DE102022212171A1 - Method for classifying plants in and/or between rows of plants in an agricultural area - Google Patents

Abstract

The invention relates to a method for classifying plants (12) in rows of plants (14) of an agricultural area, wherein in each case a coverage index and/or a second one for defined evaluation areas (20) arranged between rows of plants (14) is determined in an image information (10) by means of the computing unit and the identified evaluation area plants (18), which extend from one of the rows of plants (14) into the respective defined evaluation area (20), are classified by means of the computing unit depending on the determined coverage index.

Description

State of the art

The invention is based on a method for classifying plants in and/or between rows of plants in an agricultural area and a corresponding computing unit, a plant identification unit and an agricultural working machine according to the class of the independent claims. The present invention also relates to a corresponding computer program and a machine-readable storage medium.

The EN 10 2017 210 804 A1 discloses a method for applying a spray to a field, whereby the spray is applied depending on the degree of coverage of an evaluation area. When evaluating the image data, plants are segmented using a threshold in the NDVI. The crop rows are identified from this segmentation. A crop row hose of a certain width is placed around the crop rows. Every segmented object (plant) between the hoses is classified as a weed by definition. All objects that lie within this hose width of the rows or are connected to objects within this hose width are classified as a crop plant by definition.

Other algorithms use networks (deep learning) or classification methods to recognize plants. This requires a lot of training data, a lot of label data and offline training beforehand. In addition, these methods are very computationally intensive.

Disclosure of the invention

• - Empfangen einer Bildinformation von einem mittels einer optischen Erfassungseinheit erfassten Feldabschnitt einer landwirtschaftlichen Fläche mit Pflanzen;
• - Identifizieren von Pflanzen in der Bildinformation mittels einer Recheneinheit;
• - Identifizieren von Pflanzenreihen in der Bildinformation unter Verwendung der identifizierten Pflanzen mittels der Recheneinheit;
• - Identifizieren von Auswertebereichs-Pflanzen aus der Gruppe der identifizierten Pflanzen mittels der Recheneinheit, wobei die Auswertebereichs-Pflanzen sich in der Bildinformation von einer identifizierte Pflanzenreihe bis in einen zwischen den Pflanzenreihen angeordneten definierten Auswertebereich erstrecken;
• - Ermitteln jeweils einer Bedeckungskennzahl für die definierten Auswertebereiche, wobei die Bedeckungskennzahl
  • ◯ eine Anzahl von zusammenhängenden Pflanzenteilen der identifizierten Auswertebereichs-Pflanzen in dem jeweiligen definierten Auswertebereich und/oder
  • ◯ eine Gesamtbedeckung des jeweiligen definierten Auswertebereichs von Pflanzenteilen der identifizierten Auswertebereichs-Pflanzen in der Bildinformation repräsentiert, mittels der Recheneinheit; und
• - Klassifizieren der identifizierten Auswertebereichs-Pflanzen in Abhängigkeit von der ermittelten Bedeckungskennzahl des jeweiligen definierten Auswertebereichs mittels der Recheneinheit.

The present invention relates to a method for classifying plants in and/or between rows of plants in an agricultural area, comprising the steps:

• - Receiving image information from a field section of an agricultural area with plants captured by means of an optical detection unit;
• - Identifying plants in the image information using a computing unit;
• - Identifying rows of plants in the image information using the identified plants by means of the computing unit;
• - Identifying evaluation area plants from the group of identified plants by means of the computing unit, wherein the evaluation area plants extend in the image information from an identified row of plants to a defined evaluation area arranged between the rows of plants;
• - Determine a coverage index for each of the defined evaluation areas, whereby the coverage index
  • ◯ a number of contiguous plant parts of the identified evaluation area plants in the respective defined evaluation area and/or
  • ◯ a total coverage of the respective defined evaluation area of plant parts of the identified evaluation area plants is represented in the image information, by means of the computing unit; and
• - Classifying the identified evaluation area plants depending on the determined coverage index of the respective defined evaluation area by means of the computing unit.

