Method for controlling the operation of a machine for harvesting root crops

Abstract

A method for controlling the operation of a machine (2) for harvesting root vegetables (4) and / or separating the root vegetables (4) from further harvested material containing contaminants, comprising taking, by at least one optical image capture unit (6), at least one inspection image (8) of at least a portion of the harvested material being advanced relative to the machine frame (12) by at least one transport member (10, 10A, 10B, 10C, 10D, 10E, 10F, 10K, 10T), wherein an evaluation device generates a separation device setting signal for setting at least one operating parameter of a separation device of the machine (2) based on an inspection data set generated using or formed by the inspection image (8). The present invention further relates to a machine as described above.

Description

[Technical field]

[0001] The present invention relates to a method for controlling the operation of a machine for harvesting root vegetables and / or separating them from a further harvest, as well as to such a machine, in which at least one inspection image is taken by at least one optical image capture unit of at least a part of the harvested product that is advanced relative to the machine frame by means of at least one transport member.

[0002] The inspection images show the crop previously collected by or fed into a machine for harvesting root vegetables, with a transport member that is part of the machine and serves to advance the crop inside the machine, with at least a portion of the crop coming into direct contact with the transport member.

[0003] US 2018 / 0047177 A1 discloses a method in which the captured inspection images are used to calculate the speed of the conveying member, which is then adapted to the actual conveying speed of the conveying member.

[0004] A drawback of the known method described above is that, depending on the harvesting conditions, the root vegetables can be severely damaged or a large amount of contamination can occur in the root vegetables unloaded from the machine. Therefore, it is generally proposed in US 2018 / 0047177 to change the harvesting speed or one or more configurations of the machine depending on a server-based evaluation of three-dimensional data of the harvested product detected by the machine's sensors.

[0005] The object of the invention is to provide a method which improves the protection of root vegetables, with the greatest possible proportion of contaminants being separated from the root vegetables.

[0006] According to the invention, this object is achieved in a method as described above in that an evaluation device generates a separating device setting signal, which sets at least one operating parameter of a separating device of the machine on the basis of an inspection data set generated with the aid of the inspection images or formed by the inspection images.

[0007] The aforementioned machine is a vehicle that is self-propelled or towed during operation for harvesting root vegetables, in particular potatoes, turnips, carrots or chicory. During the implementation of the method, the machine is in particular advanced in the direction of the rows, in particular the furrows, of the root vegetables to be harvested, and these root vegetables are collected from the ground in one continuous process as part of the harvest. After the harvest has been collected, at least a part of the harvest, in particular the root vegetables and / or contaminants, is advanced at least in part by at least one conveying element relative to the machine frame of the machine. In particular, the conveying element is circulating and is formed as a conveying belt, preferably as a sorting belt or as a rotating star-shaped sorting element.

[0008] Alternatively, the machine may be a machine for separating root vegetables from contaminants in the harvest, for example from clods, stones or soil.

[0009] The separating device is part of the machine together with a number of possibly individually adjustable separating elements, and in this case preferably cooperates with one or more conveying elements. Alternatively, the separating device is part of the conveying element, at least together with this (for example in the case of a sorting belt provided with a rocking device) or together with one or more conveying elements (for example in the case of a scraping roller). During operation, the movement of the crop relative to the separating device exerts a kinetic impact on at least one component of the crop, in particular root vegetables or contaminants. The separating device is provided with circulating separating elements, for example in the form of scraping rollers, in particular in the form of rollers, in which case the different components of the crop are not moved by the separating device at least in the same direction.

[0010] The optical image capture unit is in particular arranged in a fixed position on the machine above the transport member and is oriented towards the transport member, i.e. towards the flow of the harvested product or its components, in particular root vegetables or contaminants, located between the image capture unit and the transport member during operation. The method according to the invention is in particular carried out during mechanical harvesting or separation and is preferably repeated therein.

[0011] The inspection image is in particular a multi-dimensional, preferably two-dimensional, representation, in which at least a part of the harvested product with the root crop and impurities and / or the transport member are represented. Based on the inspection image captured by the image capture unit, the inspection data set is already generated by the image capture unit or by the evaluation device. Alternatively, the inspection data set can be formed by the inspection image itself. This applies in particular to image capture units whose inspection images already have a suitable format for further analysis in the evaluation device. The inspection data set is in particular a data set that is at least temporarily present in the system, which has been created, for example, by processing, such as filtering and / or different representations, and whose information, for example color information in the form of one or more color values, is evaluated in the evaluation device. The inspection data set can be provided, for example, as an image file, a table, a matrix or a vector field. The inspection image or the inspection data set already created by the image capture unit is transmitted from the image capture unit to the evaluation device. The optical image capture unit is in particular formed as a digital photo camera or digital video camera for capturing the inspection image in two dimensions or as a line scan camera. Insofar as test images are mentioned below in connection with the processing of image information in an evaluation device, the test images can also be referred to as test data sets in this context.

[0012] The evaluation device serves to evaluate the test data sets. It comprises at least one processor and is designed as a central computing unit or as a distributed system comprising at least one processor and at least one storage device at various positions in the machine components. It is therefore a local system, where possible, in which the evaluation is carried out immediately and locally and the results are provided immediately.

[0013] The operating parameters are variable quantities, which correspond to the geometric shape of the separating device or the separating elements of the separating device, the position or orientation relative to the machine frame or the transport elements, the speed of the separating device and / or the motor power. The operating parameters make it possible to determine how or to what extent the separating device interacts with the harvested product or at least one component of the harvested product. In particular, by changing the operating parameters, it is possible to change how much of the contaminants remain in the root vegetables downstream of the separating device in relation to the transport section to which the root vegetables are to be transported inside the machine. The operating parameters are in particular independent of the transport speed of at least the transport elements serving to transport the root vegetables, on which the root vegetables are placed and which move in the same direction as the transport elements.

[0014] The operating parameters define how aggressively the separating device should behave in separating the root vegetables from the contaminants. If it is too aggressive, too many parts of the contaminants are not separated from the root vegetables. If it is too aggressive, not only the contaminants but also the root vegetables are damaged or separated off, which reduces the yield. By generating a separating device setting signal, in particular by sending said signal to the separating device controller, the operating parameters are set, preferably according to the crop or part of the crop depicted by the inspection image. By means of the separating device setting signal, the separating device controller in particular increases or decreases the operating parameters. For this purpose, the separating device controller in particular sends an electrical signal to the preferably actuator. By the separating device controller, in particular by means of the electrical signal, in particular hydraulic, pneumatic, current, voltage, force and / or torque for driving or adjusting the separating device or the operating parameters are matched. In this case, the separating device controller is in particular part of the same computing unit as the evaluation device.

[0015] In this way, a continuous optimization of the operation of the separating device can be achieved, in particular the degree of aggressiveness of the separating elements or of the separating device, so that a gentle handling of the root vegetables and at the same time an effective separation of contaminants from the harvest can be achieved.

[0016] In particular, the evaluation device evaluates the test data sets locally at the machine or at a towing vehicle directly connected to the machine in order to provide a separating device setting signal, whereby almost immediate control is possible when undesirable states are identified in the separating device, and jamming or damage is accordingly avoided.

