CN210676305U - Automatic welding smoke trapping device based on binocular vision - Google Patents
Automatic welding smoke trapping device based on binocular vision Download PDFInfo
- Publication number
- CN210676305U CN210676305U CN201921621398.5U CN201921621398U CN210676305U CN 210676305 U CN210676305 U CN 210676305U CN 201921621398 U CN201921621398 U CN 201921621398U CN 210676305 U CN210676305 U CN 210676305U
- Authority
- CN
- China
- Prior art keywords
- binocular vision
- camera
- automatic welding
- computer
- trapping device
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Measurement Of Optical Distance (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The utility model discloses an automatic welding smoke trapping device based on binocular vision, which comprises a binocular vision device, a smoke absorbing pipe, a mechanical arm and a computer, wherein the binocular vision device is used for image acquisition and transmission to the computer, and the mechanical arm is controlled by the computer to move the smoke absorbing pipe to a target position; the filter is adopted to inhibit ambient light and noise, and the signal-to-noise ratio of a visual system is improved through program operation; through automatic size selection, the single chip microcomputer and the sliding table are used for automatically controlling the position of the right camera, so that the adaptability of the welding fume collecting device to different stations, different environments and different distances is enhanced; the accuracy and stability of the vision system are improved by using ultrasonic distance measurement and gyroscope attitude measurement.
Description
Technical Field
The utility model relates to a binocular vision field, in particular to automatic cigarette entrapment device that welds based on binocular vision.
Background
The binocular vision space positioning technology is an important non-contact three-dimensional measurement technology developed based on computer vision, namely two cameras at different positions shoot the same scene, and the three-dimensional coordinate position of a space point is obtained by calculating the parallax of the space point in two images. The binocular vision space positioning technology has the advantages of high measuring speed, good real-time performance, simple structure, stable system and the like, and has wide application in production and life. In the existing binocular equipment, the base distance cannot be automatically changed in the measuring process, but the positioning and calculating precision in the depth direction are in inverse proportion to the base line distance, so that the binocular ranging with the fixed base distance cannot adapt to a complex and variable field.
The urban environmental protection industry is the sunrise industry in the world nowadays, and since the 20 th century and the 90 th era, the environmental problems are more and more emphasized by countries in the world, the clean production technology is greatly popularized, the market scale of environmental protection products and services is more and more enlarged, and the intelligent cleaning or environmental protection robot obtains unprecedented attention and support. The electric welding operation in the machining workshop generates great smoke, and the physical health of workers and the surrounding environment are seriously affected. However, at present, few robots are used in a welding workshop for cleaning welding fume, the intelligent level is low, the working efficiency is not high, and good cleaning and environmental protection effects cannot be achieved.
SUMMERY OF THE UTILITY MODEL
Aiming at overcoming the defects existing in the prior art, the utility model discloses a binocular vision-based automatic welding smoke trapping device, which comprises a binocular vision device, a smoke absorbing pipe, a mechanical arm and a computer, wherein the binocular vision device is used for collecting images and transmitting the images to the computer, and the computer controls the mechanical arm to move the smoke absorbing pipe to a target position;
binocular vision device includes pole setting, casing, light filter, left camera and right camera, set up in the casing and be used for the installation left side camera with the installation cavity of right side camera, the light filter sets up the opening part of installation cavity, right side camera is fixed to be set up in the installation cavity, left side camera horizontal slip's setting is in the installation cavity.
Further, the left camera is arranged in the installation cavity in a sliding mode through the sliding table.
Further, the left camera and the right camera are Cognex industrial cameras with the model number of In-Sightdigital CCD 800-.
Further, still include ultrasonic ranging sensor and baffle, set up on the casing lateral wall with the first spout of installation cavity intercommunication, ultrasonic ranging sensor sets up the one end of first spout, the baffle extends and sets up the connecting rod, the connecting rod passes first spout with left camera is connected, and the baffle with ultrasonic ranging sensor flushes and sets up relatively.
Further, still include the gyroscope, the gyroscope sets up on the left camera.
