WO2019000640A1 - 模块化视觉系统制作方法 - Google Patents

模块化视觉系统制作方法 Download PDF

Info

Publication number
WO2019000640A1
WO2019000640A1 PCT/CN2017/100808 CN2017100808W WO2019000640A1 WO 2019000640 A1 WO2019000640 A1 WO 2019000640A1 CN 2017100808 W CN2017100808 W CN 2017100808W WO 2019000640 A1 WO2019000640 A1 WO 2019000640A1
Authority
WO
WIPO (PCT)
Prior art keywords
robot
calibration
camera
machine tool
light
Prior art date
Application number
PCT/CN2017/100808
Other languages
English (en)
French (fr)
Inventor
杨煜俊
毕辉
张振普
郑桢才
欧伟光
Original Assignee
巨轮（广州）机器人与智能制造有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 巨轮（广州）机器人与智能制造有限公司 filed Critical 巨轮（广州）机器人与智能制造有限公司
Publication of WO2019000640A1 publication Critical patent/WO2019000640A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
    • B23Q7/00—Arrangements for handling work specially combined with or arranged in, or specially adapted for use in connection with, machine tools, e.g. for conveying, loading, positioning, discharging, sorting
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
    • B23Q17/00—Arrangements for observing, indicating or measuring on machine tools
    • B23Q17/24—Arrangements for observing, indicating or measuring on machine tools using optics or electromagnetic waves
    • B23Q17/2404—Arrangements for improving direct observation of the working space, e.g. using mirrors or lamps
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23Q—DETAILS, COMPONENTS, OR ACCESSORIES FOR MACHINE TOOLS, e.g. ARRANGEMENTS FOR COPYING OR CONTROLLING; MACHINE TOOLS IN GENERAL CHARACTERISED BY THE CONSTRUCTION OF PARTICULAR DETAILS OR COMPONENTS; COMBINATIONS OR ASSOCIATIONS OF METAL-WORKING MACHINES, NOT DIRECTED TO A PARTICULAR RESULT
    • B23Q17/00—Arrangements for observing, indicating or measuring on machine tools
    • B23Q17/24—Arrangements for observing, indicating or measuring on machine tools using optics or electromagnetic waves
    • B23Q17/2409—Arrangements for indirect observation of the working space using image recording means, e.g. a camera

