Weeding device, automatic weeding method, and automatic weeding program
The weeding device enhances efficiency by using image classification and automated weeding to distinguish and remove weeds from cultivated plants, reducing manual effort and crop damage.
Patent Information
- Application Number
- JP2020173229
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2020-10-14
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-10-14
AI Technical Summary
Conventional self-propelled weeding devices face challenges such as physical burden on workers, low efficiency, and the need for additional manual work to collect weeds, especially near flowers and crops.
A weeding device equipped with image data acquisition, a trained model for plant classification, and a weeding mechanism that automatically identifies and removes weeds, storing them in a containment section based on classification results.
Improves weeding efficiency by accurately distinguishing cultivated plants from weeds and automating the weeding process, reducing the need for manual collection and minimizing damage to crops.
Smart Images

Figure 0007585704000001 
Figure 0007585704000002 
Figure 0007585704000003
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a weeding device, an automatic weeding method, and an automatic weeding program. [Background technology]
[0002] Conventionally, self-propelled weeding devices have been used. Although such self-propelled weeding devices can perform uniform maintenance within a maintenance area, it is difficult to remove weeds near flowers or crops, so that the weeds around the flowers or crops have to be removed manually in order not to damage the flowers or crops.
[0003] Recently, a robot (device) for automatic weed removal has been proposed that takes a picture of a plant cultivation area such as farmland with a camera, captures an image of the plant, and selects a tool for applying a herbicide by determining whether the cultivated plant is a weed from the image (for example, Patent Document 1). Such a device performs a process of removing the weeds, such as spraying a herbicide, using a tool selected based on the distance between the cultivated plant and the weed. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] International Publication No. 2017 / 002093 Summary of the Invention [Problem to be solved by the invention]
[0005] However, the use of conventional self-propelled weeding devices has problems such as physical strain on the operator and time-consuming operation. In addition, devices that select tools from images of plants taken by a camera to remove weeds have problems such as low weeding efficiency because weeds remain in the cultivation area after weeding, requiring additional work to collect the weeds.
[0006] Therefore, an object of the present disclosure is to provide a weeding device, an automatic weeding method, and an automatic weeding program that have high weeding efficiency. [Means for solving the problem]
[0007] In order to solve the above-mentioned problems, a weeding device according to one embodiment of the present disclosure is characterized in that it comprises a body, wheels for propelling the body, a data acquisition unit for acquiring image data of a working area, a classification unit having a trained model capable of classifying, based on the image data, plants shown in the image shown by the image data into either cultivated plants cultivated in the working area or plants other than cultivated plants, and classifying the plants shown in the image into either cultivated plants or plants other than cultivated plants based on the image data acquired by the data acquisition unit, a storage unit provided on the body of the vehicle capable of storing plants, a weeding unit for performing weeding work by pulling out or cutting plants determined to be plants other than cultivated plants based on the classification result of the classification unit as weeds from the ground, and a collection unit for collecting weeds and storing them in the storage unit.
[0008] In addition, an automatic weeding method according to one aspect of the present disclosure is characterized in that it acquires image data of a work area, inputs the image data into a trained model that can classify plants shown in the image shown by the image data based on the image data into either cultivated plants grown in the work area or plants other than cultivated plants, obtains a classification result as an output result from the trained model that classifies the plants shown in the image into either cultivated plants or plants other than cultivated plants, performs weeding work by pulling out or cutting plants determined to be plants other than cultivated plants based on the classification result from the ground as weeds, and stores the weeds in a storage section capable of storing plants.
[0009] Furthermore, an automatic weeding program according to one embodiment of the present disclosure is characterized in having a computer execute the following steps: acquire image data of a work area; input the image data into a trained model that can classify, based on the image data, plants appearing in the image shown by the image data into either cultivated plants grown in the work area or plants other than cultivated plants; acquire, as an output result from the trained model, a classification result that classifies the plants appearing in the image into either cultivated plants or plants other than cultivated plants; perform weeding work to pull out or cut plants determined to be plants other than cultivated plants based on the classification result as weeds from the ground; and store the weeds in a storage unit capable of storing plants. Effect of the Invention
[0010] According to the present disclosure, it is possible to provide a weeding device, an automatic weeding method, and an automatic weeding program with high weed-killing efficiency. [Brief description of the drawings]
[0011] [Figure 1] FIG. 2 is a schematic diagram showing a configuration example of a weeding device according to the first embodiment. [Diagram 2] FIG. 2 is a schematic diagram showing a configuration example of a weeding arm of the weeding device according to the first embodiment. [Diagram 3] FIG. 2 is a schematic functional block diagram showing an example of a functional configuration of the weeding device according to the first embodiment. [Figure 4] FIG. 2 is a schematic diagram showing an example of a hardware configuration of the weeding device according to the first embodiment. [Figure 5A] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 5B] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 5C] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 5D] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 5E] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 5F] FIG. 1 is a schematic diagram illustrating an automatic weeding method according to a first embodiment. [Figure 6] 4 is a flowchart illustrating an automatic weeding method according to the first embodiment. [Figure 7A] FIG. 11 is a schematic diagram showing a configuration example of a weeding device according to a second embodiment. [Figure 7B] FIG. 11 is a schematic diagram showing a configuration example of a weeding device according to a second embodiment. [Figure 8] FIG. 11 is a schematic functional block diagram showing an example of a functional configuration of a weeding device according to a second embodiment. [Figure 9] 10 is a flowchart illustrating an automatic weeding method according to a second embodiment. [Figure 10] 10 is a flowchart illustrating an automatic weeding method according to a second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0012] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings. However, the embodiment described below is merely an example, and is not intended to exclude various modifications and application of technologies not specified below. The present disclosure can be implemented in various modifications (for example, combinations of the embodiments) without departing from the spirit of the disclosure. In addition, in the following description of the drawings, the same or similar parts are denoted by the same or similar reference numerals.
[0013] The weeding device described below is realized by the cooperation of hardware, which is an information processing device such as a general-purpose computer, and software, which is a computer program. The computer program can be read by one or more processors provided in the weeding device, and the processor functions as each part of the weeding device by installing the computer program on a hard disk.
[0014] 1. First embodiment Hereinafter, a weeding device 1 according to a first embodiment and an automatic weeding method and an automatic weeding program executed by the weeding device 1 will be described with reference to Figs. 1 to 6. The weeding device 1 of this embodiment is a device that uses a trained model to classify plants in a work area, and automatically (i.e., without the operator operating the device or performing other operations) weeding work by classifying plants other than cultivated plants as weeds.
