tractor vehicle. If a large database is required, in particular against the background of a large number of set parameters, the evaluation device can also evaluate the test data set at least partially, preferably also completely, on a wirelessly connected server, but here too a correspondingly fast and stable data connection is required for continuous operation.

[0041] Hybrid forms of this evaluation are also contemplated, such that a portion of the hardware used for the evaluation may reside on the local machine and another portion at a remote location.

[0042] In a further embodiment of the invention, the moisture content of the soil aggregate is additionally determined by means of possible environmental sensors, such as soil sensors, in particular moisture sensors, and is additionally used in the evaluation device to set operating parameters. The moisture sensor can operate on an electrical or optical basis. Sensors for detecting the electrical conductivity can also provide more information about the soil state, which can be used when setting the operating parameters of the screening section.

[0043] In another embodiment of the method according to the invention, the machine is equipped with a position sensor, such as a GNSS receiver, and has map material of the surface to be harvested, which is either stored directly on the machine or provided via a remote server by means of a mobile network connection. In this case, the local soil type stored in the map material in combination with the machine position can be a further auxiliary variable for setting the operating parameters of the screening section.

[0044] In one embodiment of the method, the machine can also be connected to a remote server and can additionally retrieve weather data, ie, humidity and temperature information, via the remote server. Said information can also be additionally included in the determination of the operating parameters of the screening section.

[0045] The determination of the operating parameters is preferably part of a control loop of the machine, in which in particular the weather, i.e. humidity and temperature and / or soil type and / or excavation strategy are additionally used as input or reference parameters. Other high-frequency detectable, usually one-dimensional input variables, such as the filling level of a conveying element or load factor values ​​of a drive element, such as pressure, torque or current consumption, can be included in such a control loop. The operating parameters of the machine and in particular the operating parameters of the screening section can be regulated in a high-frequency manner with the aid of such input variables. With the aid of such a control loop, the excavation depth and / or the travel speed can also preferably be regulated.

[0046] In this case, reference parameters, such as weather, i.e. humidity and temperature and / or soil type and / or excavation strategy and in particular the inventively specific features for describing the screening capacity of the soil, are used in the context of cascade control to optimize the parameters of a control loop, in particular a cascade-structured control loop, and thus to compensate for fluctuations in disturbance variables, such as soil type, which cannot be detected due to the input variables of the high-frequency control loop. Preferably, the cycle time of the high-frequency control loop is in the range of 1 to 100 milliseconds, while the detection of one or more screening capacity features from the test data set is run with a cycle time of 100 milliseconds to 30 seconds.

[0047] In particular, all input parameters that may be relevant for setting the operating parameters are linked to one another in a corresponding data pool, in particular in the form of a local or externally stored database, wherein the screening capacity characteristics and operating parameter values ​​and in particular environmental variables can also be put into an empirical or, if necessary, analytical relationship. Here, the operating parameters can also be set under different conditions with the aid of the data pool.

[0048] In an embodiment of the method according to the invention, the evaluation device constructed as an EDV device is configured so that, in addition to calculations, it also records image data and / or other sensor data and / or screening capacity characteristics and / or other intermediate calculation results and / or operating parameters and / or environmental parameters, especially in a manner linked to the position information of the machine. The data can be stored locally on the evaluation device carried in the machine or transmitted to a central server via a mobile connection. The data can be used to display the screening performance in the form of a map on the processing site. In addition, in particular when the data is recorded by multiple machines and centrally collected, the data can be used to improve the algorithmic logic for detecting the screening capacity characteristics and the possible adjustment logic for determining the operating parameters.

[0049] The objects mentioned at the outset are also achieved by a machine for harvesting root crops and / or for separating root crops, the machine comprising: at least one electromagnetic, in particular optical, or acoustic image capture unit; a transport element that is movable relative to a machine frame of the machine and is in particular designed as a belt screen; and an evaluation device; and means for setting the transport element or a further transport element, wherein the machine is suitable for carrying out the steps of the method according to the invention and the further embodiments of the method according to the invention described above and below. In particular, the means for setting the transport belt are means for setting the belt screen as described above and below.