Definitions

  • the invention relates to the field of automation devices for visually matching robots and machine tools, in particular to a vision system standardized by various functional modules, which enables the vision system to adapt to the rapid conversion of different workpieces.
  • the vision system uses cameras and software instead of human eyes to make the automation device have the function of segmenting, classifying, identifying, tracking, and discriminating decisions that are similar to humans.
  • the automation mode requires higher speed, accuracy, resolution, stability and adaptability, and it takes a long time to operate, and the manual can no longer meet such requirements, and the visual system non-contact measurement can solve this.
  • One problem can meet the needs of automated production models and protect machines and products.
  • the vision system used in conjunction with industrial robots is a customized product for a certain kind of workpiece.
  • the development cycle is long, the function is single, and the adaptability to the external environment is not strong.
  • the illumination system, camera calibration, and template processing are required.
  • Algorithms, communication interfaces, parameter settings, and image processing methods have been modified to accommodate new features of new artifacts. This has seriously affected the upgrading of customer products, does not meet the technical requirements of intelligent manufacturing, and restricts the promotion and application of vision systems.
  • Another problem is that most vision systems use more cumbersome calibration methods between cameras, robots, and machine tools.
  • the average customer cannot calibrate high-precision results, which is difficult, problem-prone, and extremely inefficient to implement.
  • the vision system is typically designed as a single, custom product for the identification of certain workpieces.
  • the embodiment of the present invention proposes a visual system standardized by each functional module, and quickly forms a new visual inspection system for identification of different workpieces through various standardized modules.
  • the camera, the robot, and the machine tool are fixedly connected. After an accurate calibration, the calibration result is saved, and after each re-installation, only a partial fine adjustment is performed, so that the relative position between the components can be accurately aligned. . Moreover, the on-site operation engineers do not need to be professionally trained, as long as the simple steps of the description process, the entire calibration process can be completed quickly, and the output parameters can be more directly controlled without relying on complicated coordinate transformation.
  • Embodiments of the present invention provide a standardized vision system having modules including an illumination system, a camera calibration, a template processing algorithm, a communication interface, a parameter setting and an image processing method, and a system measurement toolkit.
  • the camera mounting device adopts a full-scale auger adjustment with self-locking function, which greatly improves the stability and adaptability of the system, and can effectively reduce duplication of labor and improve work efficiency.
  • the resulting data results visual system provides simpler, faster, and more cost effective detection and processing.
  • the illumination system is used as a function module, and the light source installation positions of the top, the side and the back are designed, and the angle can be adjusted, so that the illumination can be realized in any direction; the light source of different colors and different specifications such as white light and red light is equipped. It is possible to make a visual illumination system from different directions, using light sources of different colors and different intensities.
  • the system calibration module includes camera calibration, coordinate transformation between the camera and the industrial robot, and coordinate transformation between the robot and the machine.
  • the calibration system is standardized, and the calibration plate used, the placement method of the calibration plate, the number of images taken, and the coordinate mode of the robot are fixed.
  • the operation of the whole process is performed.
  • the calibration is performed again. Only the relative coordinate system of the camera and the robot is locally adjusted to compensate for the error introduced by the disassembly and assembly, and the rapid calibration is realized.
  • the template processing algorithm module extracts the general characteristics of the template processing and realizes the standardization of the external interface of the module. Different types of workpieces within a certain volume range, such as hardware, food, plastic products, etc., can be identified by the same algorithm.
  • Communication interface module for different manufacturers of robots and CNC machine tools, based on the standard TCP / IP protocol, extended some specific functions for compatibility with different data interaction methods. At present, it can realize communication with equipments of ABB, FANUC, KUKA, YASKAWA, etc., and provides an expansion interface, which can customize data format and communication mode.
  • Parameter setting and image processing module uses standardized design interface and operation mode, image processing adopts color and black and white interface consistent design, adds image processing standard function tool library, can be freely split and recombined, and combines good results , saved as a new tool.
  • System toolkit module for color extraction, geometric measurement, function correction, blurring, flaw detection, barcode scanning, character recognition, edge enhancement, real-time differential, data statistics, etc. Make each tool a standard software toolkit. Different applications can be combined with any tool. Now specific features.
  • the embodiment of the invention provides a complete automation device for the vision system to cooperate with the robot and the machine tool, which has:
  • An industrial camera for obtaining an image from a target object and a mounting device thereof
  • Visual fit with machine tool loading and unloading standard units can be used to improve many industrial and manufacturing processes. These include, but are not limited to, the detection and installation of parts in the mobile phone assembly line; the sorting and processing of plastic buckles; the automatic machining of the tire mold template; the detection and handling of the motor components; the identification in the wood processing plant Logs for cutting or cutting; identification and packaging of bread production lines; brushing in moon cake production; grinding of welds caused by welding; grinding of some die-casting parts; automatic removal of sub-fruits in size or shape.
  • all of these processes require manual intervention and calculations, high requirements for field engineers, database management, and complex data processing, and are often costly, labor intensive, and not always accurate or automatic.
  • Figure 1 shows a camera and light source mounting device.
  • Figure 2 shows the workpiece processing chuck in the machine tool.
  • Figure 3 shows the development process of visual software.
  • Figure 4 shows the structure of the visual software.
  • Figure 5 shows a partial function interface
  • the invention provides a vision system for positioning and detecting various types of hardware, electronic components, plastics and rubber products, and realizing automatic processing of workpieces.
