JPH0624016B2 - Imaging type boundary detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to an imaging means for imaging a portion corresponding to a boundary between an unprocessed work site and a processed work site in a two-dimensional direction, and imaging information thereof. Differentiating means for obtaining a differential value of each pixel arranged in the two-dimensional direction based on brightness, binarizing means for extracting a pixel having a differential value larger than a set threshold, passing through the extracted pixels, and a plurality of Imaging provided with a straight line calculating means for obtaining a plurality of straight lines having gradients set in stages for each extracted pixel and a straight line extracting means for extracting a straight line having the maximum frequency of the obtained plurality of types of straight lines Expression boundary detection device.

[Prior Art] Since the brightness of the unprocessed work site and the processed work site looks different in this type of imaging-type boundary detection device, a large change in brightness occurs at the boundary portion. This is utilized to obtain information on the straight line corresponding to the boundary.

The obtained straight line information will be used as control information for automatically traveling the work vehicle along the boundary between the unprocessed work site and the processed work site. As the work vehicle travels, the work is repeated.

However, if all the inclination structures of a plurality of straight lines passing through each of the extracted pixels are obtained, the amount of calculation increases, and it becomes difficult to perform image processing at a speed corresponding to the traveling speed of the work vehicle. Conventionally, since the work vehicle is steered so as to travel along the boundary, the straight line corresponding to the boundary in the image pickup screen of the image pickup means has a greatly different inclination with respect to the length direction of the boundary. Take advantage of the fact that
The range of inclination of the straight line to be obtained is limited within the set range in the length direction of the boundary to reduce the amount of calculation for obtaining a plurality of straight lines.

For example, as shown in FIG. 7 (a), when the work vehicle is in a proper state with respect to the boundary (L), the boundary (L)
A straight line corresponding to the image pickup screen along the traveling direction of the work vehicle
The imaging field of view is set so as to be positioned on the Y-axis passing through the center of (A), and the set tilt range (() corresponding to the length direction of the boundary (L) in the proper state is set. ±
Only a plurality of straight lines within θ ₁ ) are obtained.

However, in FIG. 7 (a), a part of the boundary (L) is the imaging screen.
Although it is illustrated as being located in (A), it does not correspond exactly to the positional relationship between the actual visual field range imaged by the imaging means and the boundary (L), but on the illustrated boundary (L). So that the straight line located at corresponds to the straight line with the maximum frequency,
It is shown conceptually.

[Problems to be solved by the invention]

However, due to a control delay or the like, a delay occurs until the work vehicle converges to a state in which it is in a proper state with respect to the detected boundary.

Therefore, if the range of inclination of the straight line to be obtained is limited with reference to the case where the vehicle body is in the proper state with respect to the boundary as in the above-mentioned conventional case, the obtained straight line of maximum frequency and the straight line corresponding to the boundary do not match There are cases.

For example, as shown in FIG. 7 (b), when the boundary (L) is curved, the work vehicle follows the boundary (L) late, and the straight line corresponding to the boundary (L) The inclination, that is, the length direction of the boundary (L) is outside the set inclination range (± θ ₁ ) with respect to the reference Y-axis direction, and the boundary may be lost.

In FIG. 7 (B), the positional relationship between the imaging screen (A) and the boundary (L) is conceptual, as in the case of FIG. 7B.
The actual positional relationship between the imaging field of view and the boundary (L) is not an accurate illustration.

The present invention has been made in view of the above circumstances, and an object thereof is to reduce the number of straight lines to be obtained,
This is to prevent the boundary from being lost due to the control delay of the steering control.

[Means for solving problems]

The characteristic configuration of the image pickup type boundary detection device according to the present invention is such that the inclination range of a plurality of straight lines obtained by the straight line calculating means in the present step is the maximum extracted by the straight line extracting means in the previous step. A characteristic configuration is that inclination limiting means is provided for limiting the inclination of the straight line which becomes the frequency so as to be within the set inclination range.

The operation and effect of this characteristic configuration are as follows.

[Work]

That is, the process for detecting the straight line corresponding to the boundary is
It will be repeated as the work vehicle travels, and since there is a control delay in the steering control, the inclination of the straight line corresponding to the boundary obtained from the imaging information in the step before the current step, The inclination of the straight line corresponding to the boundary obtained from the imaging information of this time does not significantly differ from the state.