[0015] (1.1) Overall configuration of the weeding device As shown in FIG. 1, the weeding device 1 includes a body 11, wheels 12 for propelling the body 11, an imaging device 13, a storage section 14 for storing weeds, and a weeding arm 15. Each part of the weeding device 1 will be described below.
[0016] The machine body 11 includes a main body frame of the weeding device 1, and a wheel drive mechanism and a weeding arm drive mechanism (not shown) that are arranged inside the main body frame. The wheel drive mechanism has, for example, a drive unit such as an engine, and a drive train including a transmission and a drive shaft. The weeding arm drive mechanism may have any configuration as long as it is capable of performing weeding work.
[0017] The wheels 12 are attached to the vehicle body 11 to allow it to travel. The wheels 12 include front wheels 12A that are attached to the front lower part of the vehicle body 11 and support the front part of the vehicle body 11, and rear wheels 12B that support the rear part of the vehicle body 11. The wheels 12 are attached to both ends of the drive shaft of the vehicle body 11, similar to wheels for a general vehicle.
[0018] The imaging device 13 is a light receiving device having a function of detecting a two-dimensional image on an imaging surface (on a CCD array surface) by imaging, and is, for example, a camera having an imaging element such as a CCD. The imaging device 13 is attached, for example, to the front of the machine body 11, and captures an image of the work area in front of the weeding device 1 (in the direction of travel) to obtain image data of the work area. As the weeding device 1 moves within the work area on the wheels 12, the imaging device 13 can capture images of cultivated plants or plants other than cultivated plants (weeds) in a desired area of the work area to obtain image data.
[0019] The storage section 14 is provided in a part of the machine body 11, and stores the weeds pulled out by the weeding arm 15. The storage section 14 may be formed in any shape and from any material as long as it is capable of storing weeds.
[0020] The weeding arm 15 has a function of performing a weeding operation of pulling out or cutting weeds, and a storage operation of storing weeds in the storage section 14. That is, the weeding arm 15 described in this embodiment has the functions of both a weeding section and a collecting section. Fig. 2 is a schematic diagram showing a configuration example of the weeding arm 15. As shown in Fig. 2, the weeding arm 15 has, for example, a gripping arm 151 having an openable and closable first gripping arm 151A and second gripping arm 151B, and a gripping arm rotating part 152 that changes the movement direction of the first gripping arm 151A and the second gripping arm 151B. The weeding arm 15 may have three or more gripping arms.
[0021] The weeding arm 15 is controlled so that during weeding work, the weeding arm 15 performs a gripping operation in which the first gripping arm 151A and the second gripping arm 151B are closed to grip a weed, and a rotation operation in which the gripping arm rotation part 152 is rotated while the first gripping arm 151A and the second gripping arm 151B grip the weed, thereby pulling out the weed from the ground. The weeding arm 15 is controlled by the control part 36, which will be described later. The weeding arm 15 is preferable because it grips the weed with its roots during the gripping operation, making it easier to pull out the weed during the rotation operation.
[0022] (1.2) Functional configuration of the weeding device Fig. 3 is a schematic functional block diagram showing an example of the functional configuration of the weeding device 1. Each functional block shown in Fig. 3 is realized by a processor 21 of an arithmetic device 20, which will be described later in Fig. 4, executing a necessary program and controlling the operation of each part of the weeding device 1.
[0023] The weeding device 1 according to this embodiment has a data acquisition unit 31, a memory unit 32, a machine learning unit 33, a weeding area identification unit 34, a position information detection unit 35, a control unit 36, an imaging device 13, and a weeding arm 15. The weeding device 1 is also connected to an input device 16 and a display device 17, which are external devices. The input device 16 and the display device 17 may be provided as part of the weeding device 1. Each functional block will be described below.
[0024] [Data Acquisition Section] The data acquisition unit 31 acquires image data of the working area via the imaging device 13, an external storage device (not shown), or a network. At this time, the data acquisition unit 31 may perform pre-processing such as converting the acquired image data into a data format that can be handled inside the weeding device 1, converting the image data into a predetermined size, trimming the image to extract only a necessary partial image, and rotating the image. The data acquisition unit 31 may also acquire location information indicating the location where the image data was captured. When weeding work is performed after checking the state of weed growth in the entire work area, the state of weed growth in each part of the work area can be checked based on the image data and the location information corresponding to the image data. The image data acquired by the data acquisition unit 31 is stored in the storage unit 32 .
[0025] [Storage] The storage unit 32 is secured as a storage area in the auxiliary storage device 24 shown in FIG. 4, for example. The storage unit 32 stores image data of the image of the work area acquired by the data acquisition unit 31. The storage unit 32 also stores position information indicating the position of the weeding device 1 detected by the position information detection unit 35. Here, the storage unit 32 stores the position information of the weeding device 1 at the time of capturing the image of the work area in association with the image data of the image. Furthermore, the storage unit 32 stores information (position information) indicating a section in which weeds grow in clusters that has been identified by the weeding area identification unit 34 as being a weeding area.
[0026] The storage unit 32 also functions as a storage unit for the machine learning unit 33. In this case, the storage unit 32 stores learning data formed by the labeling unit 331 assigning labels (annotating) to image data acquired by the data acquisition unit 31 and stored in the storage unit 32. The storage unit 32 also stores the results output from the machine learning unit 33. The results output from the machine learning unit 33 are classification results that classify plants shown in an image into either cultivated plants cultivated in a work area or plants other than cultivated plants, based on image data of a work area such as farmland. The storage unit 32 stores the image data of the work area and the classification results in association with each other.
[0027] [Machine Learning Department] The machine learning unit 33 includes, for example, a memory unit 32, a label assignment unit 331, a learning data extraction unit 332, a learning unit 333, a learning model memory unit 334, and a classification unit 335. The machine learning unit 33 has a function of classifying, based on the image data of the image of the work area acquired by the data acquisition unit 31, cultivated plants cultivated in the work area and plants other than cultivated plants, outputting the classification result, and storing it in the memory unit 32. Each component of the machine learning unit 33 will be described below.
[0028] (Label assignment section) The label assignment unit 331 has a function of assigning labels, i.e., annotating, to image data acquired by the data acquisition unit 31 and stored in the storage unit 32, and storing the data in the storage unit 32. More specifically, the label assignment unit 331 has a function of forming learning data by associating a label with the image data to which the label has been assigned, and storing the learning data in the storage unit 32.