[0017] According to one advantageous development of the method according to the invention, the evaluation device calculates at least a first proportion of the inspection image formed by at least one image area, which at least partially depicts a defined component of the crop or of the machine, and on the basis of the first proportion, in particular a cleaning characteristic value is calculated.

[0018] Prior to calculating the first proportion, the components which are statistically represented by the first proportion are predefined. The examination image and / or the examination data set are in particular divided into a number of image regions, preferably of equal size. These image regions which at least partially show the components together form the first proportion. This proportion is in particular the proportion of these image regions which at least partially show the components in the total image region, whereby the first proportion is formed by using the ratio of the numbers of the image regions or by using their common area.

[0019] The first proportion is a measure for the size of the image area depicting the components and therefore for the density of the considered proportion of the components or inspection image in the field of view of the image capture unit. The components are in particular at least partially components of root vegetables, whereby the first proportion at least approximately represents the concentration of the root vegetables. If at least 50% to 100% of the area of ​​an image area shows the components, this image area is evaluated in particular as depicting the components and is included in the first proportion. Alternatively, the proportion can be generated by the sum of the pixels depicting the components. In particular, at least one image area can only be included in the first proportion according to a proportion, or preferably also partially in each of the different proportions. This is particularly advantageous when an unambiguous correspondence between the image area and the corresponding components is not possible, preferably within the framework of a model-based classification method. Preferably in this case, probabilities for the correspondence to the different proportions are determined. The image areas are particularly preferably included in the different proportions according to a probability or partially in each of the different proportions. This results in a more accurate depiction of the relationship between the components.

[0020] By calculating the first proportion, a characteristic value is calculated which is particularly representative of the composition of the crop. On this basis, the operating parameters can be controlled in a particularly advantageous manner, since the cleaning performance of the conveying element or of the separating device comprising the conveying element is highly dependent on the composition of the crop. In the case where the first proportion is representative of the concentration of contaminants, the operating parameters can be varied, preferably in such a way that the higher the first proportion, the better the separating action or the separating performance. The cleaning characteristic value is preferably calculated at least with the first proportion or is equal to the first proportion.

[0021] Preferably, at least one image region of the first proportion is identified as representing the defined component, in particular on the basis of an inspection data set generated with the image region, in particular on the basis of inspection values, preferably color information, contained in the inspection image and / or in the inspection sub-data set, in particular the color information comprising black and white values, gray values ​​and / or color channel values ​​of a color space.

[0022] Preferably, the test sub-datasets, the test values ​​or the color information are classified, in particular by a model-based statistical classification method. An image region is therefore counted in particular in the first proportion if the result of the classification method corresponds to a defined component of the crop or the machine. The classification method in particular uses neural networks, random forests, Bayesian classifiers, support vector machines and / or decision trees. By applying this classification method, the calculation results of the first proportions, in particular the various proportions, are particularly robust and informative with respect to the composition of the crop.

[0023] Particularly preferably, the test values ​​or color information are compared with one or more reference values ​​or reference ranges, on the basis of which the image region is or is not included in the first proportion. The reference image can preferably be captured by an optical image capture unit, like the test image, whereby the user must in particular mark different parts of the reference image as representing different components. This type of differentiation ensures a particularly reliable identification of the corresponding components on the test image. Particularly preferably, at least one of the test values ​​of the test sub-dataset, which in particular includes color information, is compared with at least one reference value, and the image region is included in particular in the first proportion if at least one test value of the test sub-dataset is within the assigned reference value range. This reference value range is delimited in particular by a maximum value and a minimum value, whereby preferably the various test values ​​must lie within the respectively assigned reference value range in order for the image region to be included in the first proportion.

[0024] According to one advantageous embodiment of the invention, the evaluation device automatically generates or automatically develops the model on which the classification method is based when exemplary image regions of the reference image, which are to be included in the first proportion, are input. Alternatively or additionally, the evaluation device automatically calculates or modifies at least one reference value range when exemplary image regions of the reference image, which are to be included in the first proportion, are input. In this way, in particular the reference values, the reference value ranges or the model or their model parameters do not have to be predefined manually by the user at least for all of them. Instead, it is sufficient to input at least one exemplary image region that shows the component when the evaluation device is started up. Using this image region, the evaluation device automatically determines at least one reference value, at least one reference value range or the model or its model parameters. In this way, the evaluation device can be independently adapted to a wide range of different application cases. In this case, the more image regions are input, the more accurately the reference values ​​or the reference value ranges or the model or their model parameters can be determined.

[0025] The method is particularly robust if the input image regions show components in different brightness and / or soil conditions. Thus, the method can be reliably applied in different application conditions. Particularly preferably, during repeated execution of the method, possibly with the operator's exemplary identification of relevant components, the evaluation device can adjust at least one reference value or reference value range, thereby drawing training data for the algorithm.

[0026] In particular, the evaluation device automatically expands the scale of the reference data by using a further sensor, such as a luminance sensor, which measures the ambient luminance, which the evaluation device associates with the test data set captured substantially simultaneously in time. Alternatively or additionally, the user of the method, i.e. in particular the driver or operator of the machine or the machine connected thereto, can manually mark at least one component on the visualized test image in order to expand the scale of the reference data of the evaluation device. Thus, based on the instruction given once by the user or based on the data stored in the evaluation device, the evaluation device can distinguish between, for example, potatoes, grass, stones, loose soil and clods, and calculate the respective proportions.

[0027] Preferably, the method according to the invention is carried out automatically after its initiation, with the exception of the possible input of training data in the form of markings of the components, which simplifies operation for the driver or operator of the machine.

[0028] Preferably, the image regions making up the first proportion are additionally identified on the basis of image sub-data sets generated from or formed by the respective adjacent image regions. In particular for this purpose, color information, which is contained in the examination sub-data sets and in particular includes black and white values ​​and / or gray values, is also used. The image regions are thus not only evaluated using the data assigned to them, but also additionally using further data assigned to the surrounding image regions. This allows brightness and / or color shifts to be determined, and thus the identification can be performed on an even broader data basis.

[0029] Preferably, different image areas are weighted differently when calculating the first proportion, so that the image areas contribute different amounts to the first proportion, so that, rather than calculating the first proportion directly using the perspective representation of the inspection image, image areas that represent components of the crop that are further away from the image capture unit can in particular be weighted more heavily than image areas that represent components that are closer to the image capture unit, so that a perspective-corrected first proportion can be formed and thus a particularly realistic representation of the composition of the crop on the transport member can be achieved.

[0030] Preferably, the entire inspection image or a coherent inspection image part is divided into partial image areas. These partial image areas in particular each comprise the same number of pixels of the inspection image, preferably exactly one pixel. An inspection image part is a part or a section of the inspection image which comprises a number of partial image areas. In particular, for calculating the first proportion, only image areas which belong to the inspection image and which represent a proportion are taken into account. For this purpose, in particular, the inspection image is defined in such a way that the inspection image depicts a sensitive zone to be monitored inside the machine. The image area which thus constitutes the first proportion in particular comprises a number of partial image areas of an inspection image part.

[0031] The test image or test image parts are in particular rasterized into a number of sub-image areas, each of which is preferably rectangular. The drawing of the sub-image areas with exactly one pixel creates a particularly large database for evaluating the state of the crop for its individual components, thus allowing a particularly fine control of the individual operating parameters. At the same time, the data volume supplied by conventional 2D digital cameras, which usually amounts to up to several million pixels, can be processed immediately by an evaluation device equipped with one or more modern processors.

