KR20210054347A - Automatic driving agricultural machine device supporting image-based path recognition and steering angle calculation techniques and operation method thereof - Google Patents

Abstract

The present invention relates to an automatic driving agricultural machine device, and more specifically, to an automatic driving agricultural machine device which is a technique for automatically recognizing a work area through image-based path recognition and steering angle calculation for an automatic driving agricultural machine to perform driving and cultivation work. An automatic driving agricultural machine device comprises: an image detection unit which shoots an image with at least one camera; a pattern learning unit which learns an image based on a deep learning algorithm and recognizes pattern information of a cultivation area based on the learned image; a reference path determining unit which determines boundaries between fields based on pattern information of cultivated areas, and determines a reference path based on boundaries between cultivated areas; a region determining unit which determines region information based on boundaries between regions; and an automatic driving control unit which calculates steering angle information based on the reference path and area information and controls the automatic driving agricultural machine to automatically drive based on the steering angle information. Through the present invention, the work area to perform driving and cultivation work may be automatically recognized.

Description

An autonomous driving agricultural machinery device supporting image-based route recognition and steering angle calculation technology and its operation method {AUTOMATIC DRIVING AGRICULTURAL MACHINE DEVICE SUPPORTING IMAGE-BASED PATH RECOGNITION AND STEERING ANGLE CALCULATION TECHNIQUES AND OPERATION METHOD THEREOF}

The present invention is a self-driving agricultural machine device, and more specifically, is a technology that automatically recognizes a work area through image-based route recognition and steering angle calculation for an autonomously-driving agricultural machine to perform driving and cultivation work.

Recently, the domestic industry is building an industrial system comparable to the level of advanced countries such as the United States, Germany, and Japan by fusion of artificial intelligence and ICT technology through the slogan of the Fourth Industrial Revolution. Such changes in the industrial structure also have an effect on the agricultural field, and the smart farm is a representative example of a new industrial system in the agricultural field.

Self-driving technology is basically applied to automobiles running on the road, but it is also widely used for military purposes and agricultural technologies. Specifically, there is an autonomous tractor as an agricultural machine to which autonomous driving technology is applied. The self-driving tractor is a tractor that moves by itself, and is configured to perform work while moving along the work path by measuring the position and size of the cultivated land and setting a work path suitable for the cultivated land.

On the other hand, research related to autonomous tractors has been actively conducted in Korea, and technologies such as automatic route generation have been developed, but compared to advanced countries, the accuracy of the measurement position, field conditions, and monitoring of major tractor elements are not considered largely, so work stability. In addition, it is difficult for the operator to grasp the current operating conditions of each tractor position by focusing on securing stability due to steering and speed when driving the tractor.

In other words, since most of the unmanned tractors are not driven by humans directly, variables are generated according to the working environment, and the driving speed and steering angle may change according to the variables, and the condition of the devices for each tractor part may change. This can cause off-road or tractor stability problems.

An object of the present invention is to provide an autonomous farming machine system capable of autonomously performing cultivation and driving by recognizing whether or not cultivation is carried out and calculating a steering angle.

The autonomous driving agricultural machine apparatus according to an embodiment of the present invention is an image detector that captures an image with at least one camera, learns an image based on a deep learning algorithm, and learns an image based on the learned image. A pattern learning unit that recognizes pattern information, a reference path determination unit that determines boundaries between regions based on pattern information of cultivated regions, and determines a reference path based on the boundaries between cultivated regions, and provides region information based on the boundaries between regions. And an autonomous driving control unit that calculates steering angle information based on the determined region determination unit and reference path and region information, and controls the autonomously driving agricultural machine to autonomously travel based on the steering angle information.

In an embodiment of the present invention, the pattern learning unit recognizes pattern information including at least one of color, shape, material, or structure through a deep learning algorithm.

In one embodiment of the present invention, the region information includes first to third region information. The first area is cultivated land, the second is uncultivated land, and the third is uncultivated.