[0029] The label assignment unit 331 may assign, for example, two labels, "cultivated plant name (e.g., rice)" and "others." As a result, a plant labeled with "others" is a plant other than a cultivated plant, and becomes a weed to be removed. In this case, for example, if rush grass grows in farmland where rice is cultivated, the rush grass is also classified as a "weed." Furthermore, the labels are not limited to the above two. For example, the types of cultivated plants and other plants shown in the image may be added as labels. In this case, the classification unit 335 classifies all plants that are added with a label other than the "cultivated plant name (e.g., rice)" as "weeds."
[0030] The label assignment in the label assignment unit 331 is performed, for example, by displaying the image data stored in the storage unit 32 on the display device 17 and having an operator who sees the displayed image assign a label to each piece of learning data by operating the input device 16. When a label is assigned to the image data, the label assignment unit 331 associates the label with the learning data and stores it in the storage unit 32.
[0031] For example, the labeling unit 331 may automatically label the image data as "cultivated plant name (e.g., rice)" if the discrimination value calculated based on a mathematical algorithm that automatically discriminates the image data is equal to or greater than a predetermined threshold value, and label the image data as "other" if the discrimination value is less than the predetermined threshold value. In this case, the image data and the labeling result may be displayed on the display device 17 to allow the worker to confirm the result and to correct any errors in the labeling.
[0032] (Learning data extraction part) The learning data extraction unit 332 has a function of extracting a plurality of pieces of learning data stored in the storage unit 32. At this time, the learning data extraction unit 332 may extract a portion of the learning data stored in the storage unit 32 as learning data (training data), and the remaining learning data may be used as verification data (verification data) and final evaluation data (test data). The training data may be further divided into a plurality of groups to create a plurality of data sets, and learning and evaluation may be performed for each data set. In this case, it is preferable that the learning data extraction unit 332 creates each data set of the training data so that the learning results are unlikely to be biased. For example, the learning data extraction unit 332 may randomly extract learning data as a data set from the learning data stored in the storage unit 32, or may use a statistical sampling method such as a systematic sampling method.
[0033] (Learning Department) The learning unit 333 has a function of generating (learning) a trained model that estimates and outputs a classification result (label) when image data is input. That is, in the present disclosure, the learning unit 333 estimates whether a plant depicted in an image shown by the input image data is a cultivated plant cultivated in the working area or a plant other than a cultivated plant, and outputs the estimated classification result (label). Here, the "input image data" refers to image data of the working area acquired by the data acquisition unit 31 and stored in the storage unit 32, and image data that is not learning data (image data to which no label is assigned).
[0034] The learning unit 333 performs supervised learning on the learning model using the learning data extracted by the learning data extraction unit 332, and generates a trained model. The learning unit 333 can use, for example, a neural network as the learning model. The learning unit 333 may be configured to perform supervised learning using the image data (partial image data) of the plant included in each of the learning data included in the data set extracted by the learning data extraction unit 332 or the image data indicated by the learning data as input data, and the label assigned to the learning data as output data. The learning unit 333 may also perform semi-supervised learning using both labeled image data and unlabeled image data as learning data, and generate a trained model.
[0035] In this case, a neural network with three layers, an input layer, an intermediate layer, and an output layer, may be used as the learning model, but a so-called deep learning method using a neural network with more layers may also be used. By using the deep learning method, more effective learning and inference can be performed. Furthermore, the learning unit 333 may be configured to perform machine learning using other learning models such as a support vector machine, a random forest, a logistic regression, a decision tree, or a k-nearest neighbor method. The trained model generated by the learning unit 333 is stored in a trained model storage unit 334. The stored trained model is used by a classification unit 335 to infer a classification result (label) based on image data.
[0036] (Learning model memory section) The learning model storage unit 334 stores the learned model generated by the learning unit 333. The learning model storage unit 334 may store a plurality of learned models corresponding to a plurality of work areas, respectively. Furthermore, the learning model storage unit 334 may store a learning model (that is, an initial model) before supervised learning using learning data is performed.
[0037] (Classification Department) The classification unit 335 has a function of performing classification using a trained model generated by the learning unit 333, based on image data of the working area captured by the imaging device 13. The classification unit 335 inputs the image data of the working area to the trained model, and obtains a classification result (an estimated result of a label) as an output of the trained model. The classification unit 335 outputs the classification result to the storage unit 32.
[0038] [Location information detection section] The position information detection unit 35 is provided in the machine body 11 and has a function of detecting the position of the machine body 11. The position information detection unit 35 has at least one of a gyro sensor 351, a GPS (Global Positioning System) 352, and a position information calculation unit 353 that calculates the position information of the weeding device 1 by mobile communication base station positioning, for example. The position information detection unit 35 outputs the detected position information to the storage unit 32 for storage.
[0039] [Weeding area specific section] The weeding area identifying unit 34 has a function of identifying an area where weeds are present based on image data of the work area. The weeding area identifying unit 34 has a weed presence determination unit 341 that determines the presence or absence of weeds in the imaging area based on the image data, and an occupancy rate calculation unit 342 that calculates a weed occupancy rate that indicates the extent to which weeds are growing. The weeding area identifying unit 34 may identify an area where weeds detected by the weed presence determination unit 341 are present as a weeding area. The weeding area identifying unit 34 may also identify a weeding area based on the weed occupancy rate calculated by the occupancy rate calculation unit 342.
[0040] The weeding area specifying unit 34 can specify the weeding area by, for example, a primary scan in which the weed presence determining unit 341 determines the presence or absence of weeds in the imaging area. The weeding area specifying unit 34 may specify the weeding area by performing a secondary scan in which the occupancy rate calculation unit 342 calculates the weed occupancy rate after the primary scan. It is preferable to specify the weeding area by the primary scan and the secondary scan, because this makes it possible to determine an area with a high priority for weeding work. The primary and secondary scans will be described below.
[0041] [Primary Scan] The weed presence determination unit 341 identifies a section of the work area where weeds grow in clusters based on the image data acquired by the data acquisition unit 31 and the classification result (label estimation result) by the classification unit 335, and sets the section where weeds grow in clusters as a weeding area. The weeding area identification unit 34 identifies the position of the section where weeds grow in clusters based on the position information associated with the image data, and outputs the position information of the section where weeds grow in clusters to the memory unit 32 as position information of the weeding area.
[0042] Here, the weed presence determination unit 341 may determine the range of the work area based on the image data of the work area, divide the work area into multiple sections, and identify the sections in which weeds grow as weeding areas. Specifically, the weed presence determination unit 341 determines the range of the work area by detecting the position of the periphery of the farmland in the image data, for example, by edge detection, and divides the area surrounded by the periphery of the farmland into multiple sections.