Therefore, for example, as shown in FIG. 7C, the range of the inclinations of the plurality of straight lines obtained from the imaging information of this time is the straight line (Lm that has the maximum frequency obtained in the step before the current step).
ax) (the straight line indicated by the thin broken line in the figure) is restricted so that it is within the set tilt range (± θ ₁ ), the straight line corresponding to the boundary (L) in the current straight line detection (in the figure) , A straight line indicated by a thick broken line is outside the set inclination range, and it is possible to avoid the possibility of losing the boundary.

〔The invention's effect〕

As a result, it has become possible to avoid losing the boundary while reducing the amount of calculation for obtaining the straight line corresponding to the boundary.

〔Example〕

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

As shown in FIGS. 3 and 4, the boundary between the uncut land (B) as an unprocessed work site and the already cut land (C) as a processed work site.
(L) in the working vehicle (V) that automatically travels along the (V), the image sensor (S ₁ ) as an imaging means for imaging the work site on the traveling front side in the two-dimensional direction, the boundary on the traveling front side ( It is provided so that a portion corresponding to L) is imaged obliquely from above.

In FIG. 3, (1) is a front wheel, (2) is a rear wheel, (3) is a lawn mowing device as a working device, and (H) is a steering handle for boarding operation.

If the image pickup field of view of the image sensor (S ₁ ) is further explained, the work vehicle (V) follows the proper condition with respect to the boundary (L) between the uncut land (B) and the already cut land (C). In this state, the boundary (L) is located at the center of the visual field of the image sensor (S ₁ ) in the lateral width direction.

As shown in FIG. 2, each operation for obtaining a straight line corresponding to the boundary (L) from the imaged information of the image sensor (S ₁ ) and the work based on the obtained straight line information So that the car (V) automatically runs along the boundary (L),
A control device (5) using a microcomputer is provided which constitutes each control means for controlling the steering actuator (4).

Incidentally, in the figure, (S ₂ ) is a distance sensor for detecting the travel distance of the work vehicle (V), and information such as a turn start position for ending one work stroke and moving to the next work stroke is shown. It is provided to detect.

That is, the work vehicle (V), the portion from one side of the square uncut land (B) surrounded by the already cut land (C) to the opposite side, as one work stroke, the work vehicle ( V) of the image sensor (S ₁ ) so that it automatically travels along the boundary (L) between the uncut land (B) and the already cut land (C) on the side along the length of the working stroke. Based on the information of the straight line corresponding to the boundary (L) detected from the imaging information, the steering is to be controlled, and
Repeating automatically turning toward the starting end of the next work stroke in the direction intersecting the work stroke as it reaches the end side of one work stroke, that is, the opposite side of the uncut land (B). Thus, the lawn mowing work within a predetermined range is automatically performed.

Next, the image processing for detecting the straight line corresponding to the boundary (L) will be described based on the flowchart shown in FIG.

That is, the imaging process is performed each time the work vehicle (V) travels a set distance, and a preset pixel density (32 × 32 pixels is set based on the imaging information of the image sensor (S ₁ ). The brightness level of each pixel is quantized.

Next, as shown in FIG. 5, the pixel (e) to be processed
Based on the values of 8 pixels (a to d, f to i) adjacent to the periphery of the image, the image is a direction intersecting the pixel (e) to be processed in the length direction of the boundary (L). The process of obtaining the absolute value of the differential value of the brightness in the X-axis direction as the differential value (IP) of each pixel is performed for each pixel arranged in the two-dimensional direction using the following equation (i).

IP (x, y) ＝ ｜ (c + 2f + i) − (a + 2d + g) ｜ …… (i) However, the pixel located at the outermost part of the screen is differentiated because there is no pixel information adjacent to it. Since the processing cannot be performed, it is excluded from the processing target.

Therefore, the pixel range to be actually processed is the quantized pixel range (1 to 32) in the X-axis direction and the Y-axis direction by subtracting one pixel located at both ends (2 to 31). ).

That is, the process of obtaining the differential value (IP) for each pixel corresponds to the differentiating means (100).

