JP2567410B2 - Boundary detection device for crop planting machine - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a crop planting machine such as a rice planting machine,
In order to use it as control information for automatically running along the boundary between the planted part and the unplanted part of the crop planted in a grid along the machine traveling direction and the width direction, Direction and the location corresponding to the boundary between the planted part and the unplanted part planted in a lattice along the lateral direction of the aircraft, from the imaging information taken from the vertical or near-two-dimensional direction from the direction, the boundary The present invention relates to a boundary detection device for a crop planting machine for extracting image information corresponding to.

[Prior Art] In the boundary detection device for a crop planting machine of this type, since the boundary is a crop row heading in the machine advancing direction that is the most unplanted side in the machine lateral width direction, in the past, for example, imaging From the information, the pixel corresponding to the crop located closest to the unplanted side is detected in each of the plurality of pixel rows along the lateral direction of the machine body arranged in the machine body traveling direction.

[Problems to be solved by the invention]

By the way, since the crops are planted in a lattice shape at intervals along the machine body traveling direction and the machine body lateral width direction, if the machine body direction is inclined with respect to the boundary length direction, The crops lined up in the same row will be inclined toward the direction of travel of the aircraft.

When the crops lined up in the same row in the machine width direction are tilted with respect to the machine body travel direction, crops other than the crops located on the boundary are the least visible on each pixel example along the machine body width direction arranged in the machine body travel direction. It may be located on the planting side, and simply extracting the pixel corresponding to the crop located on the most unplanted side on the pixel row along the machine width direction lined up in the machine advancing direction will position it on the boundary. There is a disadvantage that the image information corresponding to only the crops to be extracted cannot be extracted (see FIG. 5 (a)).

The present invention has been made in view of the above circumstances,
The purpose is to enable accurate extraction of image information corresponding to only crops on the boundary regardless of the inclination of the body.

[Means for solving problems]

The characteristic configuration of the boundary detection device for a crop planting machine according to the present invention is as follows.

That is, the crop is spaced apart from the set planting interval along the machine body advancing direction, and the location corresponding to the boundary between the planted part and the unplanted part planted in a state where the set planting width is separated along the machine body lateral width direction, An image pickup means for picking up an image from a vertical direction or a direction close thereto, and a first pixel extracting means for extracting, from the image pickup information, a pixel having information corresponding to the crop among the pixels lined up in the machine body traveling direction and the machine body lateral width direction; From the plurality of pixels extracted by the first pixel extracting means, the position in the machine lateral width direction in each of the pixel rows along the machine lateral direction aligned in the machine traveling direction is the most unplanted side. Second pixel extracting means for extracting pixels, and for extracting those pixels which are not located on the planted side of the most unplanted pixels adjacent to both adjacent unplanted pixels in the airframe traveling direction Located in including the its action and effects are as follows.

[Action]

By extracting the pixels having the information corresponding to the crop from the image information obtained by imaging the part corresponding to the boundary, the image information corresponding to only the crop from which the background is removed can be obtained. Then, from the extracted pixels, the most unplanted side pixels located on the most unplanted side on each pixel row along the lateral width direction of the aircraft arranged in the aircraft traveling direction are extracted, and the aircraft progresses at the most unplanted side pixels. Pixels located at positions that are not on the planted side of the most unplanted pixels on both sides adjacent in the direction are extracted.

In other words, the position in the machine lateral width direction of the most unplanted side pixels located on the most unplanted side on each pixel row along the machine lateral width direction aligned in the machine body traveling direction, the most unplanted side pixels adjacent to the machine body traveling direction Comparing with, the most unplanted side pixel corresponding to the crop located on the boundary is in a state where it is not located on the planted side more than the most unplanted side pixel corresponding to the crop not located on the boundary. It is possible to distinguish the most unplanted side pixel corresponding to the crop located above and the most unplanted side pixel located on the already planted side of the boundary.

〔The invention's effect〕

Therefore, the machine body is tilted, and the pixels corresponding to the crops located on the boundary and the pixels corresponding to the crops on the crop row adjacent in the machine width direction are located on different pixel rows. Even now, it has become possible to accurately detect the information corresponding to the boundary.

〔Example〕

 Embodiments of the present invention will be described below with reference to the drawings.

As shown in FIG. 2 and FIG. 3, a seedling planting device (3) is provided on the rear side of a vehicle body (V) equipped with a pair of left and right front and rear wheels (1) and (2) so that it can be raised and lowered. A rice transplanter as a planting machine is configured.