[0043] It is preferable that the weed presence determination unit 341 divides the area surrounded by the perimeter of the farmland when the farmland is viewed in plan, for example, into multiple rectangular sections of equal area. By making each section the same area, when identifying an area where weeds grow, it is possible to determine whether each section is an area where weeds grow under the same conditions. In addition, by making each section rectangular, it becomes easier to divide each section into sections of equal area, and it becomes easier to divide the entire work area into one of the sections. In this case, the plots in the edge areas of the farmland do not have to be of the same area or rectangular shape as the other plots.
[0044] In this way, in order to divide the farmland into a plurality of rectangular sections having the same area in a plan view, it is preferable to correct the inclination of the imaged object in the image, i.e., correct the shape of the farmland shown in the image, based on the photographing conditions such as the position and attitude of the weeding device 1 when the working area is photographed, and then divide the farmland into small sections. Examples of the photographing conditions include the photographing height of the image data (for example, the mounting height of the photographing device 13 from the ground) and the mounting angle to the machine body 11, the inclination of the weeding device 1 when photographing (the inclination of the ground), and the ground height of the photographing position of the weeding device 1. The photographing conditions that change depending on the photographing position can be obtained based on the position information detected by the position information detection unit 35 provided in the weeding device 1 at the time of photographing and stored in the storage unit 12, for example. The weed presence determination unit 341 outputs the position information of the weeding area from the image data of the working area and the weeding device 1, and stores it in the storage unit 32.
[0045] [Secondary Scan] The occupancy calculation unit 342 further divides the section in which weeds grow, identified by the weed presence determination unit 341, into a plurality of small sections and determines the presence or absence of weeds for each small section. The occupancy calculation unit 342 calculates the occupancy of weeds in the section in which weeds grow from the ratio of the number of small sections in which weeds exist to the total number of small sections. That is, the occupancy calculation unit 342 calculates the occupancy of weeds in the section in which weeds grow from the following formula (1). Weed occupancy rate in plots where weeds grow [%] = (number of plots with weeds / total number of plots) × 100 (1)
[0046] Furthermore, the occupancy calculation unit 342 may calculate the weed occupancy from the ratio of the area of the region where weeds grow in colonies (the area of only the weed region) in the section where weeds grow in colonies to the area of the section where weeds grow in colonies (the area of the entire section) identified by the weed presence determination unit 341. That is, the occupancy calculation unit 342 may calculate the occupancy of weeds in the section where weeds grow in colonies from the following formula (2). Weed occupancy rate in a plot where weeds grow [%] = (area of area where weeds exist / area of plot where weeds exist) × 100 (2)
[0047] The occupancy calculation unit 342 compares the calculated weed occupancy with a preset threshold, and if the weed occupancy is equal to or greater than the threshold, determines that the section in which weeds grow is a weeding area. The occupancy calculation unit 342 outputs position information of the weeding area and stores it in the memory unit 32.
[0048] The occupancy calculation unit 342 may assign priorities to multiple weeding areas. For example, when there are multiple sections with weed occupancy rates equal to or higher than a threshold, the occupancy calculation unit 342 may determine that the section with the higher weed occupancy rate has a higher priority for weeding work. The occupancy calculation unit 342 may output the position information of the weeding area and the priority in association with each other, and store them in the storage unit 32. At this time, the weed presence determination unit 341 does not need to output the position information of the weeding area.
[0049] [Control Unit] The control unit 36 controls the weeding device 1 to drive the wheels 12 to collect image data for learning data of the work area, perform a primary scan, or perform weeding work. The control unit 36 needs to make the weeding device 1 run evenly over the entire work area to collect image data or perform a primary scan. In addition, for weeding work, the control unit 36 acquires position information of the weeding area stored in the memory unit 32, and makes the weeding device 1 run to the weeding area according to the position information. If a priority is associated with the weeding area, the control unit 36 moves through the weeding areas with higher priorities in order and performs weeding work in each weeding area.
[0050] (1.3) Hardware configuration of the weeding device Fig. 4 is a schematic diagram showing an example of the hardware configuration of a functional unit for identifying weeds and identifying areas where weeds grow, among the functional units of the weeding device 1. That is, in Fig. 4, the hardware configuration of the functional unit for driving the wheels 12 to move the machine body 11 and for driving the weeding arm 15 that performs the weeding process in the weeding device 1 is omitted.
[0051] As shown in FIG. 4, the weeding device 1 includes a calculation device 20. The calculation device 20 is an information processing device including a processor 21, a ROM (Read Only Memory) 22, a RAM (Random access memory) 23, and an auxiliary storage device (memory) 24. The calculation device 20 is connected to the imaging device 13, the input device 16, and the display device 17, for example, via an external interface (I / F) 25. The calculation device 20 is connected to the computer 18, for example, via a communication interface (I / F) 26. Here, the imaging device 13 is the imaging device 13 attached to the weeding device 1 shown in FIG. 1. The imaging device 13, the input device 16, the display device 17, and the computer 18 include a user interface (not shown) for connecting to the calculation device 20. Here, the hardware resources of each part of the calculation device 20 cooperate with a predetermined program to construct each function of the data acquisition unit 31, the storage unit 32, the machine learning unit 33, the weeding area identification unit 34, the position information detection unit 35, and the control unit 36 of the weeding device 1.
[0052] The processor 21 is a processor that performs overall control of the weeding device 1. The processor 21 reads out a program stored in the ROM 22 via the bus 27, and controls the weeding device 1 in accordance with the program. The processor 21 controls the weeding device 1 so that the weeding device 1 travels by itself within a predetermined working area (farm) without relying on steering operation or the like by a driver. The ROM 22 stores a program for executing an automatic weeding method. The RAM 23 temporarily stores calculation data, various data input by an operator via the input device 16, and the like.
[0053] The auxiliary storage device 24 is backed up by, for example, a battery (not shown), and the stored state is maintained even when the power supply of the weeding device 1 is turned off. The auxiliary storage device 24 stores image data input from the imaging device 13 via the external interface 25, learning data with labels, classification results output from the machine learning unit 33, and weeding areas output from the weeding area identification unit 34 (the learning data, classification results, and weeding areas will be described later). The auxiliary storage device 24 also stores input data input by the operator from the input device 16, as well as various data input via the external interface 25 and the communication interface 26.
[0054] The imaging device 13 inputs image data to the arithmetic device 20 via the external interface 25. The photographing by the imaging device 13 may be controlled by the arithmetic device 20 of the weeding device 1 executing a program, or may be controlled by another device.
