WO1996030294A1 - Load swing detecting device for a crane - Google Patents

Abstract

A load swing at a crane is detected with high accuracy by accurately detecting the position of a suspended load through provisions of an imaging means (2) disposed in the vicinity of the top portion of a boom (1) and attitude controlled such that its visual field is directed perpendicularly downwardly, and a hook position detecting means (16) for detecting a relative position of a hook to the boom top portion by obtaining the centroid of an identification pattern by image processing an identification pattern IP plotted on an upper surface of a hook (4) disposed at a leading end of a rope and the identification pattern IP imaged by the imaging means (2).

Description

 Description Crane runout detector Technical field

 BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a load fluctuation detection device for a crane that detects a load fluctuation of a crane by imaging an identification pattern provided on a hook with an imaging camera.

 As a prior art related to the detection of load swing of a crane, there is, for example, Japanese Utility Model Laid-Open Publication No. 50-1506967:

 In this prior art, a case is provided at the end of the boom, which is rocked by a swing of a rope, and a pendulum body which is always held in a vertical posture by its own weight is provided in this case. By detecting the relative displacement, the deflection of the load at the wire tip is detected.

 However, in this conventional technology, the run-out of the rope supporting the suspended load is detected near the boom tip, which is the base of the run-out, so it is assumed that the rope ahead is a straight line. There is a problem that the rope is in a curved state as shown in FIG. 10 when the load swings, so that the position of the suspended load cannot be accurately detected in the above-described conventional technology. DISCLOSURE OF THE INVENTION An object of the present invention, which has been made in view of the above circumstances, is to provide a crane load deflection detection device capable of accurately detecting the position of a suspended load and detecting load deflection of the crane with high accuracy.

According to the present invention, there is provided an image pickup means disposed near the top of the boom and controlled in attitude so that the field of view is directed vertically downward, an identification pattern drawn on an upper surface of a hook disposed at the tip of the rope, The relative position of the hook with respect to the top of the boom is determined by performing image processing on the identification pattern imaged by the imaging means and determining the centroid position of the identification pattern. And a hook position detecting means for detecting

 According to the present invention, the position of the hook for hanging the suspended load is visually detected using the output of the imaging means. That is, the relative position of the hook with respect to the top of the boom is detected by performing image processing such as template pattern matching on the identification pattern imaged by the imaging means and obtaining the centroid position of the identification pattern.

 As described above, according to the present invention, a predetermined identification pattern is drawn on the upper surface of the hook of the crane, and image processing such as template pattern matching is performed on the image data of the identification pattern imaged by the imaging means, and the identification pattern is added. The relative position of the hook with respect to the boom top is detected by determining the center of gravity of the boom, so the hook position can be detected accurately and at high speed, and the load swing of the crane can be detected with high accuracy Become like BRIEF DESCRIPTION OF THE FIGURES

 FIG. 1 is a flowchart showing an embodiment of the present invention.

 Figure 2 is an external view of a crane to which the present invention is applied:

 Figure 3 shows the hook with the identification pattern drawn

 Figures 4 (a) and 4 (b) show the configuration of the attitude holding device of the video camera:

 Figure 5 is a block diagram showing the configuration of the control system for holding the posture.

 Figures 6 (a) and 6 (b) show examples of identification patterns

 Figure 7 is a block diagram showing the configuration of the image processing system.

 Figures 8 (a) and (b) show the search area and template:

 Figure 9 is a diagram illustrating a template for detecting the hook rotation angle.

 FIG. 10 is a diagram for explaining a conventional problem.

 Hereinafter, the present invention will be described in detail with reference to embodiments shown in the accompanying drawings.

 Fig. 2 shows the appearance of a crane to which the present invention is applied.

In FIG. 2, a video camera 2 is provided at the top of boom 1, The field of view of the camera 2 is always directed vertically downward by a posture holding device described later.

 On the upper surface of the hook block 4 suspended by the wire rope 3, for example, as shown in FIG. 3, an identification pattern IP (in this case, four black circles are arranged at equal intervals on the same circumference). Is drawn.

 That is, in this device, the identification pattern IP on the upper surface of the hook is imaged by the video camera 2, and the load deflection is detected based on the imaged data. Is the suspended load, 7 is the control box, and 8 is the hook position display device located in the driver's seat =

 FIGS. 4 (a) and 4 (b) show the configuration of a posture holding device provided near the top of the boom where the video camera 2 is disposed, and FIG. 4 (b) shows an arrow X in FIG. 4 (a). Here is a view from the direction:

 That is, the electric motor 9 is fixed to the top of the boom 1, and the turntable 10 is fixed to the output shaft 9a. A video camera 2 and an inclinometer 11 are fixed to the front surface of the turntable 10, and a circuit for driving the electric motor 9 (electric motor drive device 12, arithmetic device 13) is provided on the back surface. Fixed: Fig. 5 shows the configuration of the drive system of the electric motor 9, and the inclinometer 11 detects the angle 0 (see Fig. 4) between the optical axis direction of the video camera 2 and the vertical direction. The detected value 0 is input to the computing device 13. A computing value 13 (0 in this case) is set in the computing device 13, and the computing device 13 calculates a deviation between the command value 00 and the detection angle 0. Then, the deviation is input to the electric motor driving device 12. The electric motor driving device 12 drives the electric motor 9 in accordance with the command current i corresponding to the deviation. It is controlled so that it always faces vertically downward

 Figures 6 (a) and 6 (b) show the identification pattern IP drawn on the top of the hook. The pattern shown in (a) is an example that can detect not only the hook position but also the rotation of the hook itself. The concentric pattern shown in () is an example that can detect only the hook position

FIG. 7 shows the configuration of a data processing system for image data of the video camera 2. The image data of the video camera 2 is converted into digital data by the AZD converter 14. It is.