[0032] Preferably, the test image comprises a number of test image portions, for which the evaluation device calculates the respective first proportion, in particular a number of proportions in the image area, whereby the test image portions preferably depict the harvest of different transport elements discharging from a separating device. The test image portions thus in particular show different sections of the same transport element, one of which is located in front of the separating device or the separating element of this separating device in the conveying direction, and another section of which is located behind the separating device or the separating element of this separating device. Alternatively, the test image portions show different transport elements (e.g. a transport element for cleaned root vegetables, a transport element for sorted out contaminants) which form alternative conveying paths for different components of the harvest. By calculating the first proportion for these different test image portions, the cleaning or separating performance of the assigned separating device can be evaluated in a particularly comprehensive manner. In particular, the first proportion of the harvest flow towards the separating device can be compared with a first proportion of the downward flow of the root vegetable mixture from the separating mechanism or the separating device, and the efficiency of the separating device can thereby be determined. Alternatively, the composition of the crop of the conveying member following the separating device, i.e. discharging from the separating device, is determined, firstly for the separated and removed crop and secondly for the crop to be conveyed further. Depending on the efficiency of the separating device, in particular the operating parameters are set. Similarly, the examination image portion depicted or present in the individual examination data sets can show the portion of the conveying member in front of the separating or deflecting member of the separating device and the portion of the conveying member behind the separating or deflecting member. If it is found from the image analysis that, for example, an excessively large proportion of root vegetables is stuck in an undesirable area behind the deflecting member, then the deflecting member can be positioned differently, for example lower above the conveying member, which increases the separation performance.

[0033] According to one advantageous embodiment of the invention, the test image portions preferably show various transport elements behind the separating device, in particular a transport element for discharging the root vegetable mixture and a transport element for discharging the impurities behind the same separating device. In this case, a first proportion of the components, for example root vegetables, is preferably determined for both test image portions. Alternatively, different proportions are calculated for the various test image portions. In this way, for example, the proportion of impurities in the downward flow of the root vegetable mixture can be adjusted to the proportion of root vegetables in the flow of the sorted out impurities, on the basis of which the separating elements included in the separating device can be set with respect to their position relative to the transport elements or with respect to their speed.

[0034] Preferably, the image area of ​​the first proportion represents root vegetables or parts thereof, and the image area of ​​the second proportion represents contaminants or parts thereof. The evaluation device therefore calculates at least two different proportions. Particularly preferably, the evaluation device calculates at least four proportions, namely a first proportion for root vegetables, a second proportion for grass components, a third proportion for soil and a fourth proportion for damaged areas. The sum of these proportions is in particular less than or equal to 1. Alternatively, the first proportion can be the proportion of contaminants and the second proportion the proportion of root vegetables. Additionally or alternatively, preferably a proportion for soil clods and / or stones is calculated.

[0035] Alternatively or additionally, in a further development according to the invention, at least two image capture units and at least two transport members are provided, with a first image capture unit taking a first inspection image of the crop portion removed by the first transport member from the separating device and a second image capture unit taking a further inspection image of the crop portion removed by the second transport member from the separating device, and on the basis of at least one, preferably both, of the inspection data sets formed by or generated on the basis of both inspection images, the separating device setting signal being generated, in this case in each case being evaluated, in particular with respect to the individual proportions, as described above or as described below.

[0036] By using a plurality of ratios in the calculation of the evaluation device, a more accurate image of the composition of the crop or the occupancy of the transport element can be achieved. As an alternative to the identification of the image areas using limit values, all image areas of the test image or of the test image part are necessarily assigned to a ratio. In this case, the degree of agreement between the test sub-dataset calculated using the image areas and the reference sub-dataset is preferably evaluated, and each image area is assigned to the ratio with the greatest agreement.

[0037] According to one advantageous embodiment of the invention, the cleaning characteristic value is determined by means of the deviation of the first ratio calculated by the evaluation device from a threshold value, in particular this threshold value representing an optimal operating state of the transport member, whereby a deviation of a defined absolute value from this threshold value triggers a change of the operating parameters. The cleaning characteristic value is in particular based on a number of ratios and preferably further data, in particular sensor data.

[0038] According to one advantageous embodiment of the invention, the separating device setting signal is calculated using a number of cleaning characteristic values, in particular calculated successively in time, or at least one precalculated cleaning characteristic value is involved in the calculation of the cleaning characteristic value. In particular, a moving average value of the cleaning characteristic value is calculated, which forms the basis for the separating device setting signal, or a smoothing of the cleaning characteristic value curve is carried out, in particular by means of a low-pass filter. Through these measures, the method according to the invention becomes particularly insensitive to disturbances and therefore particularly robust to application.

[0039] According to one advantageous embodiment of the invention, at least one further sensor transmits sensor data to the evaluation device, which sensor data participate in the calculation of the separating device setting signal. The sensor is in particular a sensor for measuring the slice thickness of the crop on the conveying member, preferably a tactile or ultrasonic sensor, a sensor for measuring the drive power, for example in the form of a pressure sensor for measuring the hydraulic oil pressure, and / or a rotational speed sensor, in particular for measuring the rotational speed of the conveying member drive. In particular, the slip of the conveying member is determined using the rotational speed sensor, which is transmitted to the evaluation device in the form of sensor data. By using a humidity sensor, further information can be involved in the calculation of the separating device setting signal.

[0040] Based on this further information present in the sensor data, which goes beyond the information provided on the basis of the inspection image, an even more accurate image of the cleaning situation in the area of ​​the transport member is obtained in the evaluation device, which also makes it possible to influence the operating parameters in a more adjusted manner accordingly.

[0041] Preferably, the evaluation device triggers an increase or decrease of the operating parameter by means of various separating device setting signals. In particular, the evaluation device or separating device controller comprises a fuzzy regulator, a PID regulator or a three-point regulator, which alternatively triggers an increase, decrease or maintenance of the current operating parameter. An increase is triggered in particular when the cleaning characteristic value exceeds a first predefined threshold value, and a corresponding decrease is triggered when the cleaning characteristic value falls below a second predefined threshold value.

[0042] Preferably, the operating parameter is the distance between two conveying elements, which can act together as a separating device, or the separating elements of the separating device or the separating device and the conveying elements. In particular, the operating parameter is the distance between two conveying rollers which rotate during operation on a roller platform, between which the soil is screened off. Alternatively, the operating parameter is the distance between a conveying element formed as a screening belt and a separating element formed as a deflecting roller, which extends across the conveying element and acts to deflect the root vegetables to the side from the conveying element. In this case, the deflecting roller rotates during operation about a rotation axis which forms an angle of less than 90° with the conveying direction of the conveying element when viewed from above. Alternatively, the separating element is formed as a finger belt which circulates during operation, which is located above the conveying element and whose outwardly projecting fingers comb the crop arranged on the conveying element during operation. Alternatively, the separating elements are also formed as scraper devices that do not rotate during operation and are arranged above the grass belt that cooperates with the sorting belt and serve to scrape the root vegetables from the grass accumulated on the grass belt. The spacing can be set by means of an adjustment device, particularly hydraulic or mechanical, so that the degree of aggressiveness of the separating elements of the separating device in cooperation with the conveying element or the separating performance of the conveying element can be changed particularly simply.