6. An automatic welding fume trapping method based on binocular vision comprises the following steps:
s1: calibrating the left camera and the right camera, establishing a coordinate system of a camera model, and solving a conversion relation between an image coordinate system and a world coordinate system in the camera model;
s2: acquiring images through a left camera and a right camera, and processing the acquired images;
s3: detecting welding spots of the image shot by the left camera;
s4: performing stereo matching on the left image and the right image to obtain a parallax matrix, acquiring a depth map by using a parallax principle, and calculating a three-dimensional space coordinate;
s5: and the computer controls the mechanical arm to move according to the coordinates, and the welding fume is collected through the fume suction pipe.
Further, in step S1, the distance relationship between the left and right cameras is acquired by the ultrasonic ranging sensor and expressed as a translation vector of 3 × 1; acquiring angle variables of the left camera relative to the initial installation position in the x, y and z directions through a gyroscope, and expressing the angle variables as a 3 multiplied by 3 rotation matrix; and calculating an adjusted rotation matrix and an adjusted translation vector:
RN＝RM·RO
TN＝TO+TM
wherein R isNFor an adjusted 3 × 3 rotation matrix, RMFor a measured 3X 3 rotation matrix, ROFor a nominal 3X 3 rotation matrix, TNFor an adjusted 3 × 1 translation vector, TMFor a measured translation vector of 3X 1, TOIs a nominal 3 x 1 translation vector.
Further, in step S2, the signal-to-noise ratio is increased by using the filter, so that the gray value of the dark current noise of the camera relative to the bright point is small; then, performing single thresholding treatment on the left image and the right image, filtering noise and enhancing bright spots; and finally, extracting all brightest pixels in the image as bright point areas through gray value detection, and calculating the pixel coordinates of the area center as the bright points.
Further, the threshold is 250, and then the pixel position with the gray value larger than 250 is extracted and enhanced.
Further, in step S4, the computer adjusts the base distance of the left and right cameras through the sliding table by using the calculated area of the bright spot as a scale factor; the closer the point to the cameras, the more the bright points are, the larger the parallax of the bright points in the left camera and the right camera is, and the base distance between the left camera and the right camera is reduced; the farther a point is from the camera, the smaller its parallax in the left and right cameras, and the larger the base distance between the left and right cameras.
The utility model discloses the beneficial effect who gains:
the utility model uses the binocular vision coordinate positioning mode to match with the smoking pipe, and effectively collects the welding smoke; the protection of the camera is realized through the arrangement of the filter, the influence of high-brightness noise in the environment on the bright light at the welding point in the welding process can be weakened, the lens can be protected to a certain degree, and the processing of the computer on the image is facilitated. The filter is adopted to inhibit ambient light and noise, and the signal-to-noise ratio of a visual system is improved through program operation; through automatic size selection, the single chip microcomputer and the sliding table are used for automatically controlling the position of the right camera, so that the adaptability of the welding fume collecting device to different stations, different environments and different distances is enhanced; the accuracy and stability of the vision system are improved by using ultrasonic distance measurement and gyroscope attitude measurement.
Drawings
Fig. 1 is a schematic structural view of an automatic welding smoke trapping device based on binocular vision according to the present invention;
FIG. 2 is a schematic diagram of the mechanism of the binocular vision apparatus;
FIG. 3 is a schematic view of the internal mechanism of the housing;
FIG. 4 is a schematic diagram of a connection circuit of the single chip microcomputer;
the reference numbers are as follows:
1. binocular vision device, 2, smoking pipe, 3, robotic arm, 4, computer, 11, pole setting, 12, casing, 13, light filter, 14, left camera, 15, right camera, 16, slip table, 17, ultrasonic ranging sensor, 18, baffle, 19, connecting rod, 20, gyroscope, 121, installation cavity, 122, first spout, 123, opening, 124, recess, 125, second spout.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more clearly understood, the present invention will be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the invention.
An automatic welding smoke trapping device based on binocular vision is shown in figure 1 and comprises a binocular vision device 1, a smoke absorbing pipe 2, a mechanical arm 3 and a computer 4, wherein the binocular vision device 1 is used for collecting images and transmitting the images to the computer 4, and the computer 4 controls the mechanical arm 3 to move the smoke absorbing pipe 2 to a target position; smoking the target location. Wherein, the smoking pipe 2 can be a negative pressure collecting device.