The method described above can in particular be a computer-implemented method for classifying plants in and/or between rows of plants in an agricultural area using the steps described above.

The present invention further relates to a computing unit which is configured to execute and/or control the steps of a previously described method.

The present invention also relates to a plant identification unit with an optical detection unit for detecting a field section of an agricultural area with plants in order to obtain image information from the detected field section, and a previously described computing unit.

The present invention further relates to an agricultural working machine, in particular an agricultural field sprayer, with an agricultural working tool, in particular a spraying device, and a plant identification unit as described above, wherein the working tool, in particular the spraying device, can be controlled depending on the classified evaluation area plants.

The present invention also relates to a computer program which is designed to carry out and/or control the steps of a method described above when the computer program is executed on a computer, as well as a machine-readable memory medium on which the computer program is stored.

The method according to the invention now makes it possible to classify individual plants extending from a row of plants across this row into the intermediate area, but in particular also several overlapping plants and/or parts of plants, in a very simple and resource-saving manner, without the need for an image database or the like. According to the invention, this is done by determining a coverage index for defined evaluation areas arranged between the respective rows of plants. The coverage index represents a number of connected plant parts of identified evaluation area plants in a defined evaluation area and/or a total coverage of the respective defined evaluation area of plant parts of the identified evaluation area plants.

This does not involve the use of a conventional classification method that performs clustering based on plant-specific properties and then makes a decision about which class (crop or weed) an object most likely belongs to. A classifier is a method that ultimately calculates a probability of how well an object belongs to a class.

Instead, the method makes use of agronomic knowledge in such a way that, depending on the characteristics of the crops and/or weeds (growth characteristics, number of leaves, leaf size, etc.), it is possible to estimate how these will extend into the area between the rows of plants (e.g. driving lane), i.e. what the coverage index should look like in the evaluation area. This means that even thin but widely branched weeds (e.g. tendrils) can be classified as weeds, since with such typical weed species the number of connected plant parts to be expected in the evaluation area is large.

• - Es resultierte eine verbesserte Klassifizierung der Pflanzen, insbesondere Kulturpflanzen und/oder Beikräuter.
• - Es ist keine Sammlung von vielen Bilddaten und kein Anfertigen von Labels notwendig, um ein Netz oder Klassifikation zu trainieren, wodurch der Gesamtaufwand stark reduziert wird.
• - Es ist kein Offline-Training anhand von Bilddaten im Vorfeld notwendig.
• - Die erforderliche Rechenzeit für die Klassifizierung ist sehr gering.
• - Die Information über die klassifizierten Pflanzen, insbesondere Kulturpflanzen und Beikräuter in dem Kulturpflanzenbereich kann für darauffolgende Schritte genutzt werden, um Pflanzenreihen besser bzw. genauer zu identifizieren, da die identifizierten bzw. erkannten Beikräuter hierfür unberücksichtigt bleiben bzw. „ignoriert“ werden können.

This offers a number of advantages:

• - This resulted in an improved classification of plants, especially cultivated plants and/or weeds.
• - There is no need to collect a lot of image data or create labels to train a network or classifier, which greatly reduces the overall effort.
• - No offline training using image data is necessary beforehand.
• - The computing time required for classification is very low.
• - The information on the classified plants, in particular cultivated plants and weeds in the cultivated plant area can be used for subsequent steps to better or more accurately identify plant rows, since the identified or recognized weeds can be disregarded or "ignored" for this purpose.

An agricultural area can be understood as an area used for agriculture, a cultivation area for plants or even a plot of such an area or cultivation area. The agricultural area can therefore be arable land, grassland or pasture. The plants include cultivated plants or useful plants whose fruit is used agriculturally, for example as food, animal feed or as an energy crop, as well as weeds or weeds.