  • the vision system includes a camera mounting device, an illumination system, a system calibration, a template creation and matching algorithm, a parameter setting and image processing method, a system measurement toolkit, and a communication interface between the system and the robot and the machine tool. Standardize each function module to achieve fast conversion of different workpieces.
  • the workpiece is placed at the position to be taken, and the photoelectric sensor is used to detect whether the workpiece reaches the reclaiming area.
  • the target workpiece is classified, identified, tracked, and discriminated by the visual system.
  • the functional modules of the visual system are Standardized, it can simultaneously detect a variety of workpieces, and can realize the fast switching of different workpieces.
  • the workpiece can be grasped.
  • the installation of the robot is based on the base and the ground. The angle is convenient for feeding with the machine tool.
  • the robot puts the workpiece into the chuck of the machine. In order to achieve high-precision positioning, the chuck performs positioning correction on the workpiece.
  • the CNC machine tool performs different types on different workpieces.
  • the processed workpiece is taken out from the machine tool, and after the workpiece is inspected by the vision system, it reaches the workpiece storage area, and the photoelectric sensor detects whether the workpiece is reached, and the processed workpiece is processed according to the detection result of the vision system.
  • Classified storage At the same time, the operation safety of the whole unit is protected.
  • a laser scanner is installed at the base position of the robot, and the laser scanner cannot install a fence in the protection area.
  • Figure 1 shows a mounting device for a camera and a light source.
  • the camera and the light source mounting device must have an all-round adjustment and locking function, and the adjustment uses a screw with a small step to achieve fine adjustment. Adjustment of the camera and light source can be done separately.
  • the light source is mainly adjusted from three aspects, namely the light direction, light intensity, light color.
  • the light source mounting positions of the top, the side and the back are designed, and the angle can be adjusted, so that the lighting can be performed in any direction;
  • the light sources of different specifications are provided, and each light source is equipped with a power conditioner;
  • Light sources of different colors such as white light, red light, and blue light. It is possible to make a visual illumination system from different directions, using light sources of different colors and different intensities.
  • the camera and the robot are fixedly connected in the design. After an accurate calibration, the calibration results are saved. After each re-installation, only local fine-tuning is required. A very accurate calibration of the relative position between the various components can be achieved. Moreover, the on-site operation engineers do not need to be professionally trained, as long as the simple steps of the description process, the entire calibration process can be completed quickly, and the output parameters can be more directly controlled without relying on complicated coordinate transformation.
  • the first calibration needs to simultaneously calibrate the camera's internal and external parameters. It is necessary to collect multiple calibration images. Generally, more than 13 images are collected before the convergence result can be obtained.
  • the camera external parameters are calibrated and collected.
  • One calibration image can be used. Then, the coordinate transformation of the robot and the camera is established, the current position of the template workpiece is acquired, and the posture of the robot when the template is captured, the two are matched, and the relative change amount between the workpieces is superimposed on the posture of the robot during the detection. In the second calibration, you only need to teach the robot's grasping posture. Finally, the coordinate transformation of the robot and the machine tool is established, and the loading position of the machine tool is obtained by the robot teaching device.
  • Figure 2 shows the workpiece chuck inside the machine. Since the absolute positioning accuracy of the robot is difficult to meet the machining accuracy requirements of the CNC machine tool, if the robot is directly discharged into the machine tool, it is difficult to ensure the machine tool consistency.
  • a calibratable chuck and an electromagnetic chuck are used, which can effectively adjust the placement accuracy of the workpiece and ensure a good processing effect of the machine tool.
  • Figure 3 shows the development process of visual software.
  • the development and application of the vision system in this embodiment is determined.
  • the development and application of the vision system in this embodiment is determined.
  • the development and application of the vision system in this embodiment is determined.
  • consult a large number of literatures and related theories and algorithms understand the research status at home and abroad, and lay the theoretical foundation for the research. Again, collect the needs analysis of the subject, and carry out conceptual design and detailed design, write code to achieve various algorithms and software system-related responses, and compare the practices and characteristics of mature vision software such as VisionPro, Halcon, and NI Vision. experience.
  • research on some common cases use individual typical cases to detect the visual system, inspire and guide the project practice, and test the practice through the practical level, hands-on ability and project completion level in the actual work.
  • FIG. 4 shows the structure of the visual software.
  • the visual software mainly includes a parameter setting module, a data acquisition module, a data management module, a data processing module, a data communication module, and a software toolkit. Each module independently completes some functions, and the division of labor between modules is clear, and communication is made through each other.
  • the parameter setting uses a standardized design interface and operation mode.
  • the image processing adopts the design of color and black and white interface consistently.
  • the image processing standard function tool library is added, which can be freely split and reorganized, and the combined result is saved as a new tool. .
  • the software can implement color extraction, geometric measurement, function correction, blurring, flaw detection, barcode scanning, character recognition, edge enhancement, real-time differential, data statistics, and more.
  • Each tool is made into a standard software toolkit, and different applications can be combined to implement specific functions.
  • the general characteristics of the template processing are extracted, and the external interface of the module is standardized. Different types of workpieces within a certain volume range, such as hardware, food, plastic products, etc., can be identified by the same algorithm.
  • the establishment, matching, and detection of templates use standard function entries.
  • Figure 5 shows a partial function interface. The figure shows some of the function interfaces in the software, and the function comments that each function is responsible for.