[0055] The input device 16 is a keyboard, a pointing device, a voice input device such as a microphone, or the like, and inputs commands and input data based on the operation of an operator to the arithmetic unit 20 via an external interface 25 . The display device 17 is, for example, a liquid crystal display (LCD) or an organic electro-luminescence display (OLED), and receives and displays various data from the arithmetic device 20 via the external interface 25. The display device 17 displays, for example, each piece of data loaded onto the memory of the arithmetic device 20, data obtained as a result of executing a program or the like, image data of a working area obtained by capturing an image of the imaging device 13, data output from a machine learning unit 33 (details of which will be described later), and the like.
[0056] The computer 18 is connected to the arithmetic device 20 via a network 28 through a communication interface 26. The computer 18 is equipped with a display device and an input device (not shown), and can input commands and input data based on an operator's operation via the input device to the arithmetic device 20 via the network 28, and can receive various data from the arithmetic device 20 and display them on the display device. The network 28 may be either a wired network or a wireless network. When the network 28 is a wireless network such as a wireless LAN, Bluetooth (registered trademark), or a mobile communication network, the weeding device 1 can receive commands in real time from an operator in a remote location, such as outside the work area, and can display the execution state of a program or the like and acquired data to the operator in the remote location.
[0057] (1.4) Automatic weeding method The automatic weeding method according to this embodiment will be described below with reference to Fig. 5 and Fig. 6. Fig. 5 is a schematic diagram for explaining the automatic weeding method according to this embodiment, and Fig. 6 is a flow chart for explaining the automatic weeding method according to this embodiment. Note that the weeding device 1 automatically travels according to a program stored in advance, and performs weeding work based on image data of the working area.
[0058] The automatic weeding method according to this embodiment is carried out at least by the following method. (a) Acquire image data of the work area (b) Inputting the image data into a trained model capable of classifying plants shown in the image shown by the image data into either a "cultivated plant" cultivated in a working area or a "non-cultivated plant" based on the image data, and obtaining a classification result as an output result from the trained model that classifies the plants shown in the image into either a "cultivated plant" or a "non-cultivated plant". (c) Based on the classification results, plants that are determined to be non-cultivated plants are weeded by pulling them out or cutting them from the ground. (d) Containing the weeds in a container
[0059] The automatic weeding method will now be described in more detail. First, as shown in Fig. 5A, the weeding device 1 travels over the entire farm, which is the working area, and acquires image data of a plurality of images 21 (step S11). The imaging device 13 captures images of the working area, for example, every several meters, acquires image data of the entire working area, and inputs the image data to the data acquisition unit 31 (see Fig. 3). The imaging device 13 can adjust the imaging interval according to conditions such as the angle of view and resolution of the imaging device 13.
[0060] The weeding device 1 generates a trained model capable of classifying plants shown in the image 21 represented by the image data into either cultivated plants cultivated in the working area or plants other than cultivated plants, based on the image data (step S12). The trained model is generated in the machine learning unit 33 by generating training data by assigning labels, and performing supervised learning in which the learning model outputs the labels assigned to the training data in response to input of the training data to which the labels have been assigned. The weeding device 1 stores the trained model in the training model storage unit 334 of the machine learning unit 33.
[0061] As shown in FIG. 5B, the weeding device 1 performs a primary scan of the work area via the imaging device 13 (steps S13 to S15). The weeding device 1 inputs image data acquired by the data acquisition unit 31 to the classification unit 335 and acquires a classification result as to whether the plant growing in the work area captured immediately before is a "cultivated plant" or a "plant other than cultivated plant (i.e., weed)" (step S13). The weeding device 1 also defines the range of the work area based on image data acquired by traveling sequentially over the entire work area, and divides the work area into a plurality of sections (step S14). As a result, the weeding device 1 identifies the section where weeds grow in clusters (step S15). Here, Fig. 5B shows an example in which the work area is divided into 20 sections A1 to A20. In Fig. 5B, of the divided sections A1 to A20, only sections A9, A18, and A19 where weeds grow are given reference symbols.
[0062] 5C, the weeding device 1 performs a secondary scan of the work area via the imaging device 13 (steps S16 and S17). The weeding device 1 further divides the section where weeds grow into a plurality of small sections (step S16), determines the presence or absence of weeds in each small section, and calculates the occupancy rate of weeds in the small section where weeds grow based on a predetermined formula (step S17). Fig. 5C shows an example in which, for example, of the sections A9, A18, and A19 that were determined to have weeds growing therein in the primary scan, section A18 is divided into a number of small sections a1 to a25. In Fig. 5C, only the small sections a3, a4, a8, a9, a12 to a14, a17, a18, a23, and a24 that are growing therein are labeled with reference symbols.
[0063] 5D, the weeding device 1 specifies the weeding area (step S18). The weeding device 1 compares the weed occupancy calculated in step S17 with a preset threshold, and when the weed occupancy is equal to or greater than the threshold, determines the section in which weeds grow as a weeding area and stores the position information of the weeding area in the memory unit 32. In addition, when there are multiple sections with a weed occupancy rate above a threshold value, the weeding device 1 may determine the priority order in descending order of weed occupancy rate, and associate the position information of the weeding area with the priority and store it in the memory unit 32. As shown in Figure 5D, an example is shown in which the weeding device 1 determines that small sections a3, a8, a12 to a14, a18, and a24, which have a weed occupancy rate equal to or greater than a preset threshold, are weeding areas among the small sections where weeds are growing in clusters.
[0064] As shown in Figures 5E and 5F, the weeding device 1 performs weeding work on weeds in the area (small section) identified as the weeding area (step S19). The weeding device 1 operates the weeding arm 15 to pull out plants determined to be plants other than cultivated plants from the ground as weeds and store them in the storage unit 14. The weeding device 1 performs weeding, for example, by a gripping operation (Figure 5E) in which the first gripping arm 151A and the second gripping arm 151B of the weeding arm 15 grip weeds together with their roots, and a rotation operation (Figure 5F) in which the gripping arm rotation unit 152 is rotated while the first gripping arm 151A and the second gripping arm 151B grip the weeds to pull them out of the ground. As a result, the weeding device 1 can be automatically driven to perform weeding work.
[0065] In addition, when the weeding device 1 is configured not to perform a secondary scan of the working area, the weeding device 1 obtains the classification results of the plants growing in the working area photographed immediately before by the primary scan, and weeds the plants determined to be "weeds" with the weeding arm 15.
