JPH06137828A - Detecting method for position of obstacle - Google Patents

Abstract

PURPOSE:To stably detect the position of an obstacle in three-dimensional space by photographing a light spot projected to the obstacle with two cameras arranged at a fixed interval, and processing the image of the light spot to determine its position, and continuously carrying out this process. CONSTITUTION:The first and the second camera A, B are arranged at a fixed interval, and their visual axes are made to intersect mutually at a spot being a reference distance ahead for measuring the width of a camera photograph domain on the crossing face of camera transverse axes. A laser beam is projected to a measured body 1 by a laser projector 2, and converged into a spot-shape. Next luminescence at the laser spot is caught by the cameras A, B, and inputted into an image processing device 3 and binary-processed. In this case, the visual axis of the camera and the central axis of an image plane are made to coincide mutually. Next the spot position on a binary image is measured, and the two-dimensional coordinates of the spot position on the measured body 1 are measured from the above spot position and the previously set up arrangement of the cameras A, B. Three dimensonal coordinates are computed by a personal computer 4 on a basis of two dimensional coordinates.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an obstacle position detecting method for non-contact detection of the position of an obstacle around a robot which changes with time, among the autonomous functions essential for construction robots.

[0002]

2. Description of the Related Art According to a conventional method, an ultrasonic wave, a light wave, or the like is projected onto an obstacle, and the distance to the obstacle is measured by the time until the reflection is detected.

[0003]

In the conventional method, the measured values are not stable due to the reflectance and surface shape of obstacles. Also, at the construction site, it is very difficult to make adjustments during measurement because the surrounding environment changes.

The present invention provides an obstacle position detecting method capable of stably detecting the position of an obstacle without being influenced by the change of the surrounding environment without depending on the reflectance or surface shape of the obstacle. Is intended.

[0005]

According to the present invention, a sufficiently converged light spot projected on an obstacle is photographed by the first and second cameras arranged at a predetermined distance, and the photographed image is taken. Is image-processed to find the position of one point on the obstacle, which is the position of the spot, and the position of the entire obstacle in the three-dimensional space is found by continuously performing this processing.

Laser light or penlight light is preferably used for the above-mentioned sufficiently converged light.

Further, it is preferable that the processing is such that the spot image is binarized to measure two-dimensional coordinates, and the two-dimensional coordinates are continuously processed to calculate three-dimensional coordinates.

[0008]

According to the method of the present invention, the projected sufficiently condensed spot of light is photographed, and the position of the measuring point on the obstacle is obtained by image processing, so that stable position detection is possible.

[0009]

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

FIG. 1 shows an apparatus embodying the present invention. This device comprises a laser light projector 2 for projecting and scanning a laser light, which is light sufficiently converged on an obstacle, that is, a DUT 1, and first and second cameras A, B composed of CCD cameras. And these first and second cameras A,
Image processing device 3 connected to B and the image processing device 3
And a personal computer 4 connected to.

In the measurement, as shown in FIG. 2, the first and second cameras A and B are arranged in advance at a predetermined separation distance W so that the camera visual axes a1 and a2 intersect in front of the reference distance L. To do. At this time, the width of the camera photographing range H on the camera spindle crossing surface (hereinafter referred to as the reference surface) S is measured.

In FIG. 3, the laser projector 2 is used to measure the object 1 to be measured.
The laser light is projected onto the DUT 1 so as to converge in a spot shape (step S1).

Therefore, the cameras A and B shoot the direction of the object 1. At this time, an optical filter matching the wavelength of the laser light is attached to each of the cameras A and B. Then, the case where the emission of the laser spot is captured by the cameras A and B can be measured. The image processing device 3 inputs the images from the cameras A and B (steps S2a and S2b), and outputs the images to
The value is processed (steps S3a and S3b). At this time, the camera viewing axes a1 and a2 and the screen center axis are made to coincide with each other.
Next, the position of the laser spot in the binarized image is measured, and the laser spot emission on the DUT 1 is determined based on the relationship between the position of the measured laser spot image and the preset arrangement of the cameras A and B. The two-dimensional coordinates of the position are measured (steps S4a and S4b). The personal computer 4 calculates three-dimensional coordinates based on the measured two-dimensional coordinates (step S5).

The mode of processing will be described below.

In FIG. 4, a laser beam projector 2 projects a laser beam toward the object to be measured 1 only within the effective viewing angle ψ of the first and second cameras A and B.

The cameras A and B input the image of the DUT 1 on which the laser beam is applied. The cameras A and B are equipped with filters matching the wavelength of the laser light in order to remove ambient light.

The image processing apparatus 3 converts the images from the cameras A and B into a binary image in which only the portion exposed to the laser light is converted into white and the other portions are converted into black. Then, the laser light projection position is obtained from the center of gravity of the figure of the white pixel (laser light projection surface) of the binarized image, and this is externally output as two-dimensional coordinates (X axis, Y axis) in units of the number of pixels.

The personal computer 4 has a relationship between the two-dimensional coordinates of the laser light projection positions of the cameras A and B and the preset positions of the cameras A and B such as the camera separation distance W and the camera axis crossing distance. From the above, the position of the spot based on the camera position is calculated as the three-dimensional coordinates in the actual size unit.

In the calculation mode of the measurement coordinates, first, the horizontal angles θ1 and θ2 in FIG. 5 are obtained.

In the camera A, when the number of pixels from the screen center point C to the screen edge END is N dots on the reference plane S ahead of the camera A (hereinafter also referred to as the point A) by the reference distance L, When the measured distance between the point C and the screen edge END is D and the measured point P is on the straight line A / P, the position of the measured point P obtained in the image is irrespective of the distance from the point A. It is the position of K1 dots from the center point C.

Therefore, the angle θ1 formed by the straight lines A and C and the straight lines A and P and the angle θ2 formed by the straight lines B and C and the straight lines B and P are Required by.

Next, referring to FIG. 6, the position coordinates P of the measuring point P are shown.
Calculate (Xp, Yp). First, the unit vector n1 of the vector AB is Ask from.

The vector AE is a unit vector n1 rotated clockwise by an angle α + θ1 and multiplied by L. The unit vector of the vector AE is n2 (X ', Y')
Then, Therefore, the coordinate E (Xe, Ye) of the point E is Therefore, the equation of straight line AE is Given in.

Similarly, in FIG. 7, the unit vector n1 of the vector BA is The vector BF is obtained by rotating the unit vector m1 counterclockwise by the angle β + θ2 and multiplying it by L.
Let m2 (Xg, Yg) be the unit vector of Therefore, the coordinates F (Xf, Yf) of the point F are Therefore, the equation of straight line BF is Given in.

Therefore, in FIG. 8, the position coordinate P (Xp, Yp) of the measuring point P is the intersection of the straight lines A and E and the straight lines B and F, and can be obtained by solving the following equation.

[0026] Next, in FIG. 9, the vertical angle φ of the measurement point P is obtained.
In the camera A, the screen center point C and the screen edge E when the number of pixels from the screen center point C to the screen edge END 'is N'dots on the reference plane S ahead of the point A by the reference distance L
When the measured distance between ND ′ is D ′ and the measured point P is on the straight line A / P, the measured point P position obtained in the image is
It is located at the position of K3 dots from the center point C regardless of the distance from the point A. Therefore, the angle φ formed by the straight lines A and C and the straight lines A and P is Required by.

In FIG. 10, the vertical coordinate P (P
z) is calculated. The horizontal distance L'from point A to point P is Given by Required by. Then, the above processing is continuously performed to obtain the entire position of the DUT 1.

[0028]

As described above, according to the present invention, by determining the position of the entire obstacle from the spot of the photographed light, the change of the surrounding environment can be achieved without depending on the reflectance or the surface shape of the obstacle. It is possible to perform stable position detection without being affected by.

[Brief description of drawings]

FIG. 1 is a perspective view showing an example for carrying out the present invention.

FIG. 2 is a plan view illustrating a camera arrangement.

FIG. 3 is a control flowchart.

FIG. 4 is a plan view illustrating laser projection.

FIG. 5 is a plan view illustrating calculation of measurement points in the horizontal direction.

FIG. 6 is a plan view illustrating calculation of horizontal coordinates of a measurement point from one camera.

FIG. 7 is a plan view illustrating calculation of horizontal coordinates of a measurement point from the other camera.

FIG. 8 is a plan view for explaining coordinate calculation of measurement points.

FIG. 9 is a side view for explaining calculation of a vertical angle of a measurement point.

FIG. 10 is a side view for explaining calculation of vertical coordinates of measurement points.

[Explanation of symbols]

 A ... First camera a1, a2 ... Camera viewing axis B ... Second camera C ... Screen center point D, D '... Measured distance END, END' ... Screen edge H. .. Camera shooting range L ... Reference distance P, P '... Measuring point S ... Reference plane W ... Separation distance ψ ... Effective viewing angle θ1, θ2 ... Horizontal angle φ ..Vertical angle 1 ... Object to be measured 2 ... Laser light projector 3 ... Image processing device 4 ... Personal computer

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI Technical display location H04N 7/18 C

Claims (1)

[Claims] 1. A sufficiently converged light spot projected on an obstacle is photographed by first and second cameras arranged at a predetermined separation distance, and the photographed image is image-processed at the spot position. An obstacle position detecting method, wherein the position of one point on an obstacle is obtained, and the position of the entire obstacle in a three-dimensional space is obtained by continuously performing this processing.