In an embodiment of the present invention, the autonomous driving control unit extracts steering angle information by linearly converting the reference path into steering angle information based on the slope of the reference path.

In one embodiment of the present invention, the steering angle information is characterized in that the angle at which the spindle of the steering wheel turns when the autonomous driving agricultural machine changes direction.

In an embodiment of the present invention, the autonomous driving control unit includes information on the first route to the third route, and the first route information is information for the autonomous driving agricultural machine to cultivate the uncultivated land when only the boundary and the uncultivated exist. The second route information is information that the autonomously driving agricultural machine rotates based on the steering angle information, and the third route information includes the third route through which the autonomously driving agricultural machine cultivates uncultivated land when there are cultivated land and uncultivated land.

In an embodiment of the present invention, when only the uncultivated area and the uncultivated exist, the autonomous driving agricultural machine cultivates the uncultivated land according to the uncultivated area based on the first route information, and the autonomous driving agricultural machine is, after recognizing the uncultivated area, It rotates based on the steering angle information based on the second route information, and the autonomous driving agricultural machine cultivates the uncultivated land based on the third route information when there are cultivated land and uncultivated land.

In an embodiment of the present invention, the autonomous driving control unit cultivates the cultivated land and part of the uncultivated land based on the third path.

A method of operating an autonomous driving agricultural machine device according to an embodiment of the present invention, comprising: capturing an image with at least one camera by an image detection unit, and by a pattern learning unit learning an image based on a deep learning algorithm, and learning Recognizing the pattern information of the cultivated area based on the cultivated image, the step of determining the boundary between the areas based on the pattern information of the cultivated area by the reference path determination unit, and determining the reference path based on the boundary between the cultivated areas, and determining the area Determining area information based on the boundary between the additional areas, the autonomous driving control unit calculating steering angle information based on the reference route and area information, and controlling the autonomously driving agricultural machine to autonomously travel based on the steering angle information. .

In an embodiment of the present invention, the calculating of the steering angle information includes a step of autonomously driving by an autonomous driving farm machine based on the steering angle information, the step of calculating the slope of the reference path by the autonomous driving controller, and the slope of the reference path. And extracting and linearly converting the reference path into steering angle information.

An autonomous driving farm machine system according to an embodiment of the present invention includes an autonomous driving farm machine device and a camera for photographing a cultivated area, and the autonomous driving farm machine device includes an image detection unit, deep learning ( Deep Learning), a pattern learning unit that learns images based on algorithms and recognizes pattern information of cultivated areas based on the learned images, determines boundaries between areas based on pattern information of cultivated areas, and determines boundaries between cultivated areas. Based on the reference route determination unit for determining the reference route, the region determining unit for determining region information based on the boundary between the regions, and the steering angle information based on the reference route and region information, the autonomously driving agricultural machine is able to calculate the steering angle information based on the steering angle information. It includes an autonomous driving control unit that controls autonomous driving.

In a computer-readable recording medium including a program for executing an autonomous driving agricultural machine device according to an embodiment of the present invention, the autonomous driving agricultural machine device method comprises: capturing an image with at least one camera by an image detection unit on a computer, A pattern learning unit learns an image based on a deep learning algorithm, and recognizes pattern information of a cultivation area based on the learned image, and a reference path determination unit determines the boundary between the areas based on the pattern information of the cultivation area. And determining the reference route based on the boundary between the cultivated areas, the region determining unit determining region information based on the boundary between the regions, and the autonomous driving control unit calculating steering angle information based on the reference route and region information And controlling the autonomous driving agricultural machine to autonomously travel based on the steering angle information.

Through the present invention, it is possible to automatically recognize the work area and perform driving and cultivation work.

1 is a block diagram of an autonomous driving agricultural machine device according to an embodiment of the present invention.
2 is a block diagram of an autonomous driving agricultural machine system according to an embodiment of the present invention.
3 is a view showing a cultivation area of the cultivation area unit according to an embodiment of the present invention.
4 to 6 are diagrams illustrating a driving state of an autonomous driving agricultural machine according to route information of an autonomous driving controller according to an embodiment of the present invention.
7 is a flow chart showing the operation of the autonomous driving agricultural machine device step by step according to an embodiment of the present invention.