The method can comprise a step of detecting a field section of an agricultural area with plants using the optical detection unit. The detection step can be carried out during a crossing or a flight of the plant identification unit or the agricultural work machine. At least one further step of the method, in particular all steps of the method, can be carried out during a crossing or a flight of the plant identification unit or the agricultural work machine.

The plant identification unit can comprise a mobile unit or be arranged on a mobile unit, wherein the mobile unit can be designed in particular as a land vehicle and/or aircraft and/or trailer. The mobile unit can also be a self-propelled or autonomous robot. The plant identification unit is preferably part of an agricultural work machine. The agricultural work machine is preferably a weed control machine, in particular an agricultural field sprayer. The agricultural work tool is preferably a spraying device, but can also be a mechanical tool for weed control.

The field section can be a detection section or a detected image section of an optical detection unit. The image information can be, for example, an image of the detected field section. An optical detection unit can be understood to be, for example, a camera or a 3D camera or an infrared detection unit. The optical detection unit can be calibrated to calculate the height assignment from captured images, for example.

The method comprises a step of identifying plants in the image information or the image using a computing unit. The plants in the image information are understandably pixels, in particular connected pixels with corresponding pixel values in the image information or the image, which characterize or represent a plant object from the real world in the image. The plant object can be selected from the group consisting of: a single plant, a single plant part, several overlapping plants and/or plant parts, plant mass, biomass and combinations thereof. The step of identifying plants can comprise identifying a color component, in particular a red color component and/or an infrared component in the image information or the image. Here, plants can be detected using the optical detection unit, for example by evaluating light spectra or using a predetermined NDVI value (Normalized Differenced Vegetation Index, which is formed from reflection values in the near infrared and visible red wavelength range of the light spectrum), by distinguishing plant mass or biomass from the (earth) soil. Identifying plants also includes determining the presence of plants or parts of plants or plant mass or biomass in the field section, in particular without (initially) classifying the individual plants.

The method further comprises a step of identifying rows of plants in the image information using the identified plants by means of the computing unit. The rows of plants are plants, in particular cultivated plants, sown in rows, i.e. essentially linearly. Accordingly, the rows of plants are preferably rows of cultivated plants. Therefore, by definition, it can be assumed (initially) that plants in the row of cultivated plants are cultivated plants and plants between the cultivated plants are weeds.

The rows of plants or rows of cultivated plants extend essentially in a straight line across the agricultural area and thus also in the image information or the image. The identification of the rows of plants or rows of cultivated plants is preferably carried out using at least one of the following pieces of information: color component, in particular red color component of plants in the recorded field section, infrared component of plants in the recorded field section, plant spacing, plant row spacing, growth stage of the plants, geocoordinates of a sowing of the plants. Using this information, the rows of plants can be easily identified since, for example, cultivated plants are usually planted equidistantly along rows or the cultivated plants are further along in terms of growth stage than the weeds. In the step of identifying the rows of plants, all rows of plants in the image information or the image are preferably identified.

Since the plant rows extend essentially in a straight line, the identification of a plant row can be done, for example, by fitting a straight line or a straight plant row line into an image trajectory with a calculated NDVI index or the highest NDVI value (normalized distance or difference between red and infrared/NIR). The plant row lines can thus be considered part of the plant rows.

Preferably, in the step of identifying the plant rows, a plant row line is generated or created in each of the plant rows. The plant row line can either have been generated for identifying the plant rows or can be generated afterwards.

The method also includes a step of identifying evaluation area plants from the group of identified plants using the computing unit, wherein the evaluation area plants extend in the image information from one of the identified plant rows to a defined evaluation area arranged between the plant rows. Preferably, the evaluation area plants extend in the image information from one of the generated plant row lines to the defined evaluation area arranged between the plant row lines. In other words, from the identified plants, those are selected whose (connected) pixels extend from at least one of the identified plant rows or generated plant row lines to one of the defined evaluation areas in the image information. These plants are, by definition, evaluation area plants. Analogous to the previous description regarding the identified plants, these are pixels, in particular connected pixels with corresponding pixel values in the image information or the image, which in this case characterize or represent a "large" or elongated plant object from the real world extending transversely to the rows of plants in the image, which can include a single plant, but in particular also several overlapping plants and/or plant parts, etc.

The method further comprises a step of determining a coverage index for each of the defined evaluation areas.

In the context of this application, the term “defined” should be understood to mean specific or predetermined.

The defined evaluation area is a defined pixel area in the image information or the image, which is arranged between two rows of plants or plant row lines or extends between two rows of plants or plant row lines. The evaluation area is preferably arranged essentially centrally between two rows of plants, in particular plant row lines. The defined evaluation area preferably extends essentially parallel to two rows of plants, in particular plant row lines.

• - einer Pflanzenart bzw. Kulturpflanzenart, insbesondere der Blätterzahl, Blätterform, einem Wachstumsverhalten, einem Wachstumsstadium, einer (mittleren) Pflanzengröße oder (mittleren) Pflanzenbreite der Pflanzen bzw. Kulturpflanzen und/oder
• - den zu erwartenden Beikrautarten, insbesondere der Blätterzahl, der Blätterform, einem Wachstumsverhalten, einem Wachstumsstadium, einer Beikrautgröße oder Beikrautbreite der zu erwartenden Beikrautarten gewählt werden. Je nach Anwendungsfall kann der Auswertebereich bzw. die Auswertelinie auch nur 1 Pixel breit sein.

The evaluation area preferably extends over the entire height of the image. The evaluation area is preferably linear, ie the defined evaluation area is preferably an evaluation line. The evaluation area or the evaluation line have a defined width or extend over a defined width of the image. The width of the evaluation area can be predetermined or can be predetermined. The width of the evaluation area can be determined in particular depending on

• - a plant species or cultivated plant species, in particular the number of leaves, leaf shape, growth behaviour, growth stage, (average) plant size or (average) plant width of the plants or cultivated plants and/or
• - the expected weed species, in particular the number of leaves, the shape of the leaves, growth behavior, growth stage, weed size or weed width of the expected weed species. Depending on the application, the evaluation area or the evaluation line can also be just 1 pixel wide.

The smaller the growth size of the crops and the larger the row spacing of the crops, the wider the evaluation area can be selected in order to detect as many weeds as possible. For example, for corn with a row spacing of 75 cm and small corn plants (approx. 1 to 15 cm wide), the evaluation area can initially be 30 cm. However, if the corn plants grow larger and are 25 cm wide, the evaluation area should only be one pixel line wide. A pixel width is then between 1 mm and 5 mm wide, depending on the camera height, resolution, etc.

• - eine Anzahl von zusammenhängenden Pflanzenteilen der identifizierten Auswertebereichs-Pflanzen in dem jeweiligen definierten Auswertebereich und/oder
• - eine Gesamtbedeckung des jeweiligen definierten Auswertebereichs von Pflanzenteilen der identifizierten Auswertebereichs-Pflanzen

in der Bildinformation.A coverage index is determined for each defined evaluation area. The coverage index represents or characterizes

• - a number of contiguous plant parts of the identified evaluation area plants in the respective defined evaluation area and/or
• - a total coverage of the respective defined evaluation area of plant parts of the identified evaluation area plants

in the image information.

Accordingly, the coverage index can, on the one hand, comprise the number of (individual) connected plant parts, i.e. (individual) connected pixel segments of evaluation area plants in the respective defined evaluation area.

Alternatively or additionally, the coverage index can be the total coverage of the respective defined evaluation area of plant parts of the identified evaluation area plants, wherein the total coverage is preferably an area and/or a number of pixels. In other words, the coverage index indicates how many pixels in the evaluation area or the evaluation line are "filled" with evaluation area plant pixels.

Depending on the design of the evaluation area plant and/or the evaluation area, the connected plant parts or pixel segments and/or the total coverage can be linear or flat.

The method further comprises a step of classifying the identified evaluation area plants depending on the determined coverage index of the respective defined evaluation area by means of the computing unit. In this case, evaluation area plants of evaluation areas with a coverage index that is smaller than a defined coverage threshold are advantageously classified into a first plant class, in particular a class of cultivated plants (cultivated plant class), and/or with a coverage index that is larger than a defined coverage threshold are classified into a second plant class, in particular a class of weeds (weed class). In other words, evaluation area plants that are assigned to an evaluation area or cover it are classified depending on the coverage index of this evaluation area.

• - einem Kulturreihenabstand, einer Pflanzenart bzw. Kulturpflanzenart, insbesondere der Blätterzahl, Blätterform, einem Wachstumsverhalten, einem Wachstumsstadium, einer (mittleren) Pflanzengröße oder (mittleren) Pflanzenbreite der Pflanzen bzw. Kulturpflanzen, und/oder
• - den zu erwartenden Beikrautarten, insbesondere der Blätterzahl, der Blätterform, einem Wachstumsverhalten, einem Wachstumsstadium, einer Beikrautgröße oder Beikrautbreite der zu erwartenden Beikrautarten

gewählt werden.The coverage threshold can be specified or can be specified. The coverage threshold The value can be determined depending on

• - a crop row spacing, a plant species or crop species, in particular the number of leaves, leaf shape, growth behaviour, growth stage, (average) plant size or (average) plant width of the plants or crop plants, and/or
• - the expected weed species, in particular the number of leaves, the shape of the leaves, a growth behaviour, a growth stage, a weed size or weed width of the expected weed species

to get voted.

Preferably, before the step of classification, a pre-classification of the identified plants can be carried out independently of the coverage figures of the evaluation areas, whereby in the step of classification the pre-classified evaluation area plants retain their pre-classification or are reclassified. In other words, the plants can first be (pre)classified according to another classification method and then a validation or reclassification can be carried out depending on the coverage figures of the evaluation areas. For example, the identified plants or plant objects could be classified according to a method selected from the as yet unpublished publications EN 10 2021 213 280 or EN 10 2021 214 202 known methods in order to then be reclassified by means of the method according to the invention if a “misclassification” occurs.

Examples of a selected coverage threshold depending on a previously determined average plant width:

example 1

• Der Bedeckungsschwellenwert kann sehr stark in Richtung „sicherer Beikrauterkennung“ festgelegt werden, da die Wahrscheinlichkeit, dass Pflanzteile einer Kulturpflanze aus der Kulturreihe in den Auswertebereich hineinragen, äußerst gering ist. Demnach kann angenommen werden, dass immer, wenn ein Pflanzenteil einer Auswertebereichs-Pflanze in dem Auswertebereich angeordnet ist, es sich bei der Auswertebereichs-Pflanze um ein Beikraut handelt.

Crop row spacing 75cm, crop size 0 to 20cm:

• The coverage threshold can be set very strongly in the direction of "reliable weed detection" because the probability that plant parts of a crop from the crop row extend into the evaluation area is extremely low. It can therefore be assumed that whenever a plant part of an evaluation area plant is located in the evaluation area, the evaluation area plant is a weed.