[0043] Alternatively, the operating parameter is the penetration depth of at least one digging blade of the machine into the ground, which makes it possible in a simple manner to influence the amount of contaminants in the harvested crop.

[0044] Alternatively or additionally to what has been said so far, the operating parameter is the separation speed, in particular the circulation speed or rotation speed of the separating device or the separating element of the separating device. In particular, the separation speed is the circulation speed of the aforementioned finger belt or the rotation speed of the aforementioned deflection roller. Alternatively, the separation speed is the circulation speed of a separating device, for example in the form of a fine grass belt, which is angled and transports the contaminants upstream during operation, and which is driven in such a way that the contaminants are transported upstream as much as possible and the root vegetables roll downstream against the direction of movement of the section of the separating device facing them.

[0045] Preferably, the operating parameter is alternatively defined as the angle of attack of the conveying member, of the separating device, i.e. of at least one separating element of the separating device. In particular, the operating parameter is the angle of attack of the separating device, also called fine grass lifter. By means of this angle of attack, the inclination of the conveying plane of the fine grass belt of the separating device is changed relative to the horizontal direction, thus setting the degree of aggressiveness of the separating device.

[0046] According to an alternative preferred embodiment of the invention, the operating parameters affect the change in the air flow rate or the air mass transport rate per hour, and the corresponding separation device separates based on the air flow. In this case, the corresponding operating parameter can be, for example, the motor power represented by the motor speed. This air also separates the root vegetables from the impurities, and in particular the grass is blown away from the harvest stream and thus removed. In particular, in the case of such air separation devices that can also be used in a stationary state, the operating parameter can preferably be the speed of the associated blower or the angle of attack of an associated device, for example in the form of an air baffle, which divides the air flow into a main air flow and a cross air flow.

[0047] In accordance with one advantageous embodiment of the method according to the invention, in particular, several of the operating parameters mentioned above are set by the same or different separating device setting signals, and for this purpose, a rule set can be stored in the evaluation device, which generates corresponding signals for the desired increase or decrease in performance of the individual separating devices for the individually settable quantities.

[0048] Preferably, after triggering an operating parameter change, no further operating parameter changes are triggered for a defined period or over a defined conveying section of the conveying member. This applies in particular only to the same operating parameters and / or to at least one operating parameter of at least one separating device arranged downstream during operation. This ensures that no overadjustment of the separating member occurs and that any operating parameter changes are based on a sound data basis that already takes into account previously made operating parameter changes.

[0049] Preferably, the separating device setting signal is transmitted to the separating device controller by wire, in particular by means of a CAN bus or Ethernet, or wirelessly, whereby the separating device setting must preferably be authorized in advance by the operator via an input in the interface, whereby an existing or at least established communication transmission system can be utilized for configuring the separating element, and the reliability of the method can be increased in particular by the fact that the resulting or to be implemented setting of the separating device is displayed to the operator, in particular in the driver's cabin, instead of its automatic setting, and must be authorized via a corresponding input in the interface.

[0050] The above-mentioned object is further achieved according to the invention by a machine for harvesting root vegetables and / or separating them from further harvest contaminants, the machine having a machine frame, a transport member, an image capture unit, a separating device and an evaluation device, the machine being configured to carry out the method described above or below.

[0051] The evaluation device is preferably a graphic processor unit, in particular a processor unit based on a GPU (Graphical Processing Unit) or a GPGPU (General Purpose Graphical Processing Unit) and / or an FPGA (Field Programmable Gate Array). By configuring the evaluation device in this way, the test data sets can be evaluated in a particularly resource-saving manner, in particular locally. Obviously, the evaluation device configured as an electronic data processing device has further conventional means, for example for a current supply, an interface and a main memory.

[0052] According to one advantageous embodiment of the invention, the machine has at least one sensor coupled to the evaluation device, in particular a contact or ultrasonic sensor for measuring the slice thickness of the crop on the conveying member, a sensor for measuring the drive power, for example a pressure sensor for measuring the hydraulic oil pressure, and / or a rotational speed sensor arranged on the conveying member. These sensors allow the conveying speed signal to be calculated also on the basis of the measured physical quantities, which significantly increases the information-generating capacity of the quantities calculated by the evaluation device and reduces their vulnerability to errors. Similarly, a humidity sensor can additionally provide information that contributes to the setting of one or more of the separation devices within the framework of the analysis of the evaluation device.

[0053] Preferably, the machine has a number of image capture units for capturing at least one inspection image of the same transport member during operation. Alternatively, the machine preferably has a number of image capture units for capturing at least one inspection image of different transport members during operation, in particular discharged from one and the same separating device. Alternatively, one of the two image capture units can be directed towards the discharge area of ​​the separating device, in which separation is performed by airflow, for example. By means of the image capture units, the composition of the harvested product can be tracked, in particular the progression of the first proportion along the transport section, in particular in different parts of the transport section of the machine behind the separating device. In particular, the transport speeds of the different transport members can thus be set with different first proportions. According to the corresponding methods described above or described below, an analysis of the transport section areas captured by the individual inspection images is carried out in at least one evaluation device. Preferably, only one central evaluation device is provided for evaluating the data of the image capture units, but it is also possible to assign a specific evaluation device to each individual image capture unit. These evaluation devices can then control the respective assigned separation devices, in particular in coordination with further evaluation devices. Alternatively or additionally, a central evaluation device is responsible for setting the separation device setting signals and forwarding these signals to the machine controller.

[0054] Preferably, the image capture unit is arranged such that the inspection image shows at least two alternative transport paths of the entire transport section for the various harvest components, in particular for root vegetables and for contaminants. This allows the image capture unit to be used to monitor two transport members, with each inspection image portion of the inspection image depicting a section of the various transport members or the harvest on these transport members. In particular, one of the transport members is designed to transport the sorted out contaminants and another of the transport members is designed to transport the cleaned root vegetables. This allows a particularly comprehensive image of the cleaning performance and thus the operating state of the transport members and / or the separating device including the transport members to be captured.

[0055] Preferably, the image capture unit is arranged such that during operation the inspection image at least partially depicts at least two transport member sections separated by a separating device, in particular by a separating member of a separating device. These transport member sections are only separated by the separating member of the separating device in the display of the inspection image and are respectively included in the transport member. The separating member is closer to the image capture unit than the transport member, so that the transport member is covered by the separating member in the inspection image. Such a positioning of the image capture unit makes it possible to calculate at least a first proportion for each of the two individual inspection image sections and thus directly evaluate the effectiveness of the separating member and the assigned separating device. In particular for this purpose, the composition of the harvest before reaching the separating member is compared with the composition of at least one proportion of the harvest after passing through the separating member.

[0056] Preferably, the transport element is configured as a sorting belt or needle belt, which in operation passes in particular under at least one deflection roller, which runs across the transport element and deflects the harvested material to the side therefrom. Alternatively, the transport element is configured as a star-shaped sorting element or transport roller, in which case the transport roller is in particular included in a roller carrier.