Hereinafter, embodiments disclosed in the present specification will be described in detail with reference to the drawings. Unless otherwise specified in the text, like reference numbers in the drawings indicate like elements. The exemplary embodiments described in the detailed description, drawings, and claims are not intended to be limiting, and other embodiments may be used, and other modifications may be made without departing from the spirit or scope of the technology disclosed herein. Those skilled in the art may arrange, construct, combine, and design the components of the present disclosure, that is, the components generally described herein and described in the drawings in various different configurations, all of which are clearly devised, and the present disclosure It will be easy to understand that it forms a part of. In the drawings, in order to clearly express various layers (or films), regions, and shapes, the width, length, thickness, or shape of components may be exaggerated and expressed.

Since the description of the disclosed technology is merely an embodiment for structural or functional description, the scope of the rights of the disclosed technology should not be construed as being limited by the embodiments described in the text. That is, since the embodiments can be modified in various ways and have various forms, the scope of rights of the disclosed technology should be understood to include equivalents capable of realizing the technical idea.

Singular expressions are to be understood as including plural expressions unless the context clearly indicates otherwise, and terms such as “comprises” or “have” refer to implemented features, numbers, steps, actions, components, parts, or It is to be understood that the combination is intended to designate the existence of, and does not preclude the possibility of the presence or addition of one or more other features or numbers, steps, actions, components, parts, or combinations thereof.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the field to which the disclosed technology belongs, unless otherwise defined. As defined in a commonly used dictionary, terms should be construed as having the meaning of the related technology in context, and cannot be construed as having an ideal or excessively formal meaning unless explicitly defined in the present application.

1 is a block diagram of an autonomous driving agricultural machine device 100 according to an embodiment of the present invention.

2 is a block diagram of an autonomous driving agricultural machine system 10 according to an embodiment of the present invention.

1 and 2, the autonomous driving agricultural machine apparatus 100 includes an image detection unit 110, a pattern learning unit 120, a reference path determination unit 130, an area determination unit 140, and an autonomous driving control unit ( 150). The autonomous driving agricultural machine system 10 includes an autonomous driving agricultural machine device 100 and a camera 20.

The image detector 110 captures an image of a cultivated area with at least one camera 20. At this time, the camera 20 is an RGB camera, IP camera, HD-SDI camera, analog camera, fire detection color camera, thermal imaging camera, HD (1920x5080, HD5080p) camera, IP zoom at SD (720x486, NTSC) level resolution. It includes at least one of a speed camera or a CCTV camera. Production costs and additional technology development costs can be reduced through technology that can be used for autonomous driving using only the camera 20 without using expensive equipment such as RTK-GPS or laser sensors.

The image detection unit 110 can secure a better image quality as the installation height of the camera 20 increases, and in this way, the selection of a location where each camera 20 can eliminate the blind spot of the vehicle and the combined surroundings Since it is important to set an installation location and a viewing angle that can minimize image quality deterioration, and can be set differently according to a vehicle type and a size of the vehicle, the present invention is not limited thereto.

In addition, if the camera 20 is installed in the front, rear, left, and right of the autonomous vehicle, the image captured through the four cameras 20 can be corrected, and the 360° periphery of the autonomous farming machine can be checked. . At this time, it is advantageous to combine the images and extract the borders around the front camera 20 and the rear camera 20 to have the same height to be installed and to have the same height for the left and right cameras 20 to be installed. .

The pattern learning unit 120 learns an image based on a deep learning algorithm, and recognizes pattern information of a cultivation area based on the learned image.

The pattern learning unit 120 extracts a boundary between regions from image information captured by at least one camera 20 and identifies a pattern between the boundaries. It may also include a technology for calculating distance information between the boundaries of the cultivated area and the autonomously driving agricultural machine 30.