 In this case, as shown in Fig. 8 (a), the digital data after the AZD conversion divides the data for one screen of the video camera 2 into 512 * 512 pixels, and the grayscale data of each pixel is 8-bit. (That is, 256 gradations). Accordingly, the information amount of one screen is 8 * 512 * 512bit = 20972bit = 256kbite, and these data are stored in the image data storage device 15.

 The image processing device 16 sets a 32 × 32 bit search area for the stored image data as shown in FIG. 8 (b), and a 16 × 16 bit template in the search area. Perform the pattern matching process using:

 Hereinafter, the processing of the image processing device 16 will be described with reference to the flowchart of FIG.

 The image processing device 16 stores the identification pattern figure drawn on the upper surface of the hook as a template composed of 16 × 16-bit pixels, and the image processing device 16 reads the stored original template. (Step 100):

 Next, the image processing device 16 executes the process of enlarging or reducing the template figure according to the distance L between the video camera 2 and the hook 4: that is, the viewing angle of the video camera 2 mounted on the top of the boom is fixed. In this case, the size of the identification pattern figure in (512 * 512) in the image data of the video camera 2 changes according to the length of the distance. Therefore, in order to adjust this, the template figure is moved according to the distance. First, the image processing device 16 calculates the distance L based on the output of the rope length sensor 5, and further calculates a scaling ratio S according to the following equation (step 110).

 S = LO / L

 L0: Length of the work when S = 1

 Next, the template image is multiplied by S using the calculated enlargement / reduction ratio S (step 120).

Next, the image processing device 16 sets a search area consisting of the above 32 × 32 bits with the hook position detected last time as the center position (step 130). In the case of the first search, the rope cannot be unwound. Appropriately corresponding to the vertically suspended state Make it set to the initial searcher of:

 Then, at each position in the set searcher, the correlation with the above 16 × 16 template is sequentially examined (step 140): That is, the number of possible positions of the 16 × 16 template in the search area is 16 × 16 = 256 patterns, but the image processor 16 examines the correlation between the searcher image data and the template at all 256 template positions: As an index of the correlation height,

-1616 ₉

 Z = ∑∑ (gs (i + k, j + 1)-gt (i, j))

 j = i do 1

Is used:

 gs (i, j): gray level of pixel at row i and column j of searcher

 gt (i.j): The gradation of the template corresponding to the pixel in the "i-th row" of the searcher

 k, 1: Variables that the template can take in the search area

 0≤k, 1≤15

 According to the above correlation equation, the value of Z decreases when the correlation is high, so the position where Z is the smallest among the 256 template positions is the position with the highest possibility as the position where the identification pattern exists. Become:

 Therefore, the image processing device 16 obtains the minimum correlation data Zmin among the 256 correlation data Z (step 150), and compares this Zmin value with a predetermined bridge value Zref (step 160). If Zmin is smaller than Zref, the template position corresponding to Zmin is determined as a hook position (step 170). The determined hook position is output to the crane control device 18. Also, this hook position is displayed as a hook position display. The position is output to section 8 and the position is displayed as appropriate.If Zmin≥Zref in the previous comparison, it is determined that no hook exists in this search area, and the search area is set at another position. Set (step 180); and execute the same processing as above in the reset search area

In the above embodiment, the search area is the template recognized in the previous process. Is set as the center point, and at the normal swing speed of the suspended load, the image on the top of the hook is set so that it does not come out of the searcher during the pattern matching process. The image on the top of the hook does not go outside the searcher. That is, it is guaranteed that the target hook position always exists in the search area, so that the template position where Z is minimized can always be determined as the hook position.

 In the embodiment described above, for example, only the center of gravity of the hook is determined using the identification pattern as shown in FIG. 6 (b), but the hook is determined using the identification pattern as shown in FIG. 6 (a). In order to obtain the rotation angle of each template, as shown in Fig. 9, different templates are prepared for each of multiple rotation angles, and pattern matching processing is performed on all these templates. It is sufficient to determine that the type and position of the template that gives the smallest correlation data Z among them are the rotation angle and position of the hook to be obtained.

 It should be noted that the graphic contents of the identification patterns used in the embodiments are merely examples, and other arbitrary identification patterns may be used.

 INDUSTRIAL APPLICABILITY The present invention is used for crane load deflection detection.

Claims

The scope of the claims 1. An imaging means disposed near the top of the boom and controlled in attitude so that the field of view points vertically downward;  An identification pattern drawn on the upper surface of a hook disposed at the tip of the rope, and an image processing of the identification pattern imaged by the imaging means to obtain a centroid position of the identification pattern, thereby obtaining a relative position of the hook with respect to the boom top. Hook position detecting means for detecting a position;  Run-out detection device of a crane equipped with  2. The swing detection of a crane according to claim 1, wherein the shape of the identification pattern does not change due to rotation of the identification pattern around the centroid. 3. The load deflection detecting device for a crane according to claim 2, wherein the identification pattern is a concentric circle figure.  4. The shape of the identification pattern changes due to rotation of the identification pattern about the centroid.  2. The apparatus according to claim 1, further comprising hook rotation angle detection means for detecting a rotation angle of the hook by detecting a rotation angle of the identification pattern.  5. The crane load deflection detection according to claim 1, wherein the hook position detecting means has a template image of the identification pattern, and obtains a centroid position of the identification pattern by pattern matching with the template. Equipment:  6. It further comprises a distance detecting means for detecting the distance from the top of the boom to the upper surface of the hook,  6. The apparatus according to claim 5, wherein the hook position detecting means performs pattern matching by enlarging or reducing the template image in accordance with a detection distance of the distance detecting means.