[0057] Alternatively or additionally to what has been described above, the machine is configured as a machine for separating root crops from further crops, whereby according to one advantageous development the machine is in particular stationary, i.e. driven without continuous local forward movement of the machine during operation.

[0058] According to a further advantageous development, the machine is a potato digger or a turnip digger.

[0059] Further details and advantages of the invention can be gleaned from the diagrammatically illustrated embodiments described below. [Brief description of the drawings]

[0060] [Figure 1] 3 is a program flow chart of a method according to the present invention; [Diagram 2] 4 is an accompanying program flow chart for determining the separator setting signal. [Diagram 3] 4 is an accompanying program flow chart involving the processing of cleaning characteristic values; [Figure 4] 13 is an accompanying program flow chart for setting up the separation device. [Diagram 5] 1A to 1C are diagrams showing the appearance of an inspection image and its partial evaluation. [Figure 6] FIG. 1 is a diagram showing the object of protection according to the present invention. [Figure 7] 7A to 7C are diagrams showing the object of protection according to FIG. 6 as viewed from different sides; [Figure 8]7A to 7C are diagrams showing the object of protection according to FIG. 6 as viewed from different sides; [Figure 9] FIG. 7 is a partial view of the object of protection according to FIG. 6 together with the transport member; [Figure 10] FIG. 10 is a detailed view of one area of ​​the device according to FIG. 6, which is partially shown in FIG. [Figure 11] 11 shows the object of protection according to FIG. 10 from a different perspective; [Figure 12] FIG. 11 shows an inspection image of the image capture unit according to FIG. 10 . [Figure 13] FIG. 7 shows a separation device of the machine according to FIG. 6 together with an image capture unit. [Figure 14] FIG. 14 shows a schematic inspection image taken from the perspective of the image capture unit shown in FIG. 13. [Figure 15] FIG. 7 shows a further separating device of the machine according to FIG. 6 together with an image capturing unit. [Figure 16] FIG. 16 shows a diagrammatically drawn inspection image taken from the perspective of the image capture unit shown in FIG. 15. [Figure 17] FIG. 7 is a further detailed view of the machine according to FIG. 6 with a further image capture unit. [Figure 18] FIG. 18 shows a schematic view of an inspection image from the point of view of the image capture unit according to FIG. 17; [Figure 19] FIG. 2 is a detailed view showing a further device according to the invention. [Figure 20] FIG. 4 is a further detailed view showing a further device according to the invention.

[0061] Wherever appropriate, identical or identically acting parts are provided with the same reference symbols. The individual technical features of the exemplary embodiments described below can also be used to produce developments of the invention, but always in combination with the features of at least one of the independent claims. The subject matter mentioned in the brief description of the drawings is often only partially shown in the individual drawings.

[0062] The method according to the invention is used for controlling the operation of a machine 2 for harvesting root vegetables 4 (see Figures 6 to 8). In this method, at least one inspection image 8 is taken by at least one optical image capture unit 6, which inspection image 8 shows a harvested product, including root vegetables 4, which is advanced relative to a machine frame 12 of the machine 2 by means of at least one transport member, generally designated for the present by the reference numeral 10.

[0063] The inspection image 8 is transmitted to an evaluation device, which generates a separating device setting signal based on the inspection data set generated with or formed by the inspection image 8, for setting at least one operating parameter of the separating device of the machine 2. The drawings depicted as inspection images only show the parts relevant to the invention in a schematic manner, without any possible frames or borders. The images captured by the camera, in particular digital ones, may possibly contain further information that is not depicted in the drawings. This information can, for example, already be masked or filtered on the camera side or when setting or processing the inspection data set.

[0064] In the embodiment according to the invention, starting from the flow 1.1, the composition of the harvested product is evaluated in front of the first separating element (block 1.2) (FIG. 1) using the method described above. In addition to this, the composition of the harvested product is also calculated before the inlet and after the outlet of the further separating device (blocks 1.3 and 1.4). Finally, again at the inlet of the third separating device, the harvested product is also determined for its composition. This results in the individual root vegetable and mixture proportions A1 and A2 (blocks 1.6, 1.7, 1.8 and 1.9). Depending on the desired separation performance of the individual separating devices, the individual root vegetable or mixture proportions A1 and A2 for the individual separating devices are subtracted from each other in the evaluation device (block 1.10). Depending on this, the operating parameters of the individual separating devices are set in order to optimize the performance of the individual separating devices (block 1.11).

[0065] The determination of the separation device setting signal is shown in more detail in FIG. 2. Starting from the inspection image 8, an extraction of the relevant inspection image portion is first carried out in order to set up the inspection data set (block 2.1). For this purpose, based on the position of the image capture unit, a mask or region of interest (ROI) can be predefined (block 2.2), by means of which the intervals of the inspection image that should be taken into account and those that should not be taken into account are differentiated. Based on the relevant image sections of the inspection image and the inspection data set now provided for processing, a calculation of the proportion of the image area that shows the individual crop components is then carried out (block 2.3). For this purpose, in particular, color information can be evaluated. These values ​​can be taken from a look-up table or can also be predefined by the operator (block 2.4).

[0066] Based on the threshold definition (block 2.5), the deviation of the calculated proportion from the threshold value is calculated (block 2.6). The threshold value is, for example, an ideal value for the respective proportion considered (for example, root vegetables, contaminant 1, contaminant 2). This is followed by low-pass filtering (block 2.7) for the purpose of smoothing the determined deviation. In this case, the filter time constant defined according to block 2.8 is used. A cleaning characteristic value RS is then calculated based on the smoothed values ​​of the deviation for the individual positions along the conveying section and the individual proportions (block 2.9). In this case, simplified regulator parameters can be used, according to which, for example, the cleaning characteristic value RS_1 (A1) is set to 1 for the proportion A1 in the form of root vegetables in the first monitored conveying section if the deviation is too large and therefore there are too many contaminants in the inspection image (part) considered. If the deviation from the ideal value is sufficiently low, the cleaning characteristic value RS for the proportion A1 can be set to 0. Based on the cleaning characteristic value RS_1 (A1) in block 2.9, a separation device setting signal can be generated (block 2.10), for example by means of a three-point regulator.

[0067] Thus, for example, by using the method according to the invention, the ratio of the products present in the flow in front of the separating device and then behind the separating device is evaluated separately, for example on the output conveyor for the product, i.e. the root vegetables (4), and on the output conveyor for the impurities (5). If separate cameras monitor both conveyors, separate areas of interest are defined for the conveyor that outputs the impurities and the conveyor that outputs the root vegetables. Depending on whether an excess of potatoes (root vegetables 4) is present on the conveyor that outputs the impurities or whether an excess of impurities is present on the conveyor that outputs the product, the parameters of the separating device that affect the separation threshold are adapted accordingly. In particular, the fingers of the finger belt or the brushes of the brush belt can therefore be set higher or lower and / or the finger belt or the brush belt can be run slower or faster.