Landscapes

  • Engineering & Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Optics & Photonics (AREA)
  • Manipulator (AREA)

Abstract

一种模块化视觉系统，用于各类五金件、电子元件、塑料及橡胶制品等的定位和检测，实现模块化设计，不同应用可以快速组合。视觉系统包含相机安装装置、系统标定、模板创建及匹配算法、系统测量工具包以及系统与机器人和机床的通讯接口，并将各个功能模块标准化。

Description

模块化视觉系统制作方法 【技术领域】
本发明涉及视觉配合机器人及机床的自动化装置领域，尤其涉及各个功能模块标准化的视觉系统，前述方法使视觉系统能适应不同工件的快速转换。
【背景技术】
视觉系统是利用摄像机和软件代替人眼使得自动化设备拥有类似于人类的那种对目标进行分割、分类、识别、跟踪、判别决策的功能。
在生产过程中，自动化模式要求有更高的速度、精度、分辨力、稳定性和适应性，并且要长时间的运作，人工已经不能满足这样的需求，而视觉系统非接触性测量能够解决这一问题，可以满足自动化生产模式的需求，并且可以保障机器和产品。
当前，配合工业机器人使用的视觉系统，都是针对某种工件的定制产品，开发周期长，功能单一，对外部环境的适应能力不强；更换工件后，需要对光照系统、相机标定、模板处理算法、通讯接口、参数设置及图像处理方法进行较大改动，来适应新工件的新特性。这样严重影响了客户产品的更新换代，不符合智能制造的技术要求，制约了视觉系统的推广应用。
另一问题在于大部分视觉系统采用较繁琐的相机、机器人、机床之间的标定方法，一般客户不能标定出高精度的结果，其实施起来很困难、而且有问题，且极其低效。
在现有技术中尚未发现具有各个功能模块标准化的视觉系统配合工业机器人应用，因而需要这样的自动化设备以简化和优化许多领域中的工业处理和制造。
【发明内容】
视觉系统一般设计为功能单一的定制产品，用于某种工件的识别。而本发明实施例提出了一种各个功能模块标准化的视觉系统，通过各个标准化的模块，快速组成新的视觉检测系统，用于不同工件的识别。
使相机、机器人、机床之间固定连接，经过一次准确的标定之后，保存标定结果，在每次重新安装之后，只需做局部的微调，就可以实现各个部件之间的相对位置非常精确的校准。而且，对现场操作工程师不需要经过专业培训，只要按说明流程简单的几步操作，很快就可完成整个标定过程，对输出参数进行更直接的控制，无需依赖于复杂的坐标变换。
本发明实施例提供标准化的视觉系统，具有包括光照系统、相机标定、模板处理算法、通讯接口、参数设置及图像处理方法、系统测量工具包等模块。相机安装装置采用全方位螺旋杆调节，带自锁功能，很大程度上提高了系统的稳定性和适应性，可有效减免重复劳动，提高工作效率。所获取的数据结果视觉系统提供更简单的、更快的和效能成本更合算的检测和处理。
其中，光照系统作为一个功能模块，设计了顶部、侧面、背面的光源安装位置，而且角度可以调整，从而可以实现任意方向的打光；配备了白光、红光等不同颜色和不同规格的光源。完全可以做到从不同方向、利用不同颜色和不同强度的光源组成视觉的光照系统。
系统标定模块，包含了相机标定、相机与工业机器人之间的坐标变换、机器人与机床之间的坐标变换。将标定系统进行标准化，使用的标定板、标定板的摆放方式、所取图像数量、机器人取坐标方式都进行固定。首次标定进行整个过程的操作，设备重新拆装之后，再次进行标定时，只需对相机、机器人相对坐标系进行局部调整，补偿由于拆装引入的误差，实现快速标定。
模板处理算法模块，提取模板处理的通用特性，实现模块对外接口的标准化，一定体积范围内的不同类型工件，如五金件、食品、塑胶制品等，都可以使用相同的算法实现识别。