Then, on the front side of the machine body (V), a seedling (A) as a crop planted by the seedling planting device (3).
Boundary (L) between the already planted part (B) and the unplanted part (C)
An image sensor (S) as an image pickup means for picking up an image of a portion corresponding to the above in a two-dimensional direction is provided so that the field can be seen from the vertical direction.

The imaging field of view of the image sensor (S) will be described. As shown in FIG. 3, when the body (V) is properly aligned with the boundary (L), the boundary (L) is detected. Is positioned at the center of the image sensor (S) in the up-down direction of the field of view, and in its proper state, at least two rows of seedlings (A) are imaged in the machine width direction and in the machine travel direction. Multiple seedlings (A)
Is set so as to be imaged.

In other words, the seedlings (A), the first predetermined gap with respect to the machine body traveling direction (l ₁₎ for each (set are to 20 cm), wherein the relative vehicle body width direction first predetermined gap (l ₁ ) Larger than the second set interval (l ₂ ) (set to 30 cm), each will be planted in a rectangular grid, the fuselage (V)
The field of view of the image sensor (S) is set so as not to lose sight of the seedling (A) located on the boundary (L) even if the seedling is tilted with respect to the boundary or deviated in the lateral direction. Has been done.

Therefore, the visual field center of the image sensor (S) is located at a position where the front wheel (1) does not enter the imaging visual field, and is located at the second set interval from the outermost planting position of the planting device (3). It is a position that protrudes in the width direction of the machine by a distance corresponding to (l ₂ ).

Then, the machine body (V) follows the boundary (L) based on the detection information of the boundary detection unit that detects the boundary (L) based on the image pickup information of the image sensor (S), as described later. The steering is controlled so that the vehicle automatically travels.

Explaining the control configuration, as shown in FIG.
A seedling extraction circuit (4) for extracting a pixel having information corresponding to the seedling (A) among the pixels arranged in the machine body traveling direction and the machine body lateral direction from the image pickup information of the image sensor (S), and An image memory (5) for storing the output image information, and the boundary (L) is detected based on the image information stored in the image memory (5), and for steering based on the detected information. A control device (7) utilizing a microcomputer for driving the actuator (6) is provided.

Explaining the seedling extraction circuit (4), a subtractor (8) for subtracting a blue signal (B) from a green signal (G) of the three primary color signals of the image information output from the image sensor (S). , And a comparator (9) that binarizes the output signal of the subtractor (8) based on a set threshold value (Vth).

That is, in this embodiment, the seedling (A) is green,
The field scene where it is planted is configured to be image-processed while paying attention to the fact that it is a water surface or a mud surface,
The image sensor (S) is of a color type capable of outputting three primary color information, and only the image information corresponding to the seedling (A) is extracted from the image information.

In addition, a green component (G) and a blue component (B) of the three primary color information output from the image sensor (S) are added.
Considering the strength of the seedling (A) with respect to the part where the seedling (A) is present, the part corresponding to the mud surface, and the water surface part where natural light is reflected, in the part where the seedling (A) is present, the green component (G) The intensity is high and the intensity of the blue component (B) is low. Also, in the part corresponding to the mud surface, the green component (G)
Both the intensity of the blue component (B) and the intensity of the blue component (B) are small. Furthermore, on the water surface, the green component (G) and the blue component (B)
Both of the strengths are large.

Therefore, the blue signal (B) is subtracted from the green signal (G) to remove the influence of natural light reflected on the water surface and the image information corresponding to the mud surface. Then, by extracting and binarizing a portion in which the level of the image signal obtained by subtracting the blue color signal (B) from the green color signal (G) is equal to or higher than a set threshold value (Vth), the seedling (A) is extracted. Pixels corresponding to only are extracted.

Therefore, the seedling extraction circuit (4) first pixel extracting means (10) for extracting pixels having information corresponding to the seedling (A) among the pixels lined up in the machine advancing direction and the machine lateral width direction.
0).

Next, the boundary detection processing will be described in detail based on the flowchart shown in FIG.

That is, the boundary detection process is performed every set time, or
Each time the machine body (V) travels the set distance, the imaging process is performed every time the process is started, and the pixels corresponding to only the seedling (A) are extracted accordingly. The image information is input to the pixel memory (5).

As the image information is input, the pixel located on the most unplanted side in each pixel row along the machine lateral direction (X-axis direction) aligned in the machine traveling direction, that is, the y-axis direction in the vertical direction on the image, That is, the most unplanted side pixel is extracted, and in each of the extracted most unplanted side pixels, the pixel that is not located on the already planted side of the vertically adjacent most unplanted side pixel is extracted (fifth). Figure (a),
(See (b)).