[0068] In the case of an exemplary implementation with a three-point regulator system according to FIG. 3, the cleaning characteristic values ​​RS_1(A2)-RS_n(A2) are first considered, i.e. the cleaning characteristic values ​​for the proportion A2 representing impurities (block 3.1). For example, this means considering the proportion of impurities in the crop in the form of stones, either before and after a separating device or before and after a conveying section which includes several separating devices. The cleaning characteristic values ​​are "0" or "1" according to the above notation. Then, in block 3.2, it is checked whether the sum of the cleaning characteristic values ​​RS_i(A2), with i=1 to m, is zero. If this is not the case, then, in block 3.3, first of all, an interrogation of the memory device 3.4 is carried out for each position for the last cleaning characteristic value and, if applicable, the associated crop composition. If, as a result of the check in block 3.5, it is found that the separating device settings were changed long ago, then, in block 3.6, a decreasing operating parameter signal is generated for the individual separating devices. In this case, the cleaning characteristic values ​​called up from the memory device by interrogation can also be taken into account. The setting of the separating device or devices (block 3.7) is made less aggressive so that a smaller proportion of root vegetables remains in the mixture stream or in the harvest stream in the conveying section. The final cleaning characteristic value or values ​​RS_i (A2) are entered in the memory device 3.4, possibly together with the respective associated proportion A2 (block 3.15).

[0069] If the sum of the cleaning characteristic values ​​RS_i(A2) (with i=1 to m) is zero, then in block 3.8, attention is paid to the cleaning characteristic values ​​RS_i(A1) corresponding to the proportion A1, for example, root vegetables, or their values ​​are queried accordingly. Then, it is checked whether the sum of these cleaning values ​​is also zero (block 3.9). If so, the last cleaning characteristic value RS_i(A1) is also entered in the memory 3.4 (block 3.10). No change in the operating parameters is necessary, and a neutral operating parameter signal or no operating parameter signal is sent (block 3.11). If the sum of the cleaning values ​​for the proportion A1 is not zero, then in 3.12, the previous values ​​are again called up from the memory by interrogation, and in block 3.13, it is checked whether a considerable amount of time has already passed since the last operating parameter change. If this is the case, in 3.14, operating parameters for a more aggressive setting of the separating device are sent.

[0070] In order to realize the preset according to Fig. 3 in block 3.7, the separating device actuators must be controlled. For this purpose, according to Fig. 4, in the control unit 4.1, based on the preset in blocks 3.6, 3.11 and 3.14, and taking into account the current settings of the individual separating devices (4.2), a change in the separating device-specific operating parameters is determined. For this purpose, for example, the speed, the number of rotations, the distance or the inclination of the separating device parts are defined. In block 4.3, the manipulated variables for the separating device actuators of the individual separating devices are defined, and the separating devices are configured via these manipulated variables (block 4.4).

[0071] FIG. 5 shows an example of an inspection image 8 showing the transition from the transport member 10A to the transport member 10B at the top of the figure. Root vegetables 4 and possible contaminants 5, including stones and grass, are present in this transport section area. According to a classification defined in the training of the algorithm or predefined via a database, for example a table with color information in HSV format, the individual partial areas 16 are inspected for the presence of identical components. In this way, based on the correspondence between the individual image areas and the individual proportions, a percentage distribution of the individual proportions of root vegetables 4 and contaminants 5 in the inspection image 8 is generated, as illustrated in the lower right corner of FIG. 5. Thereby, A1 denotes the proportion of root vegetables 4 in the inspection image 8 or in the corresponding inspection data set, A2 the proportion of grass and A3 the proportion of stones. The correspondence is preferably made based on the color information of the individual pixels, i.e. the image area 19 which is associated with a certain proportion corresponds in particular to the area of ​​one pixel. The cleaning characteristic value, commonly designated by the reference character RS, is illustratively, but preferably, based on the deviation of the first proportion A1 from a threshold value R, which represents an optimal proportion distribution of the root vegetables 4 at the observation location of the transport section. For example, the cleaning characteristic value RS is set to 1 if there is a deviation of more than 50% from the cleaning value and to 0 if there is a deviation of less than 50% from the cleaning value. This value is then stored accordingly or processed in further program sequences according to FIGS. 1 to 4.

[0072] According to Fig. 6, the machine 2 according to the invention is configured as a towed potato digger, in which a number of conveyor members 10 and the separating devices assigned to them are held on a machine frame 12, which is only partially shown. Along the conveying section, a number of image capture units 6 are provided, which photograph the crop, including the root crops 4, transported on the conveyor members 10. The positions for the image capture units 6 shown in Fig. 6 are the transition from a first conveyor member 10A in the form of a sorting belt onto a second conveyor member 10B in the form of a sorting belt additionally surrounded by a grass coarse belt, and from this second sorting belt 10B to a further conveyor member 10C which comprises a further separating device. In addition to this, on the outlet side of this separating device, the conveyor member 10E leading to the sorting table is monitored by a further image capture unit 6, with a further conveyor member 10F provided for remaining impurities 5, in particular stones, being captured at the same time.

[0073] The evaluation device can be positioned at any centrally accessible location, but preferably near the sorting table. Via cable 12.1, which can be seen for example in FIG. 6, a road speed signal or information about the settings of the separating device can be sent from the evaluation device to the towing vehicle.

[0074] The machine depicted in side view in Figures 7 and 8 can be equipped with optical image capture units 6 in further positions. Additional image capture units can therefore be arranged in the immediate vicinity of the area of ​​the digging device 29 or the area of ​​the step up to the storage bin 33.

[0075] 9 and 10 show the arrangement of the optical image capture unit 6, which is located on the frame side above the first step between the transport members 10A and 10B, with its field of view directed downwards. A light source 7 provides illumination of the field of view in order to capture a well-lit inspection image 8. The transport member 10A is a sorting belt, which already sifts out part of the contaminants 5, in particular soil and / or soil clods, when coming from the digging device 29 and transfers them via the step to a further transport member 10B formed as a sorting belt. This transport member 10B additionally has a coarse grass belt, which is provided for removing grass present next to the potatoes or in the crop. Accordingly, scraper devices 32 are arranged over the entire width of the transport member 10B.

[0076] The height H of the scraper device 32 above the conveying plane of the conveying member 10B can be set by a separating device setting signal. This means that it is possible to influence the separating performance of the separating device formed as a grass belt. In addition to this, the relative speed of the sorting belt with respect to the coarse grass belt 43 can be set. For the sake of clarity, only the coarse grass belt 43 is shown in FIG. 10, but the actual conveying member 10B formed in the form of a sorting belt is not shown (see FIG. 12).

[0077] An inspection image 8 obtained from the field of view of the optical image capture unit 6, represented by a dashed line in Fig. 11, is depicted in detail in Fig. 12. An inspection data set created from or formed by this inspection image 8 is used to perform the aforementioned evaluation based on the deviation of the proportion of captured and classified objects from a threshold value R.

[0078] Leaving the transport member 10B, the still present harvest material is transferred in a transport direction 1C to a further transport member 10C, to which is assigned a separating device in the form of a number of rotating deflection rollers 24 located one above the other. Via the impact exerted by this, the harvest material is conveyed in the direction of the transport member 10D (FIG. 13).

[0079] The distance H between the transport member 10C and the lower deflection roller 24 can be set in order to vary the separation performance and therefore constitutes a settable operating parameter. If necessary, the further distance between the individual deflection rollers 24 can be changed with respect to their mutual distance for the strength of the deflection or possibly to produce a separating function in which grass is pulled out between the deflection rollers 24. Alternatively or additionally, a change in the separation performance or deflection results from the fact that the circulation speed of the deflection rollers 24 can be set.