• - für die Anzahl der zusammenhängenden Pflanzenteile: größer oder gleich 1 (somit wird schon bei einem Pflanzenteil im Auswertebereich die Auswertebereichs-Pflanze als Beikraut (um)klassifiziert)
• - für die Gesamtbedeckung: größer oder gleich 1 Pixel (somit wird schon bei einem Pixel Länge/Fläche eines zusammenhängenden Pflanzenteils im Auswertebereich die Auswertebereichs-Pflanze als Beikraut (um)klassifiziert)

Accordingly, the coverage threshold is chosen as

• - for the number of connected plant parts: greater than or equal to 1 (thus, if there is only one plant part in the evaluation area, the evaluation area plant is (re)classified as a weed)
• - for the total coverage: greater than or equal to 1 pixel (thus, even if the length/area of a contiguous plant part in the evaluation area is just one pixel, the evaluation area plant is (re)classified as a weed)

Example 2

• Der Bedeckungsschwellenwert wird als Kompromiss festgelegt, da es eine höhere Wahrscheinlichkeit gibt, dass auch Pflanzteile von Kulturpflanzen aus den Kulturreihen in den Auswertebereich hineinragen, welche nicht als Beikraut (um)klassifiziert werden sollten). Demnach wird auch geprüft, um welche Kulturpflanze es sich auf diesem Feld handelt, z.B. ob die Kulturpflanzen eher homogen wachsen (vgl. Soja) oder ob eine höhere Wahrscheinlichkeit besteht, dass die Blätter der bei Kulturpflanzen je nach Wachstum öfter in den Auswertebereich hineinragen können (vgl. Mais oder Zuckerrübe)
• Demnach wird als Bedeckungsschwellenwert gewählt
  • - für die Anzahl der zusammenhängenden Pflanzenteile: größer oder gleich 3 bis kleiner oder gleich 6
  • - für die Gesamtbedeckung: größer oder gleich 50 Pixel bei einer Auswertebereichsbreite von einem Pixel bzw. größer oder gleich 200 Pixel bei einem breiteren Auswertebereich.

Crop row spacing 75cm, crop size 20 to 50cm:

• The coverage threshold is set as a compromise, as there is a higher probability that parts of crops from the crop rows will also extend into the evaluation area, which should not be (re)classified as weeds. Accordingly, it is also checked which crop is in question in this field, e.g. whether the crops grow rather homogeneously (cf. soya) or whether there is a higher probability that the leaves of the crops can extend into the evaluation area more often depending on their growth (cf. corn or sugar beet).
• Accordingly, the coverage threshold is chosen as
  • - for the number of connected plant parts: greater than or equal to 3 to less than or equal to 6
  • - for the total coverage: greater than or equal to 50 pixels for an evaluation area width of one pixel or greater than or equal to 200 pixels for a wider evaluation area.

Example 3

• Bei kleiner werdendem Reihenabstand und größer werdenden Kulturpflanzen müssen die Bedeckungsschwellenwert für die Anzahl der zusammenhängenden Pflanzenteile und der Gesamtbedeckung weiter erhöht werden, um die Wahrscheinlichkeit einer richtigen (Um-)Klassifizierung von Kulturpflanzen zu erhöhen.

Crop row spacing 35cm, crop size 20 to 50cm:

• As row spacing decreases and crop sizes increase, the coverage thresholds for the number of connected plant parts and the total coverage need to be further increased to increase the probability of correct (re)classification of crops.

Furthermore, it is advantageous if the method comprises a step of controlling an agricultural working machine, in particular an agricultural working tool of an agricultural working machine, in particular a spraying device of a field sprayer, depending on ity of the classified evaluation area plants. This enables the method to specifically combat or regulate weeds in and/or between the rows of plants.

The computing unit is or the computing units are designed or set up for image processing so that it can carry out calculation steps or image processing steps to carry out the method according to the invention. Accordingly, each computing unit has corresponding image processing software. The computing unit can be, for example, a signal processor, a microcontroller or the like, wherein the storage unit can be a flash memory, an EPROM or a magnetic storage unit. The communication interface can be designed to read in or output data wirelessly and/or via a cable, wherein a communication interface that can read in or output cable-based data can read this data, for example, electrically or optically from a corresponding data transmission line or output it to a corresponding data transmission line. Accordingly, the method according to the invention can be implemented, for example, in software or hardware or in a hybrid form of software and hardware in the computing unit or a control unit.