If you add a description, for each pixel column along the machine width direction,
Scanning is performed from the planted side to the unplanted side along the x-axis direction, and the pixel having the maximum coordinate in the x-axis direction is extracted as the most unplanted side pixel located on the most unplanted side. .

Next, the coordinates in the x-axis direction of the most unplanted side pixel located on the most unplanted side on each pixel row along the machine width direction aligned along the y-axis direction, and the upper and lower sides of the most unplanted side pixel. When the position of the most unplanted side pixel is not on the already planted side of the vertically adjacent most unplanted side pixel, the position of the most unplanted side pixel adjacent to is compared with the coordinate in the x-axis direction. The unplanted side pixel is extracted as a pixel located on the boundary (see the chain double-dashed line in FIG. 5A).

Therefore, the process up to extracting the most unplanted side pixel located on the boundary is that the position in the machine lateral width direction in each of the pixel rows along the plurality of machine lateral width directions aligned in the machine advancing direction is the most unplanted side. Second pixel extracting means for extracting the most unplanted side pixels, and for extracting those pixels that are not located on the already planted side of the most unplanted side pixels that are adjacent to each other on both sides of the most unplanted side pixels in the machine body traveling direction. It corresponds to (101).

Then, using Hough transform or the like, a straight line connecting the pixels extracted by the second pixel extraction means (101) is obtained as a straight line corresponding to the boundary (L), and the visual field of the image sensor (S) is obtained. An inclination (θ) with respect to the y-axis passing through the center and a distance (χ) in the x-axis direction with respect to the center of the screen are obtained.

That is, when the aircraft (V) is in a proper state with respect to the boundary (L), the imaging field of view is set so that the boundary (L) is located at the center of the screen. Therefore, the amount of displacement of the vehicle body (V) in the lateral width direction with respect to the boundary (L) corresponds to the distance (χ) from the center of the screen in the direction orthogonal to the boundary (L).

Therefore, the direction (θ) of the obtained boundary (L), and
By controlling the steering so that the deviation amounts (χ) are both zero, the airframe (V) can be automatically driven along the boundary (L).

[Another embodiment]

In the above embodiment, in order to obtain the straight line corresponding to the boundary (L), instead of using the Hough transform, for example, from the coordinate values of the two pixels located at the upper and lower ends in the y-axis direction,
Inclination with respect to y-axis, that is, direction (θ) of the boundary (L)
And the distance (χ) in the x-axis direction with respect to the center of the screen may be obtained.

Further, in the above embodiment, the case where the present invention is applied to the rice planting machine is illustrated, but the specific configuration of each unit such as means for extracting pixels corresponding to the crop from the image information is the type of the target crop or Various changes can be made according to the configuration of the crop planting machine.

It should be noted that reference numerals are added to the claims for convenience of comparison with the drawings, but the present invention is not limited to the structures of the accompanying drawings by the entry.

[Brief description of drawings]

The drawings show an embodiment of a boundary detection device for a crop planting machine according to the present invention, FIG. 1 is a block diagram of a control configuration, FIG. 2 is a schematic plan view of the crop planting machine, and FIG. 3 is a side view of the same. FIG. 4 is a flow chart of boundary detection, FIG. 5 (a),
(B) is an explanatory diagram of image processing. (L ₁ ) …… planting interval, (l ₂ ) …… planting width,
(B) …… Already planted part, (C) …… Unplanted part, (L)
...... Boundary, (S) ...... Imaging means, (100) ...... First pixel extraction means, (101) ...... Second pixel extraction means.

Claims (1)

(57) [Claims] 1. An already-planted portion (B) planted in such a manner that a crop is spaced at a set planting interval (l ₁ ) along the machine body traveling direction and at a set planting width (l ₂ ) along the machine body lateral width direction. ) And an unplanted portion (C) at a location (L) that is imaged from a vertical direction or a direction close thereto and imaging information (S), and the imaging information is arranged in the machine advancing direction and the machine lateral direction. A first pixel extracting means (100) for extracting pixels having information corresponding to the crop among the pixels;
From the plurality of pixels extracted by the first pixel extraction means (100), the position in the machine lateral width direction in each of the pixel rows along the machine lateral direction arranged in the machine traveling direction is the most unplanted side. Second pixel extracting means for extracting the unplanted side pixels, and extracting the pixels which are not located on the already planted side with respect to the most unplanted side pixels adjacent to both the most unplanted side pixels adjacent to each other in the machine body traveling direction ( 101) A boundary detection device for a crop planting machine, comprising: