METHOD FOR CONTROLLING A SOIL TREATMENT DEVICE

Description

[0001] The present invention relates to a method for controlling a soil cultivation device according to the preamble of claim 1. The invention further relates to a soil cultivation device according to the preamble of claim 11.

[0002] A soil cultivation device is used for the mechanical removal and / or shredding of vegetation between crops planted in adjacent rows. The soil cultivation device comprises working tools arranged at row intervals on a support frame for removing and / or shredding vegetation in a transverse direction to the rows, which act on the vegetation present between the crop rows.

[0003] To detect and differentiate between crops and weeds, the tillage equipment includes at least one detection device. Image data of the soil, crops, and / or weeds within the row are acquired by means of at least one optical detection device. The image data generated by the at least one optical detection device is evaluated by a control device associated with the tillage equipment. The control device then directs the movements of the tillage tools to remove the weeds.

[0004] Currently, crop rows are detected using conventional RGB (red-green-blue) cameras, stereovision cameras, or 3D cameras. The generated image data is analyzed to distinguish plants from the soil. Differentiating between soil and crops is difficult with some crop species that have a bluish tint or a thick waxy coating, which significantly reduces the detectable green component. Stereovision cameras and 3D cameras are limited in their ability to distinguish between soil, vegetation, and crops when dealing with small crops or low-growing plants. Various soil cultivation devices with detection equipment, or methods for controlling them, are disclosed, for example, in DE102021133723A1, US10455826B2, and GB2598012A.

[0005] Based on this, the invention aims to further develop a method for controlling a soil cultivation device and a soil cultivation device in such a way that an improved distinction between soil and crop is achieved in order to increase the accuracy of the control of the cultivation tools.

[0006] This problem is solved by a method for controlling a soil cultivation device with the features of independent claim 1. Furthermore, the problem is solved by a soil cultivation device with the features of independent claim 11. Advantageous embodiments and further developments are described in the respective dependent claims.

[0007] According to claim 1, a method for controlling a soil cultivation device for removing and / or shredding vegetation between crops planted in rows is proposed, using cultivation tools arranged transversely to the rows at a row spacing on a support frame. Image data of the soil, the crops, and / or the vegetation are acquired by means of at least one optical detection device, and the image data are evaluated by a control device associated with the soil cultivation device. The control device adjusts the position of the cultivation tools for removing the vegetation relative to at least one of the rows depending on the evaluation. According to the invention, at least one detection device is a camera that acquires image data in the red-green-near-infrared spectrum (RG-NIR spectrum) excluding the blue spectrum.

[0008] The invention is based on the idea of ​​using the NIR spectrum instead of the blue spectrum to identify crops characterized by a bluish hue or a thick wax layer that is only permeable to a small amount of green light. This allows for differentiation between soil, vegetation, and crops, as well as determining their location, shape, and growth. The differentiation between soil and crops is of particular importance. This determination forms the basis for controlling the processing tools, thereby increasing the accuracy of the mechanical removal and / or shredding of vegetation between crops planted in rows.This method utilizes the finding that the reflection of light in the NIR spectrum by photosynthetic plants is many times higher than the reflection of light in the visible green spectrum. This results in a significantly improved distinction between crop plants and soil, as well as an improvement in differentiating between crop plants and vegetation.

[0009] In particular, a triple bandpass filter can be placed in front of the camera's image sensor to filter out wavelengths in the blue spectrum. This allows an RGB camera to be used as the optical detection device, which is more cost-effective than a stereovision or 3D camera. Since the operating principle differs when using a stereovision or 3D camera, a completely different detection result can be achieved with the RGB camera and its triple bandpass filter. The cost advantage is particularly pronounced when the tillage equipment is designed for the parallel processing of multiple rows of crops. Accordingly, the tillage equipment can include multiple optical detection devices to monitor several rows being processed.