[0066] (1.5) Image Recognition Programs An automatic weeding program executed by the weeding device 1 according to this embodiment will be described. The weeding device 1 automatically performs weeding work according to a program that causes a computer to execute at least each of the following operations (e) to (h). The following program is non-temporarily recorded on a recording medium such as a hard disk drive or memory, or an optical disc such as a DVD disc or Blu-ray (registered trademark). The following program may be distributed via the Internet. Furthermore, the following program may be recorded on a cloud server and executed via the Internet.
[0067] (e) Obtaining image data of the work area (f) inputting the image data into a trained model capable of classifying plants shown in the image represented by the image data into either cultivated plants grown in a work area or non-cultivated plants based on the image data, and obtaining, as an output result from the trained model, a classification result in which the plants shown in the image are classified into either cultivated plants or non-cultivated plants; (g) To carry out weeding work by pulling out or cutting plants that are determined to be plants other than cultivated plants based on the classification results. (h) Containing weeds in a container.
[0068] (1.6) Variations In the weeding device 1 described above, a weeding arm 15 including a gripping arm 151 having a first gripping arm 151A and a second gripping arm 151B, and a gripping arm rotating part 152 for rotating the gripping arm 151 is used for weeding and collecting weeds, but the configuration is not limited to this. For example, a weeding rotor that is cylindrical or columnar, has a mechanism for entangling weeds on its surface, and rotates relative to the ground may be provided instead of the weeding arm 15. A weeding rotor is preferable in that it can efficiently remove weeds in areas where weeds grow in clusters over the entire surface, but it becomes difficult to remove weeds in areas where cultivated plants and weeds are mixed.
[0069] In addition, in the above-mentioned weeding device 1, the weeding section that performs the weeding work of pulling out or cutting weeds and the collecting section that performs the storage work of collecting weeds and storing them in the storage section 14 are realized only by the weeding arm 15, but this configuration is not limited to this. For example, the weeding rotor may be configured to include the weeding rotor described above and a collection section that is provided separately from the weeding rotor and collects the weeds pulled out by the weeding rotor.
[0070] (1.7) Effects of the First Embodiment The weeding device 1 and the automatic weeding method and automatic weeding program executed by the weeding device 1 according to the first embodiment have the following effects. (1) The weeding device 1 classifies plants appearing in the image of the working area acquired by the data acquisition unit 31 into either cultivated plants or non-cultivated plants using a trained model, and based on the classification results, removes plants determined to be non-cultivated plants as weeds and stores them in the storage unit 14. This improves weeding efficiency because no weeds remain in the cultivation area after weeding and there is no need to collect the weeds.
[0071] (2) The weeding device 1 is equipped with a weeding area identification unit that identifies an area in the work area where weeds grow in clusters based on image data acquired by the data acquisition unit and the classification results of the classification unit, and designates the area where weeds grow in clusters as a weeding area, and the weeding unit is configured to perform weeding work in the weeding area. This allows the area where weeding work will be performed to be identified on a section-by-section basis, improving weeding efficiency.
[0072] (3) The weeding area identifying unit of the weeding device 1 defines the scope of a work area (e.g., farmland) based on image data, divides the work area into a number of sections, and identifies sections where weeds grow in clusters as weeding areas. This allows the area where weeding work will be performed to be identified on a section-by-section basis, improving weeding efficiency.
[0073] (4) The weeding device 1 is equipped with an occupancy calculation unit that further divides the weed-infested area identified by the weeding area identification unit into a number of small sections, determines the presence or absence of weeds in each small section, and calculates the occupancy rate of weeds in the weed-infested area from the number of small sections in which weeds exist relative to the total number of small sections. This makes it possible to adjust the priority of weeding based on the occupancy rate of weeds, thereby improving weeding efficiency.
[0074] (5) The weeding device 1 is equipped with an occupancy calculation unit that calculates the area of the section in which weeds grow and the area of the region in which weeds grow within the section in which weeds grow, and calculates the occupancy rate of weeds in the section in which weeds grow from the area of the region relative to the area of the section. This makes it possible to adjust the priority of weeding based on the occupancy rate of weeds, thereby improving weeding efficiency.
[0075] (6) The weeding device 1 is equipped with a weeding determination unit that compares the occupancy rate of weeds with a preset threshold value, and determines that the section in which weeds grow in clusters is a weeding area if the occupancy rate of weeds is equal to or greater than the threshold value. This allows sections with low weed occupancy to be excluded and sections with high weed occupancy to be prioritized for weeding, thereby improving weeding efficiency.
[0076] 2. Second embodiment Hereinafter, the weeding device 2 according to the second embodiment will be described with reference to FIGS. 7A, 7B, and 8. FIG. The weeding device 2 of this embodiment, like the weeding device 1, is a device that uses a trained model to classify plants in a working area and automatically (i.e., without the operator operating the device or performing other operations) weeding work by treating plants other than cultivated plants as weeds. 7A and 7B, the weeding device 2 is composed of a colony area specifying device 2A that performs operations ranging from collecting image data used as learning data to a primary scan for specifying a colony area, and a weeding device 2B that performs operations ranging from a secondary scan of the working area to weeding work based on the identification result by the colony area specifying device 2A. That is, the weeding device 2 differs from the weeding device 1 according to the first embodiment in that the functions of the weeding device 1 are realized by the colony area specifying device 2A and the weeding device 2B.
[0077] (2.1) Weeding equipment As shown in FIGS. 7A and 7B, the weeding device 2 includes a colony area specifying device 2A and a weeding device 2B, as described above. The colony area specifying device 2A differs from the weeding device 1 in that it does not have a storage section for storing weeds and a weeding arm 15. The weeding device 2B has a similar configuration to the weeding device 1 in appearance.
[0078] (2.2) Functional configuration of the weeding device Fig. 8 is a schematic functional block diagram showing an example of the functional configuration of the colony area identifying device 2A and the weeding device 2B. Each functional block shown in Fig. 8 is realized by the processor 21 of the arithmetic device 20 described in the first embodiment using Fig. 4 executing a necessary program and controlling the operation of each part of the colony area identifying device 2A and the weeding device 2B.
[0079] The colony area identifying device 2A according to this embodiment has a data acquiring unit 31A, a memory unit 32A, a machine learning unit 33A, a weeding area identifying unit 34A, a position information detecting unit 35A, and an imaging device 13A. The weeding device 1 is connected to an input device 16A and a display device 17A, which are external devices. The input device 16A and the display device 17A may be provided as part of the weeding device 1. The input device 16A and the display device 17A may be provided in common with an input device 16B and a display device 17B of the weeding device 2B, which will be described later.