Next, the coordinate value (x, y) of the pixel is set to the minimum value (x = 2, y =
Initially set to 2), it is determined whether the differential value (IP) of the pixel is greater than or equal to a preset threshold value, and thus the pixel whose differential value (IP) is larger than the set value. Through the extracted pixels using the Hough transform,
In addition, a plurality of straight lines having inclinations set in a plurality of stages will be obtained.

That is, the process of extracting the pixels whose differential value (IP) is equal to or larger than the threshold value corresponds to the binarizing means (101) which extracts the pixels whose calculated differential value is larger than the set value.

However, in the Hough transform, a plurality of straight lines passing through the extracted pixels are set to polar coordinates set in advance in a plurality of stages in the range of 0 to 180 degrees with respect to the X axis based on the following formula (ii). It is obtained as a combination of the inclination (θ) with respect to the X axis as the reference line in the system and the distance (ρ) from the origin, that is, the center of the screen (see FIG. 6).

ρ = x · cos θ + y · sin θ (ii) That is, the process of obtaining the distance (ρ) from the origin for each inclination (θ) set in a plurality of stages in the equation (ii) is the binary value. This corresponds to the straight line calculation means (102) for obtaining, for each extracted pixel, a plurality of straight lines that pass through the pixels extracted by the conversion means (101) and have different inclinations set in a plurality of stages.

However, the range of each inclination (θ) obtained by the Hough transform is
As shown in FIG. 6 and FIG. 7 (C), the straight line extraction means (described later) in the previous boundary detection processing as a step prior to the processing at the current step as the boundary detection processing based on the current imaging information ( Based on the slope (θmax) of the straight line (Lmax) that is the maximum frequency extracted in 103),
Only the straight line within the preset tilt range (± θ ₁ ) (set to ± 30 degrees) is calculated, and the tilt range to be calculated is limited.

That is, the range of the slope (θ) of the straight line obtained by the Hough transform is set based on the slope (θmax) of the maximum frequency extracted in the previous boundary detection processing as a set slope range (± θ ₁ ). The process of limiting to (1) corresponds to the tilt limiting unit (104) that limits the tilt range of a plurality of straight lines obtained by the straight line computing unit (102).

The set inclination range (± θ ₁ ) is shown as an angle in a direction orthogonal to the straight line (Lmax) having the maximum frequency in FIG. 6, and the maximum frequency obtained in FIG. The length direction of a straight line (Lmax) (a straight line indicated by a thin broken line in the figure) is used as a reference, but the inclination of the straight line in FIG. 6 is defined as the angle around the origin in the polar coordinate system. The straight line in FIG. 7 (c) is shown as a straight line in the Cartesian coordinate system in a state where the imaging screen (A) and the boundary (L) on the ground are associated with each other.
This is because the reference position of θ ₁ ) becomes different. However, the size itself of the set inclination range (± θ ₁ ) has the same value in both the polar coordinate system and the Cartesian coordinate system, only the reference position is different.

Further, in FIG. 6, the straight line indicated by the solid line is a conceptual illustration of the maximum frequency line obtained in the current boundary detection. In FIG. 7 (c), the previously obtained maximum frequency straight line (Lmax) is a thin broken line, and the maximum frequency straight line obtained this time is a thick broken line, respectively, on the current imaging surface (A) and the ground surface. It is illustrated in a state corresponding to the positional relationship with the boundary (L).

However, although the imaging screen (A) is illustrated as a rectangle,
As described above, since the visual field direction of the image sensor (S ₁ ) is provided so as to obliquely look down on the work site, as shown in FIGS. 3 and 4, the image sensor (S ₁₎ The actual image pickup range on the ground imaged by) is a trapezoidal shape, and the image pickup screen shown in FIG.
The shape of (A) does not exactly correspond to the visual field range actually imaged, but conceptually illustrates the relationship between the straight line obtained on the image and the boundary (L) on the ground. .

Next, the obtained combination of the distance (ρ) and the inclination (θ) is added to the two-dimensional histogram.