[0010] Preferably, a hyperspectral vegetation index, in particular the Triangular Vegetation Index (TVI), can be used to differentiate between soil and crops and / or vegetation. The vegetation index allows the determination of the chlorophyll content of crops and vegetation to distinguish them from the soil. The Triangular Vegetation Index (TVI) is calculated as the area of ​​a hypothetical triangle in spectral space connecting the green peak reflectance, the minimum of chlorophyll absorption, and the NIR shoulder. If chlorophyll absorption causes a decrease in red reflectance and the density of leaf tissue causes an increase in NIR reflectance, the total area of ​​the triangle increases.

[0011] According to a training course, a threshold method can be used to binarize the captured image data. Binarization transforms a color image, which is converted to grayscale in an intermediate step, into a binary image. This binary image is then used to differentiate between vegetation and cultivated plants. For this purpose, an operator can specify a size for the cultivated plant using an input / output unit, which serves as a reference point for differentiating between vegetation and cultivated plants.

[0012] A suitable binarization algorithm can be stored, or stored, in a memory unit of the control device. A simple binarization method is the threshold method. The Otsu method is cited as an example of a threshold method.

[0013] In particular, threshold values ​​used by the threshold method can be automatically preset by the control device and / or manually set and / or adjusted by an operator of the tillage equipment. While automatic presetting of threshold values ​​for the threshold method simplifies or reduces the operator's workload, manual presetting or adjustment by the operator offers the possibility of adapting the threshold values ​​used in the threshold method to existing environmental conditions during the tillage process in the field, or taking these conditions into account.

[0014] Preferably, depending on the evaluation of the image data, the lateral position of the support frame, which carries several processing tools and is oriented transversely to the rows of planted crops, can be adjusted by the control device. This allows for more precise guidance of the soil cultivation device along the rows.

[0015] According to a training course, the lateral position of cultivation tools assigned to a row of crops can be adjusted based on the evaluation of the image data. Adjusting the lateral position of these tools allows for individual tracking of each row. Different lateral distances between adjacent rows can thus be compensated for by individually adjusting the lateral position of the tools working in each row. Furthermore, the required safety distance to the crops can be minimized, enabling the removal of vegetation growing close to the crop.

[0016] The lateral adjustment of machining tools can be performed independently of each other.

[0017] In addition to a common lateral adjustment of the position of all machining tools by the lateral displacement of the support frame, a row-dependent displacement of individual machining tools can also be carried out by a corresponding control by the control device.

[0018] Preferably, depending on the evaluation of the image data, the movement sequences of one machining tool within a pair of machining tools can be modified independently of the other tool. This refers to the independent adjustment of an oscillating movement of the machining tools. The oscillating movement of each machining tool can be a pendulum motion around a substantially horizontal axis. Alternatively, the oscillating movement of each machining tool can also be a pivoting movement around a substantially vertical axis. A combination of pendulum and pivoting movements is also conceivable, depending on the design of the machining tools.

[0019] The problem initially set out is further solved by a row-dependent soil cultivation device according to claim 11.

[0020] According to claim 11, a soil cultivation device for removing and / or shredding vegetation between crops planted in rows is proposed. The soil cultivation device comprises cultivation tools arranged transversely to the rows at row spacing on a support frame, and a control device associated with the soil cultivation device, which is designed and configured to evaluate image data of the soil, the crops and / or the vegetation acquired by at least one optical detection device arranged on the soil cultivation device and, depending on the evaluation of the image data, to control the position of the cultivation tools for removing the vegetation relative to at least one of the rows.According to the invention, the at least one optical detection device is designed as a camera which is configured and set up to acquire image data in the red-green-near-infrared spectrum (RG-NIR spectrum) to the exclusion of the blue spectrum. Reference may be made to the advantages of the method according to the invention for controlling the row-dependent soil cultivation device.

[0021] The soil cultivation device can be equipped with tillage tools designed to remove and / or shred vegetation between the rows of crops. Additionally or alternatively, the soil cultivation device can be equipped with tillage tools designed to remove and / or shred vegetation within a row of crops.