通讯接口模块，针对不同厂家的机器人及数控机床，在标准的TCP/IP协议基础上，扩展了一些特定功能，用于兼容不同的数据交互方式。目前，可以实现和ABB、FANUC、KUKA、YASKAWA等厂商设备进行通讯，同时提供了扩展接口，可以定制数据格式及通讯方式。
参数设置及图像处理模块，参数设置使用标准化的设计界面及操作方式，图像处理采用彩色和黑白接口一致的设计，增加图像处理标准函数工具库，可以自由拆分和重组，并将组合好的结果，作为一个新的工具保存。
系统工具包模块，实现颜色抽取、几何测量、功能补正、模糊处理、瑕疵检测、条码扫描、字符识别、边缘强化、实时差分、数据统计等。将各个工具做成标准的软件工具包，不同的应用可任意组合工具以实 现特定功能。
实质上，本发明实施例提出视觉系统配合机器人及机床的整套自动化装置，其具有：
1、用于从目标物体获得图像的工业相机及其安装装置；
2、对目标图像进行分割、分类、识别、跟踪、判别决策的视觉软件；
3、实现搬运动作的工业机器人及物料传动装置；
4、实现工件加工的数控机床。
视觉配合机床上下料标准单元可用于改善许多工业和制造处理。这些包括但不限于在手机组装生产线上零部件的检测及安装；塑胶卡扣的分拣及加工；轮胎模具模板的数控加工自动上下料；电机组件的产线检测及搬运；木材加工厂中识别原木以进行切除或切割；面包产线的识别及包装；月饼生产中的刷浆；打磨因焊接造成的焊缝；对一些压铸件进行打磨；自动剔出尺寸或形状方面的次果等。目前，所有这些处理均需要人工干预和计算、对现场工程师能力要求高、数据库管理、及数据处理复杂，且通常成本很高、耗费大量的劳动力和时间、及并不总是准确或自动。
【附图说明】
为了更清楚地说明本发明的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图是本发明的一些重要资料，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1示出了相机及光源安装装置。
图2示出了机床内工件加工卡盘。
图3示出了视觉软件的开发过程。
图4示出了视觉软件的结构。
图5示出了部分函数接口。
【具体实施方式】
下面将结合附图对模块化视觉系统制作方法进行描述。本发明提出了一种视觉系统，用于各类五金件、电子元件、塑料及橡胶制品等的定位和检测，实现工件的自动加工。所述视觉系统包含相机安装装置、光照系统、系统标定、模板创建及匹配算法、参数设置及图像处理方法、系统测量工具包以及系统与机器人和机床的通讯接口。并将各个功能模块标准化，实现不同工件的快速转换。
第一，将工件放置于待取料位置，通过光电传感器，检测工件是否到达取料区；第二，通过视觉系统对目标工件进行分类、识别、跟踪、判别决策，视觉系统的各个功能模块是标准化的，能够同时检测多种工件，并可以实现快速切换不同工件的识别；第三，根据视觉系统及检测传感器传送过来的数据，实现对工件的抓取，机器人的安装采用底座和地面成一定的夹角，方便与机床配合上料；第四，机器人将工件放入机床卡盘内，为了实现高精度的定位，卡盘对工件进行定位校正；第五，数控机床对不同工件进行不同类型的加工操作；最后，将加工完成的工件从机床内取出，工件经过视觉系统的检测后，到达工件存放区，由光电传感器检测工件是否达到，并根据视觉系统的检测结果对加工完成的工件进行分类存放。同时对整个单元的操作安全作了保护，在机器人的底座位置安装了激光扫描器，激光扫描器不能保护区域加装了围栏。
图1示出了相机及光源的安装装置。为了适应工件的变化，相机及光源的安装装置必须具有全方位调节及锁定功能，而且调节采用步距较小的螺旋杆，实现细微调整。相机和光源的调节都可单独进行。