The computing unit can be completely or partially located on the agricultural machine or integrated into it. However, the computing unit can also be completely or partially integrated externally, e.g. in a cloud. The computing unit can therefore also be divided into different units, e.g. mobile and stationary units.

Also advantageous is a computer program product or computer program with program code that can be stored on a machine-readable carrier or storage medium such as a semiconductor memory, a hard disk memory or an optical memory and is used to carry out, implement and/or control the steps of the method according to one of the embodiments described above, in particular when the program product or program is executed on a computer or device.

• 1 eine schematische Darstellung einer Bildinformation bzw. eines Bildes eines erfassten Feldabschnitts mit einem Auswertebereich; und
• 2 ein Ablaufdiagramm eines Verfahrens gemäß einem Ausführungsbeispiel.

Embodiments of the invention are shown in the drawings and explained in more detail in the following description. It shows:

• 1 a schematic representation of image information or an image of a captured field section with an evaluation area; and
• 2 a flow chart of a method according to an embodiment.

1 shows image information 10 or an image 10 of a field section of an agricultural area with plants 12 captured by means of an optical detection unit or camera (not shown).

As from 1 As can be seen, in image 10 the plants 12 and plant rows 14 were first identified using the identified plants 12 by means of a computing unit (not shown). The plant rows 14 were identified by generating defined straight plant row lines 16 in image trajectories with the highest NDVI value of the plants 12. Then, from the group of identified plants 12, evaluation area plants 18 were identified, which in image 10 extend from one of the generated plant row lines 16 to a defined evaluation area 20 arranged between the generated plant row lines 16. In order to be able to classify these evaluation area plants 18, a coverage index for the defined evaluation area 20 was determined.

The defined evaluation area 20 is arranged essentially parallel and centrally between the two plant rows 14 or plant row lines 16. The coverage index represents or characterizes a number of connected plant parts 22 of the identified evaluation area plants 18 in the respective defined evaluation area 20.

According to the invention, the identified evaluation area plants 18 are classified depending on the determined coverage index of an evaluation area 20, i.e. the number of connected plant parts 22 or connected pixel segments 22 of the respective evaluation area plants 18 in the evaluation area 20.

In the example shown, the coverage index for the evaluation area 20 is “7”, since contiguous plant parts 22 were determined or detected in the evaluation area 20 “7”.

By definition, all plants 12 in the plant rows 14 were initially pre-classified as cultivated plants 24. However, since a defined coverage threshold of 5 was specified, from which the respective evaluation area plants 18 are to be regarded as weeds 26, the three evaluation area plants 18 identified in Figure 10 were (re)classified as weeds 26.

2 shows a flow chart of an embodiment of the approach presented here as a method 100 for classifying plants 12 in and/or between rows of plants 14 of an agricultural area. The method 100 comprises a step of receiving 102 image information 10 from a field section of an agricultural area with plants 12, which is detected by means of an optical detection unit. The method 100 further comprises a step of identifying 104 plants 12 in the image information 10 by means of a computing unit. The method 100 also comprises a step of identifying 106 rows of plants 14 in the image information 10 using the identified plants 12 by means of the computing unit. The method 100 further comprises a step of identifying 108 evaluation area plants 18 from the group of identified plants 12 by means of the computing unit, wherein the evaluation area plants 18 extend in the image information 10 from one of the identified rows of plants 14 into a defined evaluation area 20 arranged between the rows of plants 14. The method 100 further comprises a step of determining 110 a coverage index for each of the defined evaluation areas 20, wherein the coverage index represents a number of contiguous plant parts 22 of the identified evaluation area plants 18 in the respective defined evaluation area 20 and/or a total coverage of the respective defined evaluation area 20 of plant parts 22 of the identified evaluation area plants 18 in the image information 10, by means of the computing unit. The method 100 finally comprises a step of classifying 112 the identified evaluation area

Plants 18 depending on the determined coverage index of the respective defined evaluation area 20 by means of the computing unit.

QUOTES INCLUDED IN THE DESCRIPTION

This list of documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions.

Cited patent literature

• DE 102017210804 A1 [0002]
• EN 102021213280 [0035]
• EN 102021214202 [0035]

Claims (14)

Method (100) for classifying plants (12) in and/or between rows of plants (14) of an agricultural area, comprising the steps: - receiving (102) image information (10) from a field section of an agricultural area with plants (12) recorded by an optical detection unit; - identifying (104) plants (12) in the image information (10) by means of a computing unit; - identifying (106) rows of plants (14) in the image information (10) using the identified plants (12) by means of the computing unit; - identifying (108) evaluation area plants (18) from the group of identified plants (12) by means of the computing unit, wherein the evaluation area plants (18) extend in the image information (10) from one of the identified rows of plants (14) into a defined evaluation area (20) arranged between the rows of plants (14); - Determining (110) a coverage index for each of the defined evaluation areas (20), wherein the coverage index o represents a number of connected plant parts (22) of the identified evaluation area plants (18) in the respective defined evaluation area (20) and/or o represents a total coverage of the respective defined evaluation area (20) of plant parts (22) of the identified evaluation area plants (18) in the image information (10), by means of the computing unit; and - Classifying (112) the identified evaluation area plants (18) depending on the determined coverage index of the respective defined evaluation area (20) by means of the computing unit. Procedure (100) according to Claim 1 , characterized in that in the step of identifying (106) the plant rows (14) a plant row line (16) is generated in each of the plant rows (14), wherein the evaluation area plants (18) extend in the image information (10) from one of the generated plant row lines (16) into the defined evaluation area (20) arranged between the plant row lines (16). Procedure (100) according to Claim 1 or 2 , characterized in that the defined evaluation area (20) is arranged substantially centrally between two rows of plants (14), in particular plant row lines (16). Method (100) according to one of the preceding claims, characterized in that the defined evaluation region (20) is linear. Method (100) according to one of the preceding claims, characterized in that the defined evaluation area (20) extends substantially parallel to two plant rows (14), in particular plant row lines (16). Method (100) according to one of the preceding claims, characterized in that the total coverage is an area and/or a number of pixels of plant parts (22) of the respective evaluation area plants (18) in the defined evaluation area (20) of the image information (10). Method (100) according to one of the preceding claims, characterized in that in the step of classifying (112) evaluation area plants (18) of evaluation areas (20) with a coverage index which is smaller than a defined coverage threshold value are classified into a first plant class, in particular a class of cultivated plants (24), and/or with a coverage index which is larger than a defined coverage threshold value are classified into a second plant class, in particular a class of weeds (26). Method (100) according to one of the preceding claims, characterized in that before the step of classification (112) a pre-classification of the identified plants (12) is carried out independently of the coverage index of the evaluation areas (20), wherein in the step of classification (112) the pre-classified evaluation area plants (18) retain their pre-classification or are reclassified. Method (100) according to one of the preceding claims, characterized by a step of controlling an agricultural working machine, in particular an agricultural working tool of an agricultural working machine, in particular a spraying device of a field sprayer, depending on the classified evaluation area plants (18). Computing unit which is configured to execute and/or control the steps of a method (100) according to one of the preceding claims. Plant identification unit with an optical detection unit for detecting a field section of an agricultural area with plants (12) in order to obtain image information (10) from the detected field section, and a computing unit according to Claim 10 . Agricultural working machine, in particular agricultural field sprayer, with a agricultural working tool, in particular a spraying device, and a plant identification unit according to Claim 11 , wherein the working tool, in particular the spraying device, is controllable depending on the classified evaluation area plants (18). Computer program which is designed to carry out the steps of a method (100) according to one of the Claims 1 until 9 to execute and/or control when the computer program is executed on a computer. Machine-readable storage medium on which the computer program is Claim 13 is stored.

Images