[0022] In particular, a triple bandpass filter can be placed in front of the camera's image sensor, filtering out wavelengths in the blue spectrum. This triple bandpass filter removes the blue channel but allows the red, green, and NIR channels to pass through, so that the red, green, and NIR spectra are captured by the camera's image sensor. A significant advantage of using the triple bandpass filter is that an RGB camera can be used as the optical detection device, which is more cost-effective. This is particularly advantageous for a soil cultivation device, which typically works more than one row simultaneously, as each row is preferably assigned its own optical detection device to control the movements of the cultivation tools individually, i.e., row-specifically.Such an optical detection device can be used in particular for the more efficient differentiation of cultivated plants belonging to the onion family and some cabbage species, e.g. cauliflower, which have a low reflection in the green spectrum due to the thickness of an outer wax layer surrounding them.

[0023] For this purpose, the triple bandpass filter can filter the incoming wavelengths, preferably allowing only wavelengths in the ranges between 500 nm and 550 nm, between 600 nm and 660 nm, and between 800 nm and 850 nm to pass through the triple bandpass filter. The quantum efficiency for the three spectra to be detected is most suitable in these wavelength ranges.

[0024] Preferably, at least one camera can be an RGB camera. In particular, the RGB camera configuration with a triple bandpass filter enables cost-effective and more precise detection and differentiation between soil, vegetation, and cultivated plants.

[0025] In particular, the control device can be configured to carry out the method according to any one of claims 1 to 10. For this purpose, the control device can comprise a storage unit and a processing unit, which is designed and configured to execute image evaluation algorithms stored or that can be stored in the storage unit.

[0026] According to a further aspect, the use of a soil cultivation device according to one of claims 11 to 14, which is designed and configured for carrying out a method according to one of claims 1 to 10, is claimed, wherein the crop plants to be detected have an increased blue content or a wax layer that is largely opaque to wavelengths in the green spectrum and / or belong to the onion family.

[0027] The present invention is explained in more detail below with reference to an embodiment illustrated in the drawings. Fig. 1 schematically and exemplarily shows a front view of a soil cultivation device in simplified representation; Fig. 2 schematically and exemplarily shows a top view of the soil cultivation device according to Fig. 1Fig. 3 schematically and by way of example a perspective view of a tool carrier unit of a soil cultivation device according to one embodiment; Fig. 4 schematically and by way of example an optical detection device designed as an RGB camera; Fig. 5 on the left by way of example an image of a soil section taken with an RGB camera and on the right by way of example an image of the soil section with the RGB camera and a triple bandpass filter in front of it; Fig. 6 on the left by way of an analysis of the corresponding image carried out using the Excess Green Index according to Fig. 5 to distinguish between soil and plants on it, and on the right an image of an analysis of the corresponding image carried out using the Triangular Vegetation Index according to Fig. 5to distinguish between soil and the plants growing on it; Fig. 7 on the left shows a binary image of an analysis of the corresponding image on the left side carried out by a binarization method according to Fig. 6 and on the right-hand side a binary image of an analysis of the corresponding image on the right-hand side carried out by the binarization method according to Fig. 6 , each to distinguish between the soil and the plants growing on it; and Fig. 8 on the left a binary image according to Fig. 7 and on the right side a binary image according to Fig. 7 , each after a size evaluation to distinguish between vegetation and cultivated plants.

[0028] In Fig. 1Figure 1 schematically illustrates a simplified front view of a soil cultivation device 1 for removing and / or shredding vegetation between crops 2 planted in a row PR. The soil cultivation device 1 is, in particular, a hoeing machine used for the mechanical control of vegetation between crops 2 planted in rows PR. Tool carrier units 6 are arranged essentially equidistantly on frame elements 7 extending transversely QR. The frame elements 7 are arranged on a multi-part profile frame 9 running parallel to them. The profile frame 9 can be hinged. For the sake of simplicity, only three tool carrier units 6 are shown in the illustration.