[0080] The colony growth area identifying device 2A differs from the weeding device 1 in that it is not provided with the weeding arm 15 and the control unit 36 and in that the weeding area identifying unit 34 does not have a functional unit for calculating the occupancy rate of weeds. The colony growth area identifying device 2A also differs from the weeding device 1 in that the data acquiring unit 31A is formed so as to be connectable to the data acquiring unit 31B of the weeding device 2B via a communication interface not shown. Furthermore, other functional units are common to the weeding device 1, and the memory unit 32A, machine learning unit 33A, position information detection unit 35A, imaging device 13A, input device 16A and display device 17A have similar functions to the memory unit 32, machine learning unit 33, position information detection unit 35, imaging device 13, input device 16 and display device 17 of the weeding device 1, so their explanations are omitted.
[0081] The weeding device 2B according to this embodiment has a data acquisition unit 31B, a memory unit 32B, a machine learning unit 33B, a weeding area identification unit 34B, a position information detection unit 35B, a control unit 36B, a weeding arm 15, an imaging device 13B, and a communication interface. The weeding device 1 is connected to an input device 16B and a display device 17B, which are external devices. The input device 16B and the display device 17B may be provided as part of the weeding device 1.
[0082] The weeding device 2B differs from the weeding device 1 in that the weeding area specifying unit 34 does not have a functional unit for determining the presence of weeds and that the machine learning unit 33 does not have functional units (corresponding to the label assignment unit 331 and the learning unit 333) that contribute to the generation of a trained model. The weeding device 2B also differs from the weeding device 1 in that the data acquisition unit 31B is formed so as to be connectable to the data acquisition unit 31B of the colony area specifying device 2A via a communication interface not shown. In addition, other functional parts are common to the weeding device 1, and the memory unit 32B, the occupancy rate calculation unit 342B, the position information detection unit 35B, the control unit 36B, the weeding arm 15, the imaging device 13B, the input device 16B and the display device 17B have the same functions as the data acquisition unit 31, the memory unit 32, the occupancy rate calculation unit 342, the position information detection unit 35, the control unit 36, the weeding arm 15, the imaging device 13, the input device 16 and the display device 17 of the weeding device 1, so their explanations are omitted.
[0083] The colony area identifying device 2A and the weeding device 2B transmit and receive data via a network 28 through a communication interface (communication interface 26 in FIG. 3) not shown. The network 28 may be either a wired network or a wireless network. If the network 28 is a wireless network such as a wireless LAN, Bluetooth (registered trademark), or a mobile communication network, it is preferable because the colony area identifying device 2A and the weeding device 2B can transmit and receive data even when they are located apart from each other.
[0084] The weeding device 2B receives a copy of the trained model generated by the colony area identifying device 2A and stores it in the trained model storage unit 334B. As a result, the machine learning unit 33B can classify plants shown in the image represented by the image data into either "cultivated plants" or "plants other than cultivated plants (i.e., weeds)" based on the image data, similar to the colony area identifying device 2A.
[0085] Furthermore, the weeding device 2B receives the scan results (areas with weeds) of the primary scan generated by the colony area identifying device 2A and stores them in the memory unit 32B. This enables the weeding device 2B to calculate the occupancy rate of weeds in the occupancy rate calculation unit 342B using the scan results of the primary scan stored in the memory unit 32B.
[0086] In the weeding device 2 composed of the colony area identifying device 2A and the weeding device 2B, the colony area identifying device 2A and the weeding device 2B can each perform their own work by moving independently. Therefore, the work by the colony area identifying device 2A (from collecting image data for learning data to the primary scan) and the work by the weeding device 2B (from the secondary scan to the weeding work) can be performed in parallel. In addition, data can be transmitted and received between the colony area identifying device 2A and the weeding device 2B in real time via the communication interface. Therefore, the weeding work can be performed automatically without stopping the colony area identifying device 2A and the weeding device 2B for synchronization, and the weeding efficiency is further improved.
[0087] (2.3) Automatic weeding method The automatic weeding method according to this embodiment will be described below. The automatic weeding method according to this embodiment is carried out at least by the following method, as described in the first embodiment. (a) Acquire image data of the work area (b) Inputting the image data into a trained model capable of classifying plants shown in the image shown by the image data into either a "cultivated plant" cultivated in a working area or a "non-cultivated plant" based on the image data, and obtaining a classification result as an output result from the trained model that classifies the plants shown in the image into either a "cultivated plant" or a "non-cultivated plant". (c) Based on the classification results, plants that are determined to be non-cultivated plants are weeded by pulling them out or cutting them from the ground. (d) Containing the weeds in a container
[0088] Hereinafter, the automatic weeding method will be described in more detail with reference to FIGS. The flowchart shown in Fig. 9 shows the operation of the colony area identification device 2A. The colony area identification device 2A travels over the entire farm, which is the work area, to acquire multiple pieces of image data (step S21), generates a trained model (step S22), and performs a primary scan of the work area (steps S23 to S25). These operations are similar to the operations of steps S11 to S15 described in the first embodiment. Thereafter, the colony growth area specifying device 2A copies and transmits the trained model and the result of the primary scan (colony growth area) to the weeding device 2B (step S26). The colony area identifying device 2A continues these operations within the working area.
[0089] The flowchart shown in FIG. 10 illustrates the operation of the weeding device 2B. The weeding device 2B receives the trained model and the result of the primary scan (crowding area) from the colony area specifying device 2A (step S31). Next, the weeding device 2 performs a secondary scan of the work area based on the result of the primary scan (steps S32 and S33), identifies a weeding area (step S34), and performs weeding work in the area identified as the weeding area (step S35). These operations are similar to the operations from step S16 to step S19 described in the first embodiment. As a result, the weeding device 2 can be automatically driven to perform weeding work.
[0090] In addition, when the weeding device 2 is configured not to perform a secondary scan of the working area, the weeding device 2B obtains the classification results of the plants growing in the working area photographed immediately before by the primary scan, and weeds the plants determined to be "weeds" with the weeding arm 15.
[0091] (2.4) Image Recognition Programs An automatic weeding program executed by the weeding device 2 according to this embodiment will be described below. The weeding device 2 automatically performs weeding work according to a program that causes a computer to execute at least each of the operations (e) to (h) described in the first embodiment.
[0092] (2.5) Variations In the second embodiment, a weeding device 2 has been described that is composed of a colony area identification device 2A that performs everything from collecting image data to being used as learning data to a primary scan to identify a colony area, and a weeding device 2B that performs everything from a secondary scan within the working area to weeding work based on the identification results of the colony area identification device 2A, but the configuration is not limited to this. For example, a device may be provided that only collects image data to be used as learning data and generates a trained model, and has only machine learning functions. In this case, another device may perform the primary scan, secondary scan, and weeding, or weeding may be performed using a device that performs only the primary scan and a device that performs the secondary scan and weeding.