Then, the value of the inclination (θ) set in the plurality of steps is
The maximum frequency straight line (Lma obtained in the previous boundary detection process
After the process of adding the two-dimensional histogram is repeated until the angle reaches the set tilt range (+ θ ₁ ) with respect to the tilt of x), the coordinate value (x, y) of the pixel is the maximum value (x = (31, y = 31) until the coordinate value is sequentially updated, pixels with the differential value (IP) equal to or greater than the set value are extracted, and each extracted pixel is converted into a straight line passing through the pixel. The process of adding the histogram of the combination of the corresponding slope (θ) and the distance (ρ) from the origin is repeated to count the frequencies of a plurality of types of straight lines passing through all the extracted pixels.

After the frequency of straight lines for all the extracted pixels is completed, a combination of the inclination (θ) and the distance (ρ) from the origin, which is the maximum frequency, is obtained from the value added to the two-dimensional histogram. Thus, one straight line having the maximum frequency is determined, and the straight line is obtained as a straight line corresponding to the boundary (L).

That is, the process of obtaining the combination of the inclination (θ) having the maximum frequency and the distance (ρ) from the origin corresponds to the straight line extracting means (103) for extracting the straight line having the maximum frequency.

The deviation of the work vehicle (V) with respect to the boundary (L) is calculated by calculating the distance (ρ) and the inclination (θ) indicating a straight line with the obtained maximum frequency.
The actual position of the work vehicle (V) with respect to the boundary (L) is calculated by mapping from the value of to the XY coordinate system corresponding to the actual ground.

[Another embodiment]

In the above-described embodiment, the case where a plurality of straight lines passing through each extracted pixel is calculated and obtained at any time is exemplified, but for example, for each pixel, the straight line passing through the pixel is 0 degrees to 18 degrees.
The value of the distance (ρ) corresponding to each inclination (θ) set in advance in a range of 0 degree is calculated, and these values are stored in a table and stored for each extracted pixel. Of the coordinate value and the maximum frequency line (Lmax) (θma
x) and the inclination range of the value of the read distance (ρ) may be limited, and the linear calculation means (102).
In addition, the specific configuration of the tilt limiting means (104) for limiting the tilt range of the plurality of straight lines obtained by the straight line calculating means (102) can be variously changed.

Further, in the above embodiment, the present invention is a boundary (L) between the unprocessed work site (B) and the processed work site (C) in the lawn mowing work vehicle (V).
Although the case where it is applied to the means for detecting the above is illustrated, the specific form of the boundary and the work place to be detected, and the specific configuration of each means for detecting the boundary (L) can be variously changed.

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 an image pickup type boundary detection apparatus according to the present invention. FIG. 1 is a flow chart of boundary detection, FIG. 2 is a block diagram showing a control configuration, FIG. 3 is a side view of an image pickup visual field, and FIG. The figure is an explanatory view of the work place, FIG. 5 is an explanatory view of the differential processing, FIG. 6 is an explanatory view of the Hough transform, and FIGS. 7 (a), (b), and (c) are explanatory views of the captured image. . (B) …… Unprocessed work site, (C) …… Processed work site, (L) ……
Boundary, (S ₁ ) ... Imaging means, (100) ... Differentiation means,
(101) ... binarization means, (102) ... straight line calculation means, (103) ... straight line extraction means, (104) ... tilt limiting means.

Claims (1)

[Claims] 1. An image pickup means (S ₁ ) for picking up an image of a portion corresponding to a boundary (L) between an unprocessed work site (B) and a processed work site (C) in a two-dimensional direction, of the image pickup information. Differentiating means (10) for obtaining the differential value of each pixel arranged in the two-dimensional direction based on the brightness.
0), a binarizing means (101) for extracting pixels having the differential value larger than a set threshold value, and a plurality of straight lines passing through the extracted pixels and having inclinations set in a plurality of stages,
An image capturing type boundary detection device comprising: a straight line calculating means (102) for each extracted pixel; and a straight line extracting means (103) for extracting a straight line having the maximum frequency among a plurality of kinds of straight lines thus obtained. Then, the slope range of the plurality of straight lines obtained by the straight line calculating means (102) at the current step is calculated by the straight line extracting means (10) at the step before this.
Inclination limiting means (10) for limiting the inclination of the straight line having the maximum frequency extracted in 3) within the set inclination range.
4) An image pickup type boundary detection device provided.