The pattern learning unit 120 generally learns pattern information through a deep learning algorithm in order to solve the problem that the error rate between regions is high when image processing is performed based on similar colors and shapes.

At this time, the deep learning algorithm is an algorithm that learns the data extraction itself by including the preprocessing process in machine learning in the neural network architecture. In addition, machine learning algorithms derive inference results by calculating a linear combination of values related to each feature, whereas deeper algorithms can achieve a high level of abstraction through a combination of nonlinear transformation techniques. In other words, the deep learning algorithm can automatically perform a task of summarizing at least one of the core contents or functions in a large amount of data or complex data without a pre-processing process.

The pattern learning unit 120 recognizes pattern information of a cultivation area photographed by the image detection unit 110 based on the learned pattern information. In this case, the pattern information includes at least one of color, shape, material, or structure.

The reference path determination unit 130 determines a boundary between regions based on pattern information of the cultivated regions, and determines a reference path based on the boundary between the cultivated regions.

The reference path determination unit 130 divides the image obtained through the camera 20 into boundaries between regions based on preset pattern information. The traveling direction of the autonomously driving agricultural machine 30 is determined based on the divided boundary.

For example, when determining the boundary between the areas according to the color of the cultivated area, if there is a dark area, a non-dark area, and a different color area, it is determined that the first to third area boundary information exists according to the color. . At this time, the first region boundary information is a boundary between a dark color region and a non-dark color region, the second region boundary information is a boundary between a non-dark color region and a different color region, and the third region boundary information is a color. It is the boundary between dark areas and areas with different colors. The reference path determination unit 130 sets an existing path based on the first to third area boundary information.

The region determining unit 140 determines region information based on boundaries between regions.

The region information can be determined based on the boundary between the regions, and the boundary between the regions is determined based on the pattern information of the cultivated region. A boundary between regions is determined by comparing preset pattern information with a photographed image, and a region is determined by comparing the determined boundary between regions with preset region information. At this time, the preset area information is area information learned by pattern information.

The region determining unit 140 includes first to third region information. The first to third region information is classified based on the first to third region boundary information of the reference path determining unit 130. At this time, if the dark area is the first area, the dark area is the second area, and the different color area is the third area, the first area is the arable land 42, and the second area is uncultivated land ( 43), and the third area may be a boundary.

The area determining unit 140 classifies the cultivated area into at least one or more areas. The cultivation areas can be classified so that the number of travel paths included in each area becomes a predetermined value equal to each other. In addition, even when the cultivation is stopped halfway, the part in which the points of the cultivated land 42 and the uncultivated land 43 alternately appear in the cultivated area can be controlled in a small range. Therefore, the division between the cultivated land 42 and the uncultivated land 43 is easy to become clear, and the cultivation can be smoothly resumed.

The autonomous driving control unit 150 calculates steering angle information based on the reference route and the region information, and controls the autonomous driving agricultural machine 30 to autonomously travel based on the steering angle information.

The autonomous driving controller 150 extracts steering angle information by linearly converting the reference path into steering angle information based on the slope of the reference path. At this time, the steering angle information is an angle at which the spindle of the steering wheel turns when the autonomous driving agricultural machine 30 changes direction.

When the autonomous driving control unit 150 generates a work path, the most important consideration is at least one of a distance between cultivation sections, a turning method and turning distance, the number of turning tasks, or a driving path and sequence. The distance between the working sections is because it is necessary to consider the inherent working width of the autonomous farming machine 30 and the cultivator, and if this is not taken into account, the phenomenon of cultivation overlap occurs.

In addition, due to the structure of the autonomous driving agricultural machine 30, not only is the working width and length larger than that of a general vehicle, but also has a larger weight, and cannot be easily reversed or rotated, so an optimal turning path and an optimal number of turning operations must be considered. And if the calculation of the driving route and the order is poor, there is a section where the work overlaps or a space where work is not possible. Therefore, it is necessary to create a cultivation path by considering all of these factors.

The autonomous driving control unit 150 includes information on the first route to the third route, and the first route information means that when only the uncultivated area 41 and the uncultivated land 43 exist, the autonomous farming machine It is information to cultivate. The second route information is information that the autonomously driving agricultural machine 30 rotates based on the steering angle information, and the third route is the self-driving agricultural machine 30 when the cultivated land 42 and the uncultivated land 43 exist. It is information to cultivate). Details of a method for the autonomous driving controller 150 to cultivate based on information on the first route to the third route will be described later with reference to FIGS. 4 to 6.

3 is a view showing a cultivation area of the cultivation area unit according to an embodiment of the present invention.

1 to 3, the area determination unit 140 determines area information based on the boundary between areas determined by the reference path determination unit 130 in the cultivation area.

The region information can be determined based on the boundary between the regions, and the boundary between the regions is determined based on the pattern information of the cultivated region. A boundary between regions is determined by comparing preset pattern information with a photographed image, and a region is determined by comparing the determined boundary between regions with preset region information. At this time, the preset area information is area information learned by pattern information.

The region determining unit 140 includes first to third region information. The first to third region information is classified based on the first to third region boundary information of the reference path determining unit 130. For example, if a dark area is a first area, a non-dark area is a second area, and a different color area is a third area, the first area is the arable land 42, and the second area is The cultivated land 43 may be, and the third region may be an uncultivated region 41.

In this case, in the third region, the autonomously driving agricultural machine 30 may recognize the third region while driving the cultivated region based on a preset terrain pattern. By first recognizing the third area based on a preset topographic pattern, the cultivated area can be established, and the established cultivated area can be divided into a first area and a second area.

4 to 6 are views showing a state of driving the autonomous driving agricultural machine 30 according to route information of the autonomous driving control unit 150 according to an embodiment of the present invention.

4 to 6, FIG. 4 is a view of the autonomous driving agricultural machine 30 driving according to the first route information of the autonomous driving control unit 150, and FIG. 5 is the second route information of the autonomous driving control unit 150. It is a state of driving of the autonomously driving agricultural machine 30 according to, and FIG. 6 is a state of driving of the autonomously driving agricultural machine 30 according to the third route information of the autonomous driving control unit 150.

Referring to FIG. 4, the first path information is information on which autonomously oriented agricultural machinery cultivates the uncultivated land 43 when only the uncultivated area 41 and the uncultivated land 43 exist.

The first route information is generally the case of first cultivation in the cultivated area, and the direction of the autonomous driving agricultural machine 30 in the section in which the uncultivated area 43 and the uncultivated area 41 exist is indicated in the uncultivated area 41 It is determined to comply with, and based on the direction of progress, the autonomous driving agricultural machine 30 proceeds, and the cultivation of the uncultivated land 43 proceeds.

Referring to FIG. 5, the second route information is information about which the autonomously driving agricultural machine 30 rotates based on the steering angle information.

When the second path information recognizes the uncultivated area 41 during cultivation, the uncultivated land 43 is rotated based on the steering angle information in order to cultivate the uncultivated land 43 and the cultivation of the uncultivated land 43 is proceeded. At this time, the steering angle information is characterized in that the angle at which the spindle of the steering wheel turns when the autonomous driving agricultural machine 30 changes direction. The steering angle information is linearly converted to a steering angle based on the slope of the boundary between regions determined by the reference path determining unit 130 to extract steering angle information.

For example, when an autonomous vehicle traveling straight ahead needs to turn to a corner turning to the left, the steering angle of the left side is restricted while limiting the steering angle to the right of the wheel of the autonomous vehicle based on the reference route, area information, and steering angle information. You will set the angle.

In addition, if the driving direction of the autonomous vehicle is calculated based on the speed and steering angle information of the autonomous farming machine 30, it is possible to infer how long the autonomously driving farming machine 30 will contact the uncultivated area 41.

Referring to FIG. 6, the third route information is information on which the autonomously driving agricultural machine 30 cultivates the uncultivated land 43 when the cultivated land 42 and the uncultivated land 43 exist.

The third route information is the uncultivated area 41 of the cultivated land 42 and the uncultivated land 43 after cultivation proceeds with the first route information, encounters the uncultivated area 41 and rotates based on the second route information. The cultivation of the uncultivated land 43 is in progress. When cultivation is performed based on the third route information, a path is set so that the cultivated land 42 and the uncultivated land 43 overlap so that correction recognition can be performed. Accurate cultivation can be carried out through correction recognition.

7 is a flow chart showing the operation of the autonomous driving agricultural machine device 100 step by step according to an embodiment of the present invention.

Referring to FIG. 7, in step S21, the image detection unit photographs a cultivated area to be cultivated.

For example, if the camera 20 is installed in the front, rear, left, and right of the autonomous farming machine 30, the image captured through the four cameras 20 can be corrected, and the autonomous driving farming machine 30 You can check the 360° surrounding of.

In step S22, the pattern learning unit 120 learns a pattern based on the captured image.

For example, the pattern learning unit 120 learns image information captured by at least one camera 20 based on a deep learning algorithm.

In step S23, the reference path determination unit 130 divides the boundary between regions on the basis of a preset pattern on the captured image.

For example, the pattern information of the cultivation area is recognized based on the learned image, and the boundary between the areas of the cultivation area is extracted based on the recognized pattern information. In this case, the pattern information includes at least one of color, shape, material, or structure.

In step S24, the reference path determination unit 130 sets an existing path based on the boundary between regions.

For example, when the reference path determining unit 130 determines the boundary between the areas according to the color of the cultivated area, if there are areas with a dark color, a non-dark area, and a different color area, the first to third areas according to the color It is determined that area boundary information exists. The reference path determination unit 130 sets an existing path based on the first to third area boundary information.

In step S25, the region determining unit 140 determines region information based on the boundary between regions.

For example, it includes first to third region information, and the first to third region information is classified based on the first to third region boundary information of the reference path determination unit 130. In this case, the first region may be the cultivated land 42, the second region may be the uncultivated land 43, and the third region may be the uncultivated region 41.

In step S26, the autonomous driving control unit 150 drives the cultivated area based on the route information.

For example, the autonomous driving controller 150 extracts steering angle information by linearly converting the reference path into steering angle information based on the slope of the reference path. At this time, the steering angle information is the angle at which the spindle of the steering wheel turns when the autonomous driving agricultural machine 30 changes direction. The autonomous driving control unit 150 includes information on the first route to the third route and controls the spin information of the wheel of the autonomously driving agricultural machine 30 based on steering angle information required for driving in the cultivated area.

The present application is a method of operating an autonomous driving agricultural machine device on a computer, wherein the image detection unit captures an image with at least one camera, the pattern learning unit learns an image based on a deep learning algorithm, and Recognizing the pattern information of the cultivated area based on the pattern information, the reference path determining unit determining a boundary between the areas based on the pattern information of the cultivating area, determining a reference path based on the boundary between the cultivating areas, and the area determining unit In order to execute the step of determining area information based on the boundary, the autonomous driving control unit calculating steering angle information based on the reference route and area information, and controlling the autonomous driving agricultural machine to autonomously travel based on the steering angle information. Provides stored computer programs.

10: autonomous driving agricultural machine system
20: camera
30: autonomous driving agricultural machinery
41: uncultivated area
42: arable land
43: Uncultivated land
100: autonomous driving agricultural machinery device
110: image detection unit
120: Pattern Learning Department
130: reference path determination unit
140: area determination unit
150: autonomous driving control unit

Claims (12)