[0080] Likewise, the height H of the lower ends of the fingers 26 of the separating device formed as finger belts 26.1 and belonging to the transport member 10D can be set as one of several operating parameters. The height H represents the distance between the fingers 26 and the upper edge of the transport member formed as a needle belt. In addition to this, the angle of attack of the finger belts 26.1 with respect to a perpendicular to the transport plane of the transport member can be configured to be settable. The same applies to the circulation speed of the finger belts 26.1.

[0081] The image capture unit 6 shown in Fig. 13 generates an inspection image 8 shown in Fig. 14, in which an inspection image portion 8A relevant in the present example is defined by means of filtering or masking. In order to monitor the performance of the separating device, in this case the separating performance of the deflecting roller 24, an inspection image portion 8B can additionally be selected which is located behind the deflecting roller 24 as seen in the conveying direction 1C. In particular for the separating device setting, the area in front of the deflecting roller 24 is monitored. An inspection data set then results from the corresponding inspection image portion 8A.

[0082] If the associated cleaning characteristic value RS for the test image portion 8A results in an excessively low separation performance of the upper or illustrated separating device, this separating device can be set to be more aggressive. Alternatively, if the cleaning characteristic value in the test image portion 8B represents an excessively high separation performance, for example due to a too high proportion of contaminants 5 in the form of clods behind the diverting roller 24, which are still at least partially required for protecting the potatoes on the subsequent transport section, the distance H of the diverting roller 24 from the transport member 10 can be reduced, whereby the separating device can be set to be less aggressive.

[0083] 15 and 16 show a further optical image capture unit 6 arranged in the region of the conveyor belts 10C and 10D. This image capture unit 6 can be used supplementarily or alternatively to the image capture unit 6 according to FIG. 6. In particular, this image capture unit serves to monitor the function of the separating and deflecting device formed by the deflecting rollers 24. This monitoring unit is also assigned a light source 7 in order to improve the illumination of the monitored area.

[0084] A further optical image capture unit 6 with an assigned light source 7 is arranged above the sorting table with a line of sight directed towards the transport members 10E and 10F (FIG. 17). By masking, the test image portions 8A and 8B depicted in the test image 8 according to FIG. 18 are selected, which monitor, on the one hand, the transport member 10E for conveying the root vegetables 4 in the conveying direction 1E as one transport path, and, on the other hand, the transport member 10F for conveying the impurities 5 in the form of stones and / or clods in the conveying direction 1F as a further transport path. By means of the aforementioned evaluation, it is checked whether there is too much root vegetables 4 on the transport member 10F. If this is the case, then by using the method according to the invention, the separating device connected upstream is set up to be even sharper. This separating device is located above the transport member 10D formed as a needle belt and has fingers 26, which are in particular exemplarily shown as finger belts with dashed lines, but which in the illustrated drawing are arranged behind a cover 40 located in front of the fingers 26. The distance of the fingers 26 from the conveying member 10D is narrowed so that, for example, more of the harvest in the form of root vegetables 4 is conveyed onto the conveying member 10E via the associated chute 41. If too many impurities 5 in the form of stones and / or clods are identified on the conveying member 10E, for example, the circulation speed of the deflecting rollers 24 can be reduced so that less impacts act on the impurities 5 in order to better deflect any stones that may occur in the direction of the conveying member 10F. In that way, the impurities 5 slide down better via the chute 42 towards the conveying member 10F.

[0085] Fig. 19 shows the use of the method according to the invention in a further separating device, here in the form of a grass fine lifter. The inspection image 8 shown in Fig. 19 accordingly shows a conveying member 10K which enters in the conveying direction 1K and is guided via a step towards a separating device formed as a grass fine lifter and with a grass belt 30. This separating device continues to convey the fine grass possibly present in the harvest stream in the direction 31, while the potatoes and heavier contaminants of the harvest stream that reach the grass belt 30 fall back into the gap between the conveying member 10K and the grass fine lifter due to the inclination position of the grass belt 30, which can be preset via a settable angle α. If in the inspection image portions 8A and 8B the proportion of root vegetables 4 or impurities 5 is outside the standard value, the separating device can be set to be more aggressive or correspondingly less aggressive, for example by setting the angle of attack α or by changing the circulation speed of the grass belt 30, thereby resulting in the desired separation performance.

[0086] In the soil removing roller according to FIG. 20, which is preferably used in the field of storage technology, the distance H between the individual transport members 10T is made adjustable. The transport members 10T thus simultaneously constitute a separating device. In the individual test image portions 8A and 8B of the test image 8, the proportion of contaminants is preferably calculated and used to set the performance of the separating device. In this case, a perspective correction based on the "fisheye" display of the image capture unit 6 is performed, as described above.

[0087] Further alternative embodiments of the invention can be implemented, for example, as a self-propelled turnip digger or as a cleaning section in potato or turnip storage technology.

Claims

1. A method for controlling the operation of a machine (2) for separating root vegetables (4) from a contaminating harvest, comprising the steps of:

1. A method for capturing, by means of at least one optical image capture unit (6), at least one inspection image (8) of at least a portion of said harvested crop being advanced relative to a machine frame (12) by at least one transport member (10, 10A, 10B, 10C, 10D, 10E, 10F, 10K, 10T), comprising: an evaluation device generating a separation device setting signal for setting at least one operating parameter of a separation device of the machine (2) based on at least one color information contained in the inspection image (8); By setting the operating parameters, the separation performance of the separation device between root vegetables and impurities is changed. A method comprising:

2. 2. The method according to claim 1, wherein the evaluation device evaluates at least one color information contained in the inspection image (8) locally at the machine (2) or at a directly connected towing vehicle.

3. 3. The method according to claim 1 or 2, wherein the evaluation device calculates at least a first proportion (A1) of the inspection image (8) formed from at least one image area (19), the at least one image area (19) at least partially depicting a defined component of the crop or the machine (2).

4. 4. The method of claim 3, further comprising calculating a cleaning characteristic value (14) based on said first ratio (A1).

5. 5. The method according to claim 3 or 4, further comprising identifying said at least one image area (19) comprising said first proportion (A1) as indicative of said defined component of said crop or said machine (2).

6. 6. The method according to claim 5, further comprising identifying the at least one image area (19) that constitutes the first proportion (A1) as representing the defined component of the crop or the machine (2) based on at least one color information of the image area (19).

7. 7. The method according to claim 3, further comprising classifying at least one color information of the image area (19) by a statistical classification method and classifying an image area (19) if the result of the classification method corresponds to the defined component of the crop or the machine (2).

8. 8. The method according to claim 7, further comprising classifying at least one color information of the image region (19) by a model-based statistical classification method, and classifying an image region (19) if a result of the classification method corresponds to the defined component of the crop or the machine (2).

9. 8. The method according to claim 7, further comprising classifying an image area (19) into said first proportion (A1) if the result of said classification method corresponds to said defined component of said crop or said machine (2).

10. 10. The method according to claim 3, further comprising the steps of: comparing at least one color information of the image region (19) with at least one reference value (R); and including an image region (19) at least if the at least one color information of the image region (19) is within an associated reference value range.

11. 11. The method according to claim 10, further comprising the step of: counting an image area (19) into the first proportion (A1) if at least one color information of the image area (19) is within an associated reference value range.