As shown in fig. 2-3, the binocular vision device includes a vertical rod 11, a housing 12, a filter 13, a left camera 14 and a right camera 15, a mounting cavity 121 for mounting the left camera 14 and the right camera 15 is arranged in the housing 12, the filter 13 is arranged at an opening of the mounting cavity 121, the right camera 15 is fixedly arranged in the mounting cavity 121, and the left camera 14 is horizontally slidably arranged in the mounting cavity 121. Therefore, the housing 12 and the optical filter 13 are matched with each other, so that the installation cavity 121 is equivalent to a sealed space, welding fume in the environment can be effectively prevented from entering the installation cavity 121, and the left camera 14 and the right camera 15 are polluted to cause damage to the cameras. In addition, the left camera 14 can be movably arranged, so that the distance between the left camera 14 and the right camera 15 can be adjusted, the distance is larger when a short-distance object is observed, the coordinate calculation result is more accurate, the distance is smaller when a long-distance object is observed, and the coordinate calculation result is more accurate. Wherein, Cognex industrial cameras can be used as the left camera 14 and the right camera 15, and the model is In-Sight Digital CCD 800-.
In one embodiment, as shown in fig. 1 to 4, the device further includes a sliding table 16, the sliding table 16 is disposed in the installation cavity 121, the left camera 14 is installed on the sliding table 16, and the sliding table 16 is controlled by the single chip microcomputer to horizontally reciprocate. Wherein, slip table 16 uses the brand to be THK, and the model is that STM32F103C8T6 can be used to the CSKR33 singlechip.
In an embodiment, as shown in fig. 1 to 4, the ultrasonic ranging device further includes an ultrasonic ranging sensor 17 and a baffle 18, a first sliding groove 122 communicated with the mounting cavity 121 is formed in a side wall above the housing 12, the ultrasonic ranging sensor 17 is disposed at one end of the first sliding groove 122, the baffle 18 extends to form a connecting rod 19, the connecting rod 19 passes through the first sliding groove 122 to be connected with the left camera 14, and the baffle 18 is flush with and opposite to the ultrasonic ranging sensor 17. The ultrasonic distance measuring sensor 17 may be a HY-SRF05 product. Through the mutual cooperation of the ultrasonic ranging sensor 17 and the baffle 18, the distance between the left camera 14 and the right camera 15 can be measured, and then the displacement data measured by the ultrasonic ranging module can be correspondingly adjusted without repeated calibration.
In one embodiment, as shown in fig. 1-4, an opening 123 is opened on one side wall of the housing 12 for allowing the filter to pass through, and a groove 124 for fixing the filter 13 is disposed at a corresponding position on the other side wall of the housing 12. Preferably, the upper wall and the lower wall of the housing 2 at the opening 123 are provided with symmetrical second chutes 125, and the upper side and the lower side of the optical filter 3 are placed in the second chutes 125. The filter 13 is supported and guided to move horizontally by the second chute 125. Further, the sealing property of the mounting cavity 121 and the stability of the filter 13 during movement are improved. Meanwhile, the optical filter 13 can weaken the influence of high-brightness noise in the environment on the bright light at the welding point in welding, and can protect the lens to a certain extent. Of course, the filter 13 may use an attenuation sheet.
In one embodiment, as shown in fig. 1-4, a gyroscope 20 is also included, with the gyroscope 20 being disposed on the left camera 14. Wherein, the type of the gyroscope is GY-9250, and the type of the chip is MPU-9250.
Wherein, ultrasonic ranging sensor 17 and gyroscope all carry out data acquisition through the singlechip to data transfer to computer 4 carries out data processing through the computer.
The utility model also discloses an automatic welding smoke trapping method based on binocular vision, which comprises the following steps,
s1: calibrating the left camera and the right camera, establishing a coordinate system of a camera model, and solving a conversion relation between an image coordinate system and a world coordinate system in the camera model;
s2: acquiring images through a left camera and a right camera, and processing the acquired images;
s3: detecting welding spots of the image shot by the left camera;
s4: performing stereo matching on the left image and the right image to obtain a parallax matrix, acquiring a depth map by using a parallax principle, and calculating a three-dimensional space coordinate;
s5: and the computer controls the mechanical arm to move according to the coordinates, and the welding fume is collected through the fume suction pipe.