[0093] (2.6) Effects of the Second Embodiment The weeding device 2 and the automatic weeding method and the automatic weeding program executed by the weeding device 2 according to the second embodiment have the following effects in addition to the effects described in the first embodiment. (7) The weeding device 2 is composed of a colony area identifying device 2A and a weeding device 2B. This allows the work by the colony area identifying device 2A (from collecting image data for learning data to the primary scan) and the work by the weeding device 2B (from the secondary scan to the weeding work) to be performed in parallel, further improving weeding efficiency. Also, since data can be sent and received in real time between the colony area identifying device 2A and the weeding device 2B via the communication interface, there is no need to stop the colony area identifying device 2A and the weeding device 2B for synchronization, and this does not hinder the improvement of weeding efficiency.
[0094] Although the embodiments of the present invention have been described above, the technical scope of the present invention is not limited to the technical scope described in the above-mentioned embodiments. Various modifications or improvements can be made to the above-mentioned embodiments, and it is clear from the claims that such modifications or improvements can also be included in the technical scope of the present invention. [Explanation of symbols]
[0095] 1,2,2B Weeding equipment 2A Colony Area Identification Device 11 Aircraft 12 wheels 12A front wheel 12B rear wheel 13, 13A, 13B Imaging device 14 Storage unit 15 Weeding arm 16,16A,16B Input devices 17,17A,17B Display device 18 Computers 20 Arithmetic unit 21 Processors 24 Auxiliary storage 25 External Interface 26 Communication Interface 27 Bus 28 Network 31, 31A, 31B Data acquisition section 32,32A,32B Storage section 33, 33A, 33B Machine Learning Department 34, 34A, 34B Weeding area specific section 35, 35A, 35B Position information detection unit 36, 36B Control section 151 Grasping Arm 151A First gripping arm 151B Second gripping arm 152 Grip arm rotation part 331 Labeling Department 332 Learning Data Extraction Unit 333 Learning Department 334,334B Learning model memory section 335 Classification Department 341 Weed Presence Determination Unit 342,342B Occupancy rate calculation section 35 Location information detection unit 351 Gyro Sensor 352 GPS 353 Location information calculation unit Sections A1 to A20 a1~a25 Small plots
Claims
1. The aircraft and Wheels that run the aircraft; A data acquisition unit that acquires image data of a working area; A classification unit that has a trained model capable of classifying a plant shown in an image represented by the image data into either a cultivated plant cultivated in the working area or a plant other than the cultivated plant based on the image data, and classifies the plant shown in the image into either the cultivated plant or the plant other than the cultivated plant based on the image data acquired by the data acquisition unit; a weeding area specifying unit that specifies a section of the work area in which weeds grow, the weeds being plants determined to be plants other than the cultivated plants, based on the image data acquired by the data acquisition unit and the classification result of the classification unit, and specifies the section in which the weeds grow as a weeding area; A storage section provided on the body and capable of storing plants; A weeding unit that performs weeding work by pulling out or cutting the weeds from the ground; A collection unit that collects the weeds and stores them in the storage unit; Equipped with The weeding unit performs weeding work in the weeding area. Weeding equipment.
2. The weeding area specification unit defines a range of the work area based on the image data, divides the work area into a plurality of sections, and specifies the section in which the weeds grow as a weeding area.
2. A weeding device according to claim 1.
3. The weed-growing area specified by the weeding area specifying unit is further divided into a plurality of small sections, the presence or absence of the weeds is determined for each small section, and an occupancy rate calculation unit is provided which calculates an occupancy rate of the weeds in the weed-growing area from the number of small sections in which the weeds exist relative to the total number of the small sections. A weeding device according to claim 2.
4. A position information detection unit capable of acquiring a position and an attitude of the device itself is provided, The occupancy calculation unit corrects an inclination of an image capturing object in the image based on at least one of an imaging height of the image data and a position and an orientation of the own device acquired from the position information detection unit, and divides the image into a plurality of small sections based on the corrected image. A weeding device according to claim 3.
5. an imaging device for capturing the image and generating the image data; The occupancy rate calculation unit corrects the inclination of the object to be imaged by setting the installation height of the imaging device as the imaging height of the image data. A weeding device according to claim 4.
6. an occupancy calculation unit that calculates an area of a section in which the weeds grow and an area of a region in which the weeds grow in the section in which the weeds grow, and calculates an occupancy rate of the weeds in the section in which the weeds grow from the area of the region relative to the area of the section; A weeding device according to claim 2.
7. A weeding determination unit is provided for comparing the occupancy rate of the weeds with a preset threshold value, and determining that the section in which the weeds grow is the weeding area when the occupancy rate of the weeds is equal to or greater than the threshold value. A weeding device according to any one of claims 1 to 6.
8. Obtain image data of the working area; Based on the image data, the image data is input to a trained model capable of classifying a plant shown in the image data into either a cultivated plant cultivated in the working area or a plant other than the cultivated plant, and a classification result in which the plant shown in the image is classified into either the cultivated plant or the plant other than the cultivated plant is obtained as an output result from the trained model; Based on the acquired image data and the classification result, a section in the work area in which weeds, which are plants determined to be plants other than the cultivated plants, grow is identified, and the section in which the weeds grow is designated as a weeding area; In the weeding area, a weeding operation is performed to pull out or cut the weeds from the ground; The weeds are accommodated in a storage section capable of accommodating the plants. Automatic weeding method.
9. acquiring image data of a working area; Inputting the image data into a trained model capable of classifying a plant shown in an image represented by the image data into either a cultivated plant cultivated in the working area or a plant other than the cultivated plant based on the image data, and acquiring a classification result in which the plant shown in the image is classified into either the cultivated plant or the plant other than the cultivated plant as an output result from the trained model; Identifying a section of the work area in which weeds, which are plants determined to be plants other than the cultivated plants, grow in colonies based on the acquired image data and the classification result, and designating the section in which weeds grow in colonies as a weeding area; In the weeding area, a weeding operation is performed to pull out or cut the weeds from the ground; accommodating the weeds in a storage section capable of accommodating the plants; An automatic weeding program that causes a computer to execute the following.
Citation Information
Patent Citations
Weeding and weeder
JP1992011803A
Method for discriminating and removing weeds
JP1997006949A
Weed recognizing device of weeding robot
JP1997081750A
Method and system for managing farmland
JP2013059296A
Robotic vehicles and methods of using robots for automated processing of plant organisms
JP2018525976A