通常要取得一个好的打光效果，主要从三个方面对光源进行调整，分别是光照方向、光照强度、光色。在本实施例中设计了顶部、侧面、背面的光源安装位置，而且角度可以调整，从而可以实现任意方向的打光；配有不同规格的光源，且每个光源都配有电源调节器；配备了白光、红光、蓝光等不同颜色的光源。完全可以做到从不同方向、利用不同颜色和不同强度的光源组成视觉的光照系统。
系统正常工作之前，要对所有相关部分进行标定，设计中使相机、机器人之间固定连接，经过一次准确的标定之后，保存标定结果，在每次重新安装之后，只需做局部的微调，就可以实现各个部件之间的相对位置非常精确的校准。而且，对现场操作工程师不需要经过专业培训，只要按说明流程简单的几步操作，很快就可完成整个标定过程，对输出参数进行更直接的控制，无需依赖于复杂的坐标变换。首先，对相机 进行标定，首次标定需对相机的内参和外参同时标定，需要采集多张标定图像，一般要采集13张以上，才可得到收敛结果；再次标定时，只需对相机外参进行校准，采集1张标定图像即可。然后，建立机器人和相机的坐标变换，获取模板工件的当前位置，及机器人抓取模板时的姿态，将二者相匹配，在做检测时将工件之间的相对变化量，叠加到机器人的姿态中；再次标定时，只需示教出机器人的抓取姿态即可。最后，建立机器人和机床的坐标变换，通过机器人示教器获取机床的上料位置。
图2示出了机床内部的工件卡盘。由于机器人的绝对定位精度很难达到数控机床的加工精度要求，如果采用机器人直接向机床内放料，很难保证机床的加工一致性。本实施例中采用了可校正卡盘和电磁吸盘，可有效调整工件的放置精度，保证机床的良好加工效果。
图3示出了视觉软件的开发过程。首先，根据现有的文献资料，结合研究条件和自身研究优势，确定本实施例中视觉系统开发与应用。其次，查阅大量文献和相关理论、算法，了解国内外研究现状，奠定课题研究的理论基础。再次，收集课题的需求分析，并进行概念设计和详细设计，编写代码实现各种算法及软件系统相关的各个响应，同时比较研究VisionPro、Halcon、NI Vision等成熟视觉软件的做法和特点，总结成功经验。然后，对一些常用案例进行研究，利用个别典型案例检测视觉系统，启发并引导项目实践，并通过在实际工作中的实践水平、动手能力与项目完成程度来检验实践。最后，结合理论实验、比较研究和个案研究得出总结与启示，针对视觉软件在机床上下料标准单元中的应用，最终形成一个完善的应用模式，并对其效果进行新一轮的验证以进一步检测前期成果，进而得出研究结论，撰写相应的文档。
图4示出了视觉软件的结构。视觉软件主要包括参数设置模块、数据采集模块、数据管理模块、数据处理模块、数据通讯模块、软件工具包组成。每个模块独立完成部分功能，模块间分工明确，相互之间通过接口通讯。
针对不同厂家的机器人及数控机床，在标准的TCP/IP协议基础上，扩展了一些特定功能，用于兼容不同的数据交互方式。目前，可以实现和ABB、FANUC、KUKA、YASKAWA等厂商设备进行通讯，同时提供了扩展接口，可以定制数据格式及通讯方式。
参数设置使用标准化的设计界面及操作方式，图像处理采用彩色和黑白接口一致的设计，增加图像处理标准函数工具库，可以自由拆分和重组，并将组合好的结果，作为一个新的工具保存。
软件可以实现颜色抽取、几何测量、功能补正、模糊处理、瑕疵检测、条码扫描、字符识别、边缘强化、实时差分、数据统计等。将各个工具做成标准的软件工具包，不同的应用可任意组合工具以实现特定功能。
提取模板处理的通用特性，实现模块对外接口的标准化，一定体积范围内的不同类型工件，如五金件、食品、塑胶制品等，都可以使用相同的算法实现识别。模板的建立、匹配、检测都采用标准的函数入口。
图5示出了部分函数接口。图中展示了软件中的部分函数接口，及各个函数负责完成的功能注释。