[0029] The representation in Fig. 2 schematically and exemplarily shows a top view of the soil cultivation device 1 according to Fig. 1The profile frame 9 is connected to a base frame 11 by parallelogram elements 10. An actuator 12, preferably a linear actuator, enables lateral movement of the profile frame 9 relative to the base frame 11 by displacing one of the parallelogram elements 10. The tool carrier units 6, arranged on the frame elements 7, are moved laterally by the profile frame 9. Wheels 8 are arranged on the base frame 11.

[0030] The soil cultivation device 1 comprises cultivation tools 3 arranged transversely QR to the row and a control device 4 associated with the soil cultivation device 1. The control device 4 is designed and configured to evaluate image data acquired by at least one optical detection device 5 arranged on the soil cultivation device 1 with regard to the crop plants 2 and / or vegetation, and, depending on the evaluation of the image data, to control the position, in particular the lateral position, of the cultivation tools 3 for removing the vegetation relative to the rows PR. The lateral movement of the profile frame 9 relative to the base frame 11 allows for adjustment of the position of the cultivation tools 3 relative to the rows PR with crop plants 2.

[0031] At least one optical detection device 5 is assigned light sources 14. The light sources 14 can be designed, in particular, as LED spotlights. The optical detection device 5 has a detection range 13, which points in the direction of travel FR. This enables the anticipatory detection of the cultivated plants 2 arranged in a row PR.

[0032] In the illustrated embodiment, at least one optical detection device 5 is arranged in the outer edge region of the soil cultivation device 1. A further optical detection device 5 can be arranged on the opposite side of the soil cultivation device 1. This improves the accuracy of the control of the lateral movement of the profile frame 9 relative to the base frame 11.

[0033] Each processing tool 3 is assigned at least one protective element 15. The protective element 12 is oriented essentially parallel to the PR row. The protective element 15 prevents soil from being thrown onto the crop plants 2 by the processing tool 3.

[0034] The processing tools 3 of the soil cultivation device 1 can be designed differently. In the case of the Fig. 1 and Fig. 2 In the illustrated soil cultivation device 1, which is used for the mechanical control of vegetation between the crops 2 planted in rows PR, the cultivation tools 3 are fixedly attached to the tool carrier unit 6. The cultivation tools 3 can be designed, for example, as hoe blades, angle blades, or goosefoot shares. Combined arrangements of hoe blades and angle blades on the tool carrier unit 6 are also conceivable.

[0035] An alternative embodiment of the soil cultivation device 1 is shown in Fig. 3 depicted. In Fig. 3 A row-dependent soil cultivation device 1 is shown schematically and exemplarily, which is additionally set up for removing and / or shredding vegetation between crop plants 2 planted in the row PR.

[0036] The soil cultivation device 1 comprises cultivation tools 3 arranged transversely QR to the row PR, as well as cultivation tools 36 oscillating transversely to the row PR. The tool carrier units 6 are laterally adjustable by an actuator, i.e., transversely to the direction of travel FR. For this purpose, the tool carrier units 6 can be arranged on a frame element 7 of the soil cultivation device 1 extending transversely to the direction of travel FR. The frame element 7 of the soil cultivation device 1 can be connected by at least one Fig. 3Actuator 12, not shown in the image, here preferably two linear actuators, designed to be laterally displaceable.

[0037] The tool carrier units 6 are equipped with the processing tools 3 described above, which are used to remove and / or shred vegetation between the rows PR of crops 2. Additionally, oscillating processing tools 36 are provided, which remove and / or shred vegetation occurring within the respective row PR.

[0038] The control device 4 is designed and configured to evaluate image data acquired by the at least one optical detection device 5 arranged on the soil cultivation device 1 with regard to the crop plants 2 and / or the vegetation and, depending on the evaluation of the image data, to control the movement sequences of the cultivation tools 36 for the removal of the vegetation in addition to the lateral position of the cultivation tools 3 and 36.

[0039] Viewed in the direction of travel FR, behind the machining tools 3, the oscillating machining tools 36 are arranged on both sides of a longitudinal beam of the tool carrier units 6. Each machining tool 36 has a blade 37 that can be cyclically pivoted about a pivot axis perpendicular to the ground by means of a drive.