[0050] This object is also achieved by a computer program product comprising instructions which cause a machine for harvesting root crops and / or for separating root crops according to the invention described above or below to execute the method steps described above or below. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Other advantages or details of the present invention can be derived from the following description of the drawings. Schematically shown:

[0052] Figure 1 shows a side view of a machine according to the invention,

[0053] Figure 2 Show according to Figure 1 a stereogram of a part of an object,

[0054] Figure 3 Show according to Figure 1 the part of the object detected by the optical image detection unit,

[0055] Figure 4 showing the selection of a region of the test image,

[0056] Figure 5 Show according to Figure 4 The possibility of classification of the area

[0057] Figure 6 A mechanism for influencing a belt screen is shown,

[0058] Figure 7 Another mechanism for influencing a belt screen is shown,

[0059] Figure 8 Another mechanism for influencing a belt screen is shown,

[0060] Fig. 9 Another mechanism for influencing a belt screen is shown,

[0061] Fig.10 A flow chart showing a method according to the invention is shown,

[0062] Fig.11 A further diagram shows a further method sequence according to the invention. DETAILED DESCRIPTION

[0063] The various technical features of the embodiments described below can also be combined with the above-mentioned embodiments to realize the subject matter according to the present invention. Whenever it makes sense, functionally equivalent elements are provided with the same reference numerals.

[0064] The machine 2 is currently used for harvesting root crops in the form of potatoes and is therefore designed as a potato digger. The material picked up in the area of ​​the pick-up 4 in the form of soil or soil aggregates, root crops, weeds and / or stones is transported in the conveying direction 1A via a conveying element in the form of a belt screen 10 supported behind the machine frame 6 and other frame parts 8. The belt screen 10 is directly connected to the pick-up 4 ( Figure 2 ). The material transported in the direction 1A by means of the belt screen 10 is detected by a first optical image detection unit 12 in the form of an RGB camera, which is fixed to the machine frame part 9 and is obliquely pointed toward the belt screen 10 in the direction of the pick-up 4. The lighting device 14 illuminates the belt screen in the area of ​​the second image detection unit 12, which is arranged behind the first image detection unit 12 in the conveying direction 1A.

[0065] Machine 2 is set by Figure 2 The area detected by the first image detection unit 12 on the right is Figure 3 , there is no earth filling here. In particular, the individual screening bars 16 of the belt screen 10 can be seen, which are supported directly downstream of the horizontal digging blade 18 of the pick-up 4. Below the belt screen 10, there is a setting device on the machine frame 6 for setting the operating parameters for operating the belt screen 10. Here, the setting device can be Figure 3 The roller 20 shown can be part of a further setting mechanism according to the design of the machine 2 according to the invention. The image section 22 ( Figure 4 ), which contains as little root crops 24 and admixtures 26 (here: weeds) as possible. Image section 22 currently has at least 90% soil aggregate 28 and is already oriented in a rectangular shape, while the rest of the image is still shown with a slight perspective distortion.

[0066] Weeds and soil aggregates are identified by means of a pixel-by-pixel classification, for example based on color values ​​including grayscale values ​​and / or values ​​representing the actual color, recorded by the optical image capture unit 12. The color values ​​are compared with reference values ​​or reference value ranges. This form of differentiation enables a qualitative identification of the components on the test image and the association of pixels with a class of (harvested) material (soil / soil aggregates, weeds, root crops, stones), in particular within a predeterminable or predefined threshold value.

[0067] If a region 22 is identified, a test data set or a test data set portion representing said region is fed to the neural network in a version that may be adapted to the input requirements of the neural network. The neural network, in particular the CNN, associates the size of at least one soil aggregate with the image region and, in a further embodiment of the invention, also the ratio of the size distribution of different belt screen sections in the image with the image region. Figure 4 The detail of FIG. 1 shows the size of the soil blocks with particularly large soil blocks, as can be seen from Figure 5 In the right-hand part of the figure, further size classes of soil aggregates are shown, moving from the right-hand part of the figure to the left, which were identified by the neural network and with which the neural network was pre-trained.

[0068] Depending on the aggregate size thus defined, the operating parameters can be varied, e.g. Figure 6 The frequency of the movement of the vibrating beater 30 or the amplitude of the deflection shown in FIG. The vibration movement of the vibrating beater transmits pulses to the belt screen 10, which causes the soil aggregates, in particular soil clods, to be comminuted. Alternatively or additionally, Figure 7 The frequency of the rotor beater 32 shown in Figure 8 The position of the triangular wheel 34 shown in FIG. 1 is changed relative to the machine frame 6 carrying the belt screen. Fig. 9 ) changes its distance from the belt 3 of the belt screen machine 10, so that the screen bar unit formed by two screen bars 16 connected via connectors 38 moves, and thereby changes the net width of the opening between the successive screen bars of the successive screen bar units.

[0069] according to Fig.10The process of the method according to the invention begins with a first method step 40, in which a test data set 42 is generated by means of the image acquisition unit 12, which is then quantitatively divided in step 43 by means of a pixel-based classification, so that an association 44 of individual image regions or pixels of the test data set with potatoes, weeds, dirt or soil, etc. can be performed. Then, in step 46, a region or image part 22 is selected which only has dirt and corresponding soil aggregates. The region 22 is analyzed in step 50 by means of a CNN, via which the image is classified according to Figure 5 The size class is associated with the image part in the result 52. Then, in step 54, the operating parameters are changed if necessary by outputting or initiating setting signals, so that the screening performance of the belt screen 10 is adapted.

[0070] According to step 54, the screening performance of the belt screen 10 and thus also the screening route are preferably set by the control loop 60 ( Fig.11 ), in which the machine controller 62 including the evaluation device accesses a database 64 which is present locally or externally and, from the association rules, obtains a digging depth setting 66 and / or a travel speed 68 for setting the operating parameters of the belt screen according to step 54. In addition, a variety of other information can be processed in the machine controller 62. This includes information on the material level detection 70 at the start of screening and / or information on the material level detection 72 at the end of screening and / or information on the pressure 74 of a possible separation device and / or information on the blockage detection 76 of a possible separation device. Finally, the actual values ​​of the individual operating parameters of the individual functional units can be detected in step 78 and processed as input information for the machine controller 62.

[0071] Typically, the evaluation device 80 is part of the machine controller 62. Additional input information for the machine controller 62 is, besides the estimate 52 of the soil aggregate size, an excavation strategy 82 which can be preset by the operator and / or environmental information about the weather and soil type, which comes from the detection 84. The circle 86 symbolizes the influence of the size class identification 52, the excavation strategy detection 82 and the environmental variable detection 84 performed by the evaluation device on the excavation performance of the machine 2 as depicted by steps 70 to 78. For example, in an excavation strategy that takes into account the maximum gain, the amplitude of the vibrating beaters, the screen bar spacing and the belt speed are maximized when the largest aggregate size class is identified, while in the case of a milder strategy the amplitude is increased less strongly and the belt speed is reduced at the same time.

Claims

1. A method for operating a machine (2) for harvesting root crops and / or for separating root crops (24) from other material transported together, the other material comprising at least soil in the form of loose earth and / or soil aggregates (28), wherein at least one test image of at least a portion of the material being moved forward by means of at least one transport element relative to a machine frame (6) of the machine (2) is recorded by at least one electromagnetic or acoustic image capture unit (12), and an evaluation device generates a setting signal for setting at least one operating parameter of the transport element and / or another transport element of the machine (2) based on at least one test data set generated on the basis of the test image and / or formed by the test image, At least one characteristic for describing the screening capacity of the soil transported together is determined by the evaluation device and used to set the operating parameters, the operating parameters of the transport element constructed as a belt screen (10) are the belt screen speed, the pickup belt screen speed, the set height of at least one triangular roller, the set height of the descending step, the frequency of the knocker, the amplitude of the knocker, the position of the knocker and / or the net width of the belt screen, and the characteristic includes one or more values, which describe the size, shape, hardness or color of one or more soil aggregates (28) and / or one or more statistical distributions of the size, shape, hardness or color of multiple soil aggregates (28).