Claims (12)

  1. 模块化视觉系统制作方法，其特征在于，包括:用于从目标物体获得图像的工业相机及其安装装置；用于提高采集图像质量的光照系统；对目标图像进行分割、分类、识别、跟踪、判别决策的视觉软件；实现搬运动作的工业机器人及物料传动装置；各种光电传感器及外围配套设备；实现工件加工的数控机床。
  2. 根据权利要求1所述的方法，其中所述从目标物体获得图像的工业相机及其安装装置，具有全方位调节及锁定功能，而且调节采用步距较小的螺旋杆，实现细微调整。相机的调节可单独进行。
  3. 根据权利要求1所述的方法，其中所述提高采集图像质量的光照系统，可从三个方面对光源进行调整，分别是光照方向、光照强度、光色；设计了顶部、侧面、背面的光源安装位置，而且角度可以调整，从而可以实现任意方向的打光；配有不同规格的光源，且每个光源都配有电源调节器；配备了白光、红光、蓝光等不同颜色的光源；完全可以做到从不同方向、利用不同颜色和不同强度的光源组成视觉的光照系统。
  4. 根据权利要求1所述的方法，其中所述实现搬运动作的工业机器人及物料传动装置，机器人底座与物料传动装置固定连接，物料传动装置和相机安装装置固定连接，机器人以一定的倾斜角度安装在底座上，物料输送带采用双层设计，节省上料和下料所占用的空间。
  5. 根据权利要求1所述的方法，其中所述各种光电传感器及外围配套设备，包含:安全防护激光扫描器、到位检测光幕、防护栏、电箱、线缆、定位销、真空吸盘、防护罩、防水气缸、电磁阀、消声器等。
  6. 根据权利要求1所述的方法，其中所述实现工件加工的数控机床，由于机器人的绝对定位精度很难达到数控机床的加工精度要求，如果采用机器人直接 向机床内放料，很难保证机床的加工一致性。采用了可校正卡盘和电磁吸盘，可有效调整工件的放置精度，保证机床的良好加工效果。
  7. 根据权利要求1所述的方法，其中所述对目标图像进行分割、分类、识别、跟踪、判别决策的视觉软件，包含:系统标定、模板创建及匹配算法、参数设置及图像处理方法、系统测量工具包以及系统与机器人和机床的通讯接口。并将各个功能模块标准化，实现不同工件的快速转换。
  8. 根据权利要求7所述的方法，其中所述系统标定，包含:相机标定、相机与工业机器人之间的坐标变换、机器人与机床之间的坐标变换。将标定系统进行标准化，使用的标定板、标定板的摆放方式、所取图像数量、机器人取坐标方式都进行固定。首次标定进行整个过程的操作，设备重新拆装之后，再次进行标定时，只需对相机、机器人相对坐标系进行局部调整，补偿由于拆装引入的误差，实现快速标定。
  9. 根据权利要求7所述的方法，其中所述模板创建及匹配算法，提取模板处理的通用特性，实现模块对外接口的标准化，一定体积范围内的不同类型工件，如五金件、食品、塑胶制品等，都可以使用相同的算法实现识别。
  10. 根据权利要求7所述的方法，其中所述参数设置及图像处理方法，参数设置使用标准化的设计界面及操作方式，图像处理采用彩色和黑白接口一致的设计，增加图像处理标准函数工具库，可以自由拆分和重组，并将组合好的结果，作为一个新的工具保存。
  11. 根据权利要求7所述的方法，其中所述系统测量工具包，包含:颜色抽取、几何测量、功能补正、模糊处理、瑕疵检测、条码扫描、字符识别、边缘强化、实时差分、数据统计等。将各个工具做成标准的软件工具包，不同的应用可任意组合工具以实现特定功能。
  12. 根据权利要求7所述的方法，其中所述系统与机器人和机床的通讯接口，针对不同厂家的机器人及数控机床，在标准的TCP/IP协议基础上，扩展了一些特定功能，用于兼容不同的数据交互方式。目前，可以实现和ABB、FANUC、KUKA、YASKAWA等厂商设备进行通讯，同时提供了扩展接口，可以定制数据格式及通讯方式。
PCT/CN2017/100808 2017-06-27 2017-09-07 模块化视觉系统制作方法 WO2019000640A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710501367.5A CN107139003A (zh) 2017-06-27 2017-06-27 模块化视觉系统制作方法
CN201710501367.5 2017-06-27

Publications (1)

Publication Number Publication Date
WO2019000640A1 true WO2019000640A1 (zh) 2019-01-03

Family

ID=59785133

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/100808 WO2019000640A1 (zh) 2017-06-27 2017-09-07 模块化视觉系统制作方法

Country Status (2)

Country Link
CN (1) CN107139003A (zh)
WO (1) WO2019000640A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3650180A1 (en) * 2018-11-07 2020-05-13 Wolfgang Grigo System for optical inspection
CN109822564B (zh) * 2019-01-14 2022-06-07 巨轮(广州)智能装备有限公司 一种视觉系统的构建方法
CN110200440B (zh) * 2019-06-06 2021-08-03 福建泉州市剑井贸易有限公司 智能家居用护理床
CN115401527B (zh) * 2022-10-08 2024-04-23 瑞安市博业激光应用技术有限公司 可视示教系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2005072917A1 (en) * 2004-01-30 2005-08-11 Wisematic Oy Machine vision controlled robot tool system
CN101456159A (zh) * 2008-08-15 2009-06-17 齐齐哈尔华工机床制造有限公司 火花识别对刀方法及磨削加工自动化系统
CN102120307A (zh) * 2010-12-23 2011-07-13 中国科学院自动化研究所 一种基于视觉信息的工业机器人磨削系统及方法
CN102514002A (zh) * 2011-12-29 2012-06-27 江苏大学 一种数控车床单目视觉上下料机器人系统及方法
CN103406905A (zh) * 2013-08-20 2013-11-27 西北工业大学 一种具有视觉伺服及检测功能的机器人系统
US20150203304A1 (en) * 2014-01-22 2015-07-23 Axium Inc. Vision-Assisted Robotized Depalletizer
CN205572431U (zh) * 2016-01-08 2016-09-14 杭州自动化技术研究院有限公司 一种智能视觉机器人的拼装系统