[0040] Each of the machining tools 36 is assigned a drive, the drives being independently controllable by the control device 4. The drives are here, and preferably, designed as double-acting hydraulically or pneumatically actuated cylinders.

[0041] Each knife 37 has a substantially arc-shaped or crescent-shaped contour. Viewed in the longitudinal direction, i.e., the direction of travel FR, the knives 37 are arranged one behind the other at intervals. This prevents collisions from occurring when the machining tools 36 perform their oscillating pivoting motion.

[0042] To control the double-acting, hydraulically or pneumatically actuated cylinders that oscillate the processing tools 36, each cylinder is assigned a valve block, which is controlled by the control device 4. The translational movement generated by the respective cylinder is transformed into a rotational movement of the pivot axis by an intermediate kinematic mechanism that connects the cylinder to the pivot axis of the processing tool 36. The respective pivot angle at which the processing tool 36 is pivoted is independently adjustable and can be changed during operation of the soil cultivation device 1.

[0043] In Fig. 4A simplified schematic and exemplary representation of an optical detection device 5 designed as an RGB camera 20 is shown. The RGB camera 20 comprises a lens 21, an image sensor 22, and an interface 23 for data transmission to the control device 4. The RGB camera 20 can optionally be equipped with processing electronics 24, which can be used for image processing.

[0044] A triple bandpass filter 25 is connected upstream of the image sensor 22 of the RGB camera 20. The triple bandpass filter 25 is designed and configured to filter out wavelengths in the blue spectrum. This triple bandpass filter 25 filters out the blue channel, but allows the red and green channels, as well as the NIR channel, to pass through, so that the red, green, and NIR spectra are captured by the image sensor 22 of the RGB camera 20. For this purpose, the triple bandpass filter 25 can filter the incoming wavelengths, preferably allowing only wavelengths in the ranges between 500 nm and 550 nm, between 600 nm and 660 nm, and between 800 nm and 850 nm to pass through the triple bandpass filter 25.

[0045] The control device 4 comprises an image processing system which serves to convert and evaluate the image data. The process of image data evaluation by the control device 4 is described below using the following example: Figs. 5 to 8 Explained step by step.

[0046] In Fig. 5 On the left is an example image 26 of a soil section 27 taken with the RGB camera 20, and on the right is an example image 28 of the same soil section 27 taken with the RGB camera 20 and the image sensor 22 with a triple bandpass filter 25 in front of it. Images 26 and 28 are a comparison of images of cauliflower plants as cultivated plants 2.

[0047] The representation in Fig. 6 The left side shows a grayscale image 29 of an analysis of image 26 carried out using the Excess Green Index (ExG) according to Fig. 5 to distinguish between soil and plants on it, and on the right a grayscale image 30 of an analysis of the corresponding image 28 carried out using the Triangular Vegetation Index (TVI) according to Fig. 5to distinguish between soil and the plants growing on it. This comparison of the analyzed images 29, 30 already shows a significant deviation in the quality of the distinction between soil and plant, i.e., vegetation and cultivated plant 2.

[0048] In Fig. 7 On the left side is a binary image 31 of an analysis of the corresponding grayscale image 29 on the left side carried out by a binarization method according to Fig. 6 and on the right-hand side a binary image 32 of an analysis of the corresponding grayscale image 30 on the right-hand side carried out by the binarization method according to Fig. 6, each to distinguish between the soil and the plants on it 2, 33. Using the binarization method, the grayscale image 30 analyzed using the Triangular Vegetation Index (TVI) is converted into the binary image 32 or black and white image, whereby only the elements identified as plants on the soil are shown.

[0049] The representation in Fig. 8 The left side shows a binary image 34, which is based on the size evaluation of the left binary image 31 according to Fig. 7 based and on the right side a binary image 35, which is based on the size evaluation of the right binary image 32 according to Fig. 7This is based on the following: For this purpose, an operator can specify a size for the crop 2 using an input / output unit 16, which serves as a benchmark for differentiating between the vegetation 33 to be removed and / or shredded and the crop 2, which are identified in the binary image 32 shown on the right by the image evaluation. It is also conceivable to select from a database of different crop 2 stored or stored in a memory unit 17 of the control device 4.