2. The method according to claim 1, characterized in that The features are determined by the evaluation device on the basis of an input data set generated by or formed by the test data set by means of a neural network analysis, a histogram analysis and / or a structure from motion analysis.

3. The method according to claim 2, characterized in that The neural network is designed as a convolutional neural network, which classifies each input data set into one of a plurality of classes, wherein the classes represent values ​​of different screening ability characteristics.

4. The method according to any one of claims 1 to 3, characterized in that The components of the material present in the test image are determined using a classification method.

5. The method according to claim 1, characterized in that In order to determine the features by the evaluation device, a region of at least 75% of the test image or of the test data set is selected (22).

6. The method according to claim 5, characterized in that In order to determine the features by the evaluation device, at least 90% of a region of the test image or of the test data set is selected (22).

7. The method according to claim 5, characterized in that In order to determine the features by the evaluation device, at least 95% of a region of the test image or of the test data set is selected (22).

8. The method according to claim 4, characterized in that In order to determine the features by the evaluation device, a region (22) of the test image or of the test data set is selected which contains only soil aggregates (28) and is continuous and rectangular.

9. The method according to any one of claims 4 to 8, characterized in that The portion of the test data set representing the region (22) is directly input into a neural network analysis, a histogram analysis and / or a motion recovery structure analysis or processed as an input data set, and in the neural network analysis, histogram analysis and / or motion recovery structure analysis, the features used to set the operating parameters are associated with the region (22).

10. The method according to any one of claims 1 to 2, characterized in that The evaluation device evaluates the test data set at least partially locally on the machine (2) or on a directly coupled traction vehicle.

11. The method according to any one of claims 1 to 2, characterized in that The evaluation device evaluates the test data set on a wirelessly connected server.

12. The method according to any one of claims 1 to 2, characterized in that The moisture content of the soil aggregate (28) is determined by means of a moisture sensor and used in the evaluation device to set the operating parameters.

13. The method according to any one of claims 1 to 2, characterized in that The determination of the operating parameters is part of a control loop of the machine (2), wherein weather and / or soil type and / or excavation strategy are additionally used as input parameters.

14. The method according to claim 13, characterized in that The excavation depth and / or the travel speed are additionally regulated by means of the control circuit.

15. The method according to any one of claims 1 to 2, characterized in that The operating parameters are set using a data pool in which characteristics and operating parameters as well as environmental variables are stored in a linked manner.

16. The method according to claim 15, characterized in that The data pool is in the form of a database.

17. The method according to claim 1, characterized in that The electromagnetic image detection unit (12) is an optical image detection unit.

18. A machine (2) for harvesting root crops (4) and / or for separating root crops (4), the machine comprising: At least one electromagnetic or acoustic image detection unit (12); a transport element movable relative to a machine frame of the machine (2); and an evaluation device; and a mechanism for setting the transport element or another transport element, wherein the machine is suitable for performing the steps of the method according to any of the preceding claims.

19. The machine (2) according to claim 18, characterized in that The transport element is in the form of a belt screen (10).

20. The machine (2) according to claim 18, characterized in that The image capture unit (12) is an optical image capture unit.

21. A computer program product comprising instructions causing a machine according to claim 18 to perform a method according to any one of claims 1 to 17.

Citation Information

Patent Citations

  • Method for operating a machine for harvesting and / or separating root crops, an associated machine and an associated computer program product

    CN115135133B

  • Method for regulation of a working unit of a harvesting machine

    EP1763988A1

  • Agricultural machines with image processing system

    EP3299996A1

  • Method And Arrangement For The Optical Evaluation Of Harvested Crop In A Harvesting Machine

    US20140050364A1

  • Automated Control Systems and Methods for Underground Crop Harvesters

    US20170013773A1