12. 10. The method according to claim 7, wherein the evaluation device automatically develops a model underlying the classification method and / or automatically calculates or modifies at least one reference value range when exemplary image regions (19) of a reference image, which are to be included in the first proportion (A1), are input.

13. 13. Method according to any one of claims 3 to 12, characterized in that different image regions (19) are weighted differently when calculating the first proportion (A1).

14. 14. The method according to claim 3, wherein the entire inspection image (8) or a coherent inspection image portion (8A) is divided into a plurality of sub-image regions (16).

15. The method of claim 14, wherein the sub-image regions (16) each include the same number of pixels.

16. The method of claim 15, wherein each of the sub-image regions (16) comprises exactly one pixel.

17. 17. The method according to claim 3, wherein the inspection image (8) comprises a plurality of inspection image portions (8A, 8B), for each of the plurality of inspection image portions (8A, 8B) the evaluation device calculates a first proportion (A1).

18. 18. The method according to claim 17, wherein for the plurality of examination image portions (8A, 8B), the evaluation device calculates a plurality of proportions (A1, A2, A3) in an image area (19).

19. 18. Method according to claim 17, characterized in that the inspection image portions (8A, 8B) depict the crops of different transport members leaving one separating device.

20. 20. The method according to claim 3, wherein the image area (19) constituting the first proportion (A1) represents a root vegetable (4) or a part of the root vegetable (4) and the image area constituting the second proportion (A2) represents an impurity (5) or a part of the impurity (5).

21. 5. The method according to claim 4, further comprising determining the cleaning characteristic value (14) by means of the deviation of the first proportion (A1) calculated by the evaluation device from a threshold value (R).

22. 10. The method according to claim 3, wherein the separating device setting signal is calculated using a plurality of calculated cleaning characteristic values ​​(RS) or wherein at least one pre-calculated cleaning characteristic value (RS) is involved in the calculation of the cleaning characteristic value (RS).

23. 23. The method according to claim 22, wherein the separating device setting signal is calculated using a plurality of cleaning characteristic values ​​(RS) calculated successively in time or at least one precalculated cleaning characteristic value (RS) is involved in the calculation of the cleaning characteristic value (RS).

24. At least two image capture units (6) and at least two conveying members (10) are provided; a first image capture unit (6) captures a first inspection image of the harvest portion removed by the first conveying member (10) from the separating device and a second image capture unit (6) captures a further inspection image of the harvest portion removed by the second conveying member (10) from the separating device; generating the separation device setting signal based on at least one of at least one color information included in both of the inspection images; 24. The method according to any one of claims 1 to 23.

25. 25. The method according to claim 1, wherein at least one further sensor transmits sensor data which are involved in the calculation of the separating device setting signal to the evaluation device.

26. 26. The method according to claim 25, wherein the at least one further sensor is an ultrasonic or contact sensor for measuring a slice thickness of the crop on the transport member, a sensor for measuring the drive power, a humidity sensor and / or a rotation speed sensor.

27. 27. The method according to claim 1, wherein the evaluation device uses various separation device setting signals to trigger an increase or decrease of the operating parameter.

28. 28. The method according to claim 27, further comprising the step of: after triggering an operating parameter change, not triggering a further operating parameter change for a defined period of time or over a defined transport section of the transport member (10, 10A, 10B, 10C, 10D, 10E, 10F, 10K, 10T).

29. 29. The method according to claim 1, wherein the operating parameters are the distance (H) between two conveying members (10T), the distance (H) between a separating member (32) or the separating device and the conveying member or a further conveying member (10C).

30. 30. The method according to any one of the preceding claims, wherein the operating parameter is a separation speed of a separating member (24, 26, 30) or of the separating device.

31. 31. The method of claim 30, wherein the operating parameter is the circulation or rotation speed of the separating member (24, 26, 30) or the separating device.

32. The method according to any one of the preceding claims, wherein the operating parameter is an angle of attack (α) of the conveying member (10) or the separating device.

33. 33. The method according to claim 1, wherein the operating parameter is a motor power and / or an angle of attack of an associated device.

34. 34. The method of claim 1, wherein the separation device configuration signal is transmitted to a separation device controller by wire or wirelessly.

35. 35. The method of claim 34, wherein the isolation device configuration signal is communicated to the isolation device controller by wire using a CAN bus or Ethernet.

36. 35. The method of claim 34, wherein the configuration of the separation device must be previously authorized by an operator via an input in an interface.

37. 37. The method according to any one of claims 1 to 36, wherein the separating device setting signals and / or the operating parameters are plotted for an operator and / or used to automatically set the operating parameters.

38. 38. The method of claim 1, wherein the optical image capture unit captures only one or two dimensional information.

39. A machine (2) for separating root vegetables (4) from a contaminated harvest, comprising: The machine comprises at least a machine frame (12), a transport member (10), an image capture unit (6), a separating device and an evaluation device and is adapted to carry out the method according to any one of claims 1 to 38. Machine (2).

40. 40. The machine (2) of claim 39, wherein the evaluation device comprises a graphics processor unit.

41. The machine (2) of claim 40, wherein the evaluation device comprises a GPGPU and / or FPGA based processor unit.

42. 42. Machine (2) according to any one of claims 39 to 41, characterized in that at least one sensor is provided which is coupled to the evaluation device.

43. 43. The machine (2) according to claim 42, wherein the at least one sensor coupled to the evaluation device is an ultrasonic or contact sensor for measuring the slice thickness of the crop on the conveying member (10), a sensor for measuring the drive power, a rotational speed sensor arranged on the conveying member (10) and / or a humidity sensor.

44. 44. The machine (2) according to any one of claims 39 to 43, further comprising a plurality of image capture units (6) for capturing at least one inspection image (8) of the same conveying member (10) or the same separating device or of different conveying members (10) or separating devices, respectively, during operation.

45. 45. The machine (2) according to claim 44, further comprising a plurality of image capture units (6) for taking at least one inspection image (8) of each of the conveying members (10) or separating devices which, during operation, are delivered from one and the same separating device.

46. A machine (2) according to any one of claims 39 to 45, wherein the image capture unit (6) is arranged such that, during operation, the inspection image (8) at least partially depicts each of at least two conveying member sections separated by at least one separating member (24, 32).

47. A machine (2) according to any one of claims 39 to 46, wherein the image capture unit (6) is arranged such that the inspection image (8) shows a transport member (10) of the machine (2) forming at least two alternative transport paths for various harvested product components.

48. 48. The machine (2) according to claim 47, wherein the image capture unit (6) is arranged such that the inspection image (8) shows a conveying path for root vegetables (4) and a conveying path for contaminants.

49. The machine (2) according to any one of claims 39 to 48, wherein the conveying member (10) is formed as a sorting belt (10A, 10B, 10E), as a needle belt (10C, 10D), as a star-shaped sorting member (10P, 10Q, 10S) or as a digging blade.

50. 50. The machine (2) of claim 49, wherein the conveying member (10) during operation travels beneath at least one diverting roller (24) extending across the conveying member (10) and diverting the harvested material from the conveying member (10).

51. 47. The machine (2) according to any one of claims 39 to 46, wherein the transport member (10) is formed as a transport roller (10T) included in a roller bed (10M).

Images