[0050] The comparison of the two binary images 31 and 32 in Fig. 7 illustrates the qualitative difference in the differentiation between the soil and the vegetation on it 33 and the cultivated plants 2.

[0051] Crucially, the use of the RGB camera 20 with a preceding triple bandpass filter 25 as at least one detection device 5, which captures image data in the red-green-near-infrared spectrum (RG-NIR spectrum) while excluding the blue spectrum, significantly improves the differentiation between crops exhibiting a bluish hue and soil compared to the classic method of acquisition using an RGB camera and subsequent image data analysis using the Excess-Green Index. Examples of crops with a bluish hue include onions, leeks, and some types of cabbage. Cauliflower and other cabbage varieties have a thicker wax layer than other crops such as corn or sugar beets.As a result, the reflection in the green spectrum is usually weaker, so that the detection of such crops using image data that represents the crop in the RG-NIR spectrum is significantly improved.

[0052] The binarization of the in Fig. 6 The image data contained in grayscale image 30 is processed using a thresholding method. Binary image 32 is generated from a color image that was converted to grayscale image 30 in an intermediate step. Binary image 32 is used to differentiate between vegetation 33 and crop 2. For this purpose, an operator can specify a size for crop 2 using the input / output unit 16, which serves as a reference point for differentiating between vegetation 33 and crop 2.

[0053] A suitable algorithm for binarization can be stored or stored in the memory unit 17 of the control device 4, which can be executed or is executed by a computing unit 18 of the control device 4. The Otsu method is given as an example of a threshold method.

[0054] An angle sensor is arranged on the profile frame 9 and / or on at least one of the parallelogram elements 10, from whose measurement signals the current position, i.e. the lateral displacement of the profile frame 9 relative to the base frame 11, can be determined.

[0055] In particular, threshold values ​​used by the threshold method can be automatically preset by the control device 4 and / or manually set or adjusted by an operator of the soil cultivation device 1. While automatic presetting of threshold values ​​for the threshold method simplifies or relieves the operator, manual presetting or adjustment by the operator offers the possibility of adapting the threshold values ​​used in the threshold method to existing environmental conditions during the cultivation process in the field or taking these into account.

[0056] Depending on the evaluation of the image data, the control device 4 adjusts the lateral position of the frame element 7, which carries several processing tools 3, 36 and is oriented transversely to the row of planted crops 2. For this purpose, the control device 4 activates at least one actuator 12 to move the frame element 7 laterally.

[0057] Furthermore, depending on the evaluation of the image data, the control device 4 can adjust the lateral position of processing tools 36 assigned in pairs to a row of crop plants 2. For this purpose, the control device 4 controls the drives of the processing tools 36, in particular independently of each other.

[0058] According to another aspect, depending on the evaluation of the image data, the movement sequences of one processing tool 36 of a pair of processing tools 36 can be changed independently of the other processing tool 36. Reference symbol list 1 Soil cultivation equipment 34 Binary image 2 cultivated plant 35 Binary image 3 Machining tool 36 Machining tool 4 Control device 37 Knife 5 Detection device 6 Tool carrier unit 7 Frame element FR Direction of travel 8 wheel PR Row 9 Profile frame QR transverse direction 10 Parallelogram element 11 Basic frame 12 actuator 13 Detection area 14 light bulbs 15 Protective element 16 Input-Output Unit 17 Storage unit 18 computing unit 19 20 RGB camera 21 lens 22 Image sensor 23 interface 24 Processing electronics 25 Triple bandpass filter 26 Picture 27 floor section 28 Picture 29 grayscale image 30 grayscale image 31 Binary image 32 Binary image 33 vegetation

Claims

1. A method for controlling a soil cultivation device (1) for removing and / or chopping vegetation (33) between crop plants (2) planted in rows by means of processing tools (3, 36) arranged, at a row distance in a transverse direction (QR) to the rows, on a frame element (7), wherein image data of the ground, the crop plants (2) and of the vegetation (33) are detected by means of an optical detection device (5) and the image data are evaluated by a control device (4) assigned to the soil cultivation device (1), through which the position of the processing tools (3, 36) for removing and / or chopping the vegetation (33) is controlled relative to at least one of the rows (PR) depending on the evaluation, characterised in that as at least one detection device (5) a camera (20) is used, through which image data are detected in the red-green-near infrared spectrum (RG-NIR spectrum) under the exclusion of the blue spectrum.

2. The method according to Claim 1, characterised in that a triple band pass filter (25) is connected upstream of an image sensor (22) of the camera (20), by way of which wavelengths in the blue spectrum are filtered out.

3. The method according to Claim 1 or 2, characterised in that for distinguishing between ground and crop plants (2) and / or vegetation (33), a hyper-spectral vegetation index, in particular, the triangular vegetation index (TVI) is used.

4. The method according to any one of the Claims 1 to 3, characterised in that binarisation of the captured image data is carried out by means of a thresholding method.

5. The method according to Claim 4, characterised in that threshold values used by the thresholding method are automatically set by the control device (4) and / or manually set and / or adapted by an operator of the soil cultivation device (1).

6. The method according to Claim 4 or 5, characterised in that the threshold values used by the thresholding method are determined by the control device (4) depending on the type of the planted crop plants (2) and / or manually set by an operator of the soil cultivation device (1).

7. The method according to any one of the Claims 1 to 6, characterised in that depending on the evaluation of the image data the lateral position of the frame element (7) carrying multiple processing tools (3) oriented transversely to the rows (PR) of the planted crop plants (2) is adapted by the control device (4).

8. The method according to any one of the Claims 1 to 7, characterised in that the lateral position of processing tools (3, 36) assigned to a row (PR) with crop plants (2) is adapted depending on the evaluation of the image data.

9. The method according to Claim 8, characterised in that the lateral adaptation of processing tools (36) is carried out independently of one another.

10. The method according to any one of the Claims 1 to 9, characterised in that depending on the evaluation of the image data, the movements of a processing tool (36) of a pair of processing tools (36) is changed independently of the other processing tool (36).

11. The soil cultivation device (1) for removing and / or chopping vegetation (33) between crop plants (2) planted in rows, including processing tools (3, 36) arranged at a row distance in the transverse direction (QR) to the rows on a frame element (7), and a control device (4) assigned to the soil cultivation device (3), which is embodied and equipped for evaluating image data of the ground, the crop plants (2) and / or the vegetation (33) detected by at least one optical detection device (5) arranged on the soil cultivation device (1) and depending on the evaluation of the image data control the position of the processing tool (3, 36) for removing the vegetation (33) relative to at least one of the rows (PR), characterised in that the at least one detection device (5) is designed as a camera (20), which is embodied and equipped for capturing image data in the red-green-near infrared spectrum (RG-NIR spectrum) with the exclusion of the blue spectrum.

12. The soil cultivation device (1) according to Claim 11, characterised in that a triple band pass filter (25) is connected upstream of an image sensor (22) of the camera (20), which filters out wavelengths in the blue spectrum.

13. The soil cultivation device (1) according to Claim 11 or 12, characterised in that the at least one camera (20) is embodied as RGB camera.

14. The soil cultivation device (1) according to any one of the Claims 11 to 13, characterised in that the control device (4) is equipped for carrying out the method according to any one of the Claims 1 to 10.

15. Use of a soil cultivation device (1) according to any one of the Claims 11 to 14, which is embodied and equipped for carrying out a method according to any one of the Claims 1 to 10, characterised in that the crop plants (2) to be detected have an elevated blue component or a wax layer which for wavelengths in the green spectrum is largely impermeable, and / or belong to the family of bulbous plants.