No matter in the installation process of the visual equipment, due to improper manual operation, collision, extrusion, deformation and the like, errors can occur in the calibration parameters of the left camera and the right camera, or when the sliding table automatically controls the camera to change the base distance, ideal horizontal movement cannot be guaranteed, deviation of an inclination angle cannot occur, and errors can be inevitably introduced to the calibration parameters. However, recalibration requires a large amount of manpower and material resources, the rotation matrix has a large influence on subsequent image matching based on epipolar constraint, and the matching requires that the coplanar row alignment of the left and right images is realized by using internal and external parameters, the rotation matrix, the translation vector and the like of the camera. Preferably, in step S1, the distance relationship between the left and right cameras is acquired by the ultrasonic distance measuring sensor and expressed as a translation vector of 3 × 1; acquiring angle variables of the left camera relative to the initial installation position in the x, y and z directions through a gyroscope, and expressing the angle variables as a 3 multiplied by 3 rotation matrix; and calculating an adjusted rotation matrix and an adjusted translation vector:
RN＝RM·RO
TN＝TO+TM
wherein R isNFor an adjusted 3 × 3 rotation matrix, RMFor a measured 3X 3 rotation matrix, ROFor a nominal 3X 3 rotation matrix, TNFor an adjusted 3 × 1 translation vector, TMFor a measured translation vector of 3X 1, TOIs a nominal 3 x 1 translation vector.
Preferably, in step S2, the filter 13 may be an attenuator, and the filter is used to increase the signal-to-noise ratio, so that the dark current noise of the camera itself is small relative to the gray value of the bright point; then, performing single thresholding treatment on the left image and the right image, filtering noise and enhancing bright spots; and finally, extracting all brightest pixels in the image as bright point areas through gray value detection, and calculating the pixel coordinates of the area center as the bright points. The attenuation sheet is used for filtering out the environmental noise darker than the electric welding arc, such as lighting light, object reflection, natural light and the like. In a preferred embodiment, the threshold in the computer is 250, and pixel locations with gray values greater than 250 are extracted and enhanced.
In step S4, the disparity matrix is determined by using a stereo matching algorithm, and the specific process is as follows:
(1) constructing a small window, similar to a convolution kernel;
(2) covering the left image by using a window, and selecting all pixel points in the area covered by the window;
(3) selecting pixel points of a coverage area on the same polar line of a right image aligned with the left image in a coplanar row;
(4) subtracting the right coverage area from the left coverage area, and solving the sum of absolute values of gray differences of all pixel points;
(5) moving the window of the right image along the epipolar line, repeating the processes (3) and (4) (here, searching is performed only in the range of the same epipolar line);
(6) the window with the smallest difference value in this range is found, i.e. the best matching pixel block of the left and right images is found.
When the binocular parallax error correction method is used, the computer adjusts the base distance of the left camera and the right camera by taking the calculated area of the bright spots as a scale factor, and the measurement error in the depth direction is inversely proportional to the base distance according to the binocular parallax principle. The closer the point to the cameras, the more the bright points are, the larger the parallax of the bright points in the left camera and the right camera is, and the base distance between the left camera and the right camera is reduced; the farther a point is from the camera, the smaller its parallax in the left and right cameras. Therefore, in the case of distance measurement, the left and right cameras should maintain a large base distance and parallax to ensure the accuracy and stability of coordinate calculation.
In addition, the binocular vision coordinate positioning method is disclosed in chinese patent CN 104933718B, and will not be elaborated herein.
The above are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention; the scope of the appended claims should be accorded the broadest interpretation so as to encompass all such modifications and equivalent arrangements as is within the spirit and scope of the present invention.
Claims (5)
1. The automatic welding smoke trapping device based on binocular vision is characterized by comprising a binocular vision device, a smoke absorbing pipe, a mechanical arm and a computer, wherein the binocular vision device is used for image acquisition and transmission to the computer, and the computer controls the mechanical arm to move the smoke absorbing pipe to a target position;
binocular vision device includes pole setting, casing, light filter, left camera and right camera, set up in the casing and be used for the installation left side camera with the installation cavity of right side camera, the light filter sets up the opening part of installation cavity, right side camera is fixed to be set up in the installation cavity, left side camera horizontal slip's setting is in the installation cavity.
2. The binocular vision based automatic welding smoke trapping device is characterized in that the left camera is arranged in the mounting cavity in a sliding mode through a sliding table.
3. The binocular vision based automatic welding smoke trapping device as claimed In claim 1, wherein the left camera and the right camera are Cognex industrial cameras with model number In-Sight Digital CCD 800-.