An image detector for capturing an image with at least one camera;
A pattern learning unit that learns the image based on a deep learning algorithm and recognizes pattern information of a cultivation area based on the learned image;
A reference path determining unit determining a boundary between regions based on the pattern information of the cultivated regions and determining a reference path based on the boundary between the cultivating regions;
A region determining unit that determines region information based on the boundary between the regions; And
An autonomous driving control unit that calculates steering angle information based on the reference path and the region information, and controls autonomously driving agricultural machines to autonomously travel based on the steering angle information;
Self-driving agricultural machine device comprising a. The method of claim 1,
The pattern learning unit recognizes pattern information including at least one of color, shape, material, or structure through the deep learning algorithm. The method of claim 1,
The region information includes first to third region information,
The first area is arable land,
The second area is uncultivated land,
The third area is an uncultivated area, the autonomous driving agricultural machine device. The method of claim 1,
The autonomous driving control unit,
An autonomous driving agricultural machine device for extracting the steering angle information by linearly converting the reference path into the steering angle information based on the slope of the reference path. The method of claim 4,
The steering angle information is an autonomous driving farm machine apparatus, characterized in that when the autonomous driving farm machine changes direction, the spindle of the steering wheel turns. The method of claim 4,
The autonomous driving control unit includes first route to third route information,
The first route information is information on which the autonomous driving agricultural machine cultivates the uncultivated land when there is only an uncultivated area and an uncultivated field,
The second route information is information that the autonomously driving agricultural machine rotates based on steering angle information,
The third route information includes a third route through which the autonomous driving agricultural machine cultivates uncultivated land when there are cultivated land and uncultivated land. The method of claim 6,
When only the uncultivated area and the uncultivated exist, the autonomous driving agricultural machine cultivates the uncultivated land according to the uncultivated area based on the first route information,
The autonomous driving agricultural machine rotates based on steering angle information based on the second route information after recognizing the uncultivated area,
The autonomous driving agricultural machine is an autonomous driving agricultural machine device that cultivates the uncultivated land based on the third route information when the cultivated land and the uncultivated land exist. The method of claim 7,
The autonomous driving control unit cultivates the cultivated land and a part of the uncultivated land based on the third path. In the operation method of the autonomous driving agricultural machinery device,
Capturing an image with at least one camera by an image detector;
Learning the image based on a deep learning algorithm by a pattern learning unit and recognizing pattern information of a cultivation area based on the learned image;
Determining, by a reference path determination unit, a boundary between regions based on pattern information of the cultivated regions, and determining a reference path based on the boundary between the cultivated regions;
Determining, by an area determining unit, area information based on the boundary between the areas; And
Calculating, by an autonomous driving control unit, steering angle information based on the reference path and the region information, and controlling the autonomous driving agricultural machine to autonomously travel based on the steering angle information;
The operation method of the autonomous driving agricultural machine device comprising a. The method of claim 9,
The step of calculating the steering angle information,
Autonomously driving, by the autonomously driving agricultural machine, based on the steering angle information;
Calculating, by the autonomous driving controller, a slope of the reference path; And
And extracting and linearly converting the reference route into steering angle information based on a slope of the reference route. Autonomous driving agricultural machinery devices; And
Including a camera for photographing the cultivated area,
The autonomous driving agricultural machinery device,
An image detector for capturing an image with at least one camera;
A pattern learning unit that learns the image based on a deep learning algorithm and recognizes pattern information of a cultivation area based on the learned image;
A reference path determining unit determining a boundary between regions based on the pattern information of the cultivated regions and determining a reference path based on the boundary between the cultivating regions;
A region determining unit that determines region information based on the boundary between the regions; And
An autonomous driving control unit that calculates steering angle information based on the reference path and the region information, and controls autonomously driving agricultural machines to autonomously travel based on the steering angle information;
Self-driving agricultural machine system comprising a. On the computer
Capturing an image with at least one camera by an image detection unit,
A pattern learning unit learning the image based on a deep learning algorithm, and recognizing pattern information of a cultivation area based on the learned image,
Determining, by a reference path determination unit, a boundary between regions based on pattern information of the cultivated regions, and determining a reference path based on the boundary between the cultivated regions,
Determining, by an area determining unit, area information based on the boundary between the areas; and
A computer program stored in a medium to execute the step of an autonomous driving controller calculating steering angle information based on the reference route and the region information, and controlling the autonomous driving agricultural machine to autonomously travel based on the steering angle information.

Images