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20110001460U (ko) * 2009-08-05 2011-02-11 김교월 반도체의 아이씨의 트리밍 검사장치 및 그사용법
JP5653166B2 (ja) * 2010-10-22 2015-01-14 キヤノン株式会社 ロボットセル
CN102706274B (zh) * 2012-04-25 2014-08-06 复旦大学 工业结构化场景中机器视觉精确定位机械零件的系统
CN104786226A (zh) * 2015-03-26 2015-07-22 华南理工大学 抓取在线工件的机器人位姿及运动轨迹定位系统与方法
CN205021618U (zh) * 2015-10-15 2016-02-10 常州铭赛机器人科技股份有限公司 视觉调节机构
CN106770343A (zh) * 2017-03-28 2017-05-31 慧眼自动化科技（广州）有限公司 涂胶产品视觉检测系统及检测方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2005072917A1 (en) * 2004-01-30 2005-08-11 Wisematic Oy Machine vision controlled robot tool system
CN101456159A (zh) * 2008-08-15 2009-06-17 齐齐哈尔华工机床制造有限公司 火花识别对刀方法及磨削加工自动化系统
CN102120307A (zh) * 2010-12-23 2011-07-13 中国科学院自动化研究所 一种基于视觉信息的工业机器人磨削系统及方法
CN102514002A (zh) * 2011-12-29 2012-06-27 江苏大学 一种数控车床单目视觉上下料机器人系统及方法
CN103406905A (zh) * 2013-08-20 2013-11-27 西北工业大学 一种具有视觉伺服及检测功能的机器人系统
US20150203304A1 (en) * 2014-01-22 2015-07-23 Axium Inc. Vision-Assisted Robotized Depalletizer
CN205572431U (zh) * 2016-01-08 2016-09-14 杭州自动化技术研究院有限公司 一种智能视觉机器人的拼装系统

Also Published As

Publication number Publication date
CN107139003A (zh) 2017-09-08

Similar Documents

Publication Publication Date Title
WO2019000640A1 (zh) 模块化视觉系统制作方法
CN108380509B (zh) 基于机器视觉的led灯盘分拣与检测系统
CN110370286B (zh) 基于工业机器人和单目相机的定轴运动刚体空间位置识别方法
CN107590835B (zh) 一种核环境下机械臂工具快换视觉定位系统与定位方法
CN103206949B (zh) 一种自检自定位通用型螺丝机及其定位方法
CN111571190B (zh) 三维立体式可视化自动装配系统及方法
CN103776841B (zh) 合成革缺陷自动检测装置及检测方法
CN106853639A (zh) 一种手机电池自动化装配系统及其控制方法
CN106809730A (zh) 一种基于视觉的集装箱自动对接起吊系统及起吊方法
CN106444674B (zh) 冲压五轴直立式机械手多机互联控制方法及冲压生产线
CN106152935A (zh) 适用于无特征空间曲面毛坯件的激光测量系统及其加工定位方法
CN105690393A (zh) 一种基于机器视觉的四轴并联机器人分拣系统及其分拣方法
CN111645111A (zh) 一种基于工业机器人的智能制造系统及方法
CN101645136A (zh) 图像识别检测系统
CN106647621A (zh) 一种贵金属数控加工定位方法及系统
CN206373896U (zh) 一种六轴工业机器人与智能视觉柔性线装置
Xia et al. Workpieces sorting system based on industrial robot of machine vision
CN113878576B (zh) 一种机器人视觉分拣工艺程序编制方法
CN109752627A (zh) 一种基于机器视觉的馈线终端单元插口状态自动识别的方法
CN105478529A (zh) 全自动冲压生产线制品表面质量在线检测系统及方法
CN110640741A (zh) 一种规则形状工件匹配的抓取工业机器人
CN113641155A (zh) 一种高温锻件检测控制系统
CN111626995B (zh) 一种针对工件的智能嵌件检测方法和装置
CN114252452A (zh) 一种小型回转体外观缺陷和轮廓尺寸在线检测装置及方法
CN114913346B (zh) 一种基于产品颜色及形状识别的智能分拣系统及方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17915898

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17915898

Country of ref document: EP

Kind code of ref document: A1