4. The binocular vision based automatic welding smoke trapping device is characterized by further comprising an ultrasonic ranging sensor and a baffle, wherein a first sliding groove communicated with the installation cavity is formed in the side wall of the shell, the ultrasonic ranging sensor is arranged at one end of the first sliding groove, a connecting rod is arranged in the baffle in an extending mode and penetrates through the first sliding groove to be connected with the left camera, and the baffle and the ultrasonic ranging sensor are flush and arranged oppositely.
5. The binocular vision based automatic welding fume trapping device according to claim 1, further comprising a gyroscope disposed on the left camera.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201921621398.5U CN210676305U (en) | 2019-09-26 | 2019-09-26 | Automatic welding smoke trapping device based on binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201921621398.5U CN210676305U (en) | 2019-09-26 | 2019-09-26 | Automatic welding smoke trapping device based on binocular vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN210676305U true CN210676305U (en) | 2020-06-05 |
Family
ID=70901388
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201921621398.5U Active CN210676305U (en) | 2019-09-26 | 2019-09-26 | Automatic welding smoke trapping device based on binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN210676305U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110560443A (en) * | 2019-09-26 | 2019-12-13 | 太仓光电技术研究所 | automatic welding smoke trapping device and method based on binocular vision |
| WO2023071955A1 (en) * | 2021-10-27 | 2023-05-04 | 维沃移动通信有限公司 | Photographing device |
-
2019
- 2019-09-26 CN CN201921621398.5U patent/CN210676305U/en active Active
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110560443A (en) * | 2019-09-26 | 2019-12-13 | 太仓光电技术研究所 | automatic welding smoke trapping device and method based on binocular vision |
| WO2023071955A1 (en) * | 2021-10-27 | 2023-05-04 | 维沃移动通信有限公司 | Photographing device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108982546B (en) | Intelligent robot gluing quality detection system and method | |
| CN109544679B (en) | Three-dimensional reconstruction method for inner wall of pipeline | |
| CN108592788B (en) | 3D intelligent camera system facing spraying production line and workpiece online measurement method | |
| CN210676305U (en) | Automatic welding smoke trapping device based on binocular vision | |
| CN110246175A (en) | Intelligent Mobile Robot image detecting system and method for the panorama camera in conjunction with holder camera | |
| CN109483531A (en) | It is a kind of to pinpoint the NI Vision Builder for Automated Inspection and method for picking and placing FPC plate for manipulator | |
| CN101506614A (en) | Method and apparatus for 3-dimensional vision and inspection of ball and like protrusions of electronic components | |
| CN111220126A (en) | Space object pose measurement method based on point features and monocular camera | |
| CN110425996A (en) | Workpiece size measurement method based on binocular stereo vision | |
| CN112629441A (en) | 3D curved surface glass contour scanning detection method and system based on line structured light dynamic vision | |
| CN103398669A (en) | Multi-axis linkage visual inspection method and equipment used for measuring free-form surface | |
| CN111127540B (en) | Automatic distance measurement method and system for three-dimensional virtual space | |
| CN116958146B (en) | Acquisition method and device of 3D point cloud and electronic device | |
| CN115880296B (en) | Machine vision-based prefabricated part quality detection method and device | |
| CN113310433A (en) | Virtual binocular stereo vision measuring method based on line structured light | |
| CN114581379B (en) | Sealant detection method and device | |
| CN109696240A (en) | Detection method, device and the readable storage medium storing program for executing of the semiconductor laser overall situation uniformity | |
| CN114383543B (en) | WAAM molten pool three-dimensional reconstruction method | |
| CN110560443A (en) | automatic welding smoke trapping device and method based on binocular vision | |
| CN114750154A (en) | Dynamic target identification, positioning and grabbing method for distribution network live working robot | |
| Ye et al. | Weld seam tracking based on laser imaging binary image preprocessing | |
| CN116393982B (en) | Screw locking method and device based on machine vision | |
| CN110706333B (en) | Reconstruction method based on manual calibration of pipeline position and leakage point thereof | |
| KR20100081881A (en) | Data matching device and method, and robot using these | |
| CN108180825A (en) | A kind of identification of cuboid object dimensional and localization method based on line-structured light |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |