Example analysis of key technology and application of machine vision

Abstract: this paper briefly introduces the concept related to machine vision, the components of the system, and introduces the key technology related to machine vision, such as lighting source, lens, camera, image acquisition card, image processing card, the integrated visual components, etc., and finally briefly introduces in industrial detection and sorting of agricultural products, robot navigation and other fields of application.

Keywords: machine vision; The LED light source; The CCD camera; Digital signal processing; Visual sensor quote machine vision mainly use computer to simulate human vision function, it extract information from images of the object, and to deal with these information, to understand, finally used for the actual test, measurement and control. A typical industrial machine vision application system includes a light source, optical system, image capture system, image digital module, digital image processing module, intelligent judgment decision-making module and mechanical execution module, as shown in figure 1 in [1]

 

 

 

 

 

 

In this system, first select the appropriate light source and light at right Angle measured target, and then USES the camera to get the image signal of the measured target, through A/D conversion will simulation signal into digital signal transmitted to the dedicated image processing system, based on pixels distribution, brightness and color information, for A variety of operations to extract target features, and then according to the default output criterion for judgment as A result, the control drive actuator corresponding processing.

 

Machine vision is a comprehensive technology, including digital image processing technology, mechanical engineering technology, control technology, lighting technology, optical imaging technology, sensor technology, analog and digital video technology, computer hardware and software technology, man-machine interface technology, etc. Machine vision emphasizes the practicability, can adapt to the harsh environment of industry field, want to have a reasonable price, general industrial interface, high fault tolerance and security, and has strong versatility and portability. It is more of an emphasis on real-time, high speed and high precision.

 

1 machine vision key technology analysis in machine vision system, the key technology of lighting technology, optical lens, camera (CCD), image acquisition card, image processing card (also available image acquisition card + image processing software) and actuator etc.

 

Light source (1) to identify the machine in the application system, the key is to choose of detected optics can be used for lighting light, good light source and lighting scheme is often a key of success or failure in the whole system, it is not simply illuminated object, but to be as much as possible to highlight the object characteristics and the need to detect objects in the part and the secondary part, should as far as possible produce obvious difference between increasing contrast; Also should guarantee the enough overall brightness at the same time, the object's position changes will not affect the imaging quality.

 

Machine vision application system commonly used in the transmission light and reflected light. Light source choice must conform to the required geometry, lighting brightness, uniformity, light spectrum characteristics and so on, should also consider the lighting efficiency and service life. Table 1 lists the main light source of the related properties [2]. Application technology of modern computer (total DiErErQi), (2) optical lens optical lens is equivalent to the lens of the eye. A lens imaging quality fit and unfit quality, that is, its good or not of aberration correction, can be measured through aberration size. Common aberration are spherical aberration, coma, astigmatism, field, distortion, and so on six kinds of color. For prime lens and zoom lens, the prime lens aberration must be of the same class is smaller than a zoom lens, because the zoom lens must compromise, make all sorts of image quality under different focal length are relatively good, are not allowed to appear in the zoom range imaging under the focal length of a bad situation. So in machine vision applications, according to the state of the measured target should first choose prime lens [3]. In addition to considering the magnification of the image size, focal length, Angle, field size, aperture size of concrete of factors such as. Lens and camera installation interface, of course, is also one of the important factors should be considered.

 

 

 

 

 

 

(3) cameras and image acquisition card camera and image acquisition card together to complete the subject object image acquisition and digital. High quality of the image information is the correct judgment and decision making system of the original basis, is another key to the success of the whole system. At present in the machine vision system, CCD camera, with its advantages of small size, reliable performance and high resolution has been widely used. According to the use of CCD camera CCD devices can be divided into two major categories of line formation and formation. A linear CCD camera can only acquire the information of image line, being filmed objects form from the camera must be in a straight line forward, to get the complete image, therefore very suitable for the uniform motion of the material flow at a certain speed of image detection; And plane array CCD camera can get whole image information at a time.

 

The linear CCD camera products mainly include two categories, category by Basler L301bc and L301kc represented.

 

 

 

 

 

 

Its characteristic is: at the same time, R, G, B sensor output of the corresponding subject material information is not in the same position, but three different positions, which has the problem of dislocation of R, G and B.

 

 

 

 

 

 

 

 

 

 

 

Another by Dalsa company TR - 33 and TR - 34 series cameras, as shown in figure 3, this series USES the prism spectroscopic techniques put the material on the same position reflected light red components to the R sensor output, the green component to the G sensor output, blue component to the sensor output, B is the same time, the R, G, B sensor output information is of the same subject materials location of the color information.

 

The first kind of product price is quite cheap, the detection accuracy and real-time demand in system is not very high applications more widely used; The second category of product price is more expensive, is mainly used for high accuracy, strong real-time property, must obtain high quality image of the system. In the machine vision system, image acquisition card is to control the camera taking pictures, complete with digital image acquisition, coordinate the important equipment of the whole system.

 

It usually has the following function module:

 

(1) image signal receiving and A/D conversion module, responsible for the image signal of amplifier and digital; (2) the camera control input/output interface, mainly responsible for coordinating the camera for taking pictures synchronous or asynchronous reset, timing, pictures, etc.; (3) bus interface, is responsible for the internal bus output high-speed digital data through a computer, usually a PCI interface, transfer rate can be up to 130 MBPS, fully competent for high precision image real-time transmission, and takes up less CPU time; (4) display module, is responsible for the high quality of image real-time display; (5) communication interface, is responsible for the communication.

 

 

 

 

 

 

At present, the image acquisition card is a lot of more phyletic, according to the different classification methods, there are black and white image and color image acquisition card, analog signal and digital signal acquisition card, have a composite signal and the RGB components in signal acquisition card input. When choosing image acquisition card, the main consideration should be given to the system functional requirements, the image acquisition precision and the matching factors such as the output signal to the camera.

 

 

 

 

 

 

 

 

 

 

 

(4) image signal processing card image signal processing is the core of the machine vision system, it is equivalent to the human brain, complete the image processing and operation, namely the realization of the algorithm. Machine vision system development, the algorithm is both emphasis and difficulty.

 

Along with the computer technology, microelectronics technology and the rapid development of large scale integrated circuit technology, in order to improve the real-time performance of system, the image processing, a lot of work can be done with the aid of hardware such as DSP, dedicated image signal processing card, etc.; Software mainly completes in the algorithm is very complicated, not mature, still need to continuously explore and change parts of it. In image signal processing time must note: in order to meet the continuous complete processing system for material image real-time demand, must make a frame image processing time is less than or equal to a frame of image acquisition time, namely, image processing speed is greater than or equal to the speed of image acquisition.

 

(5) actuator for a system, finally realizes the system function is done by actuator, it was the last of the system is also the most key link. For different applications, actuators can be mechanical and electrical system, hydraulic system, pneumatic system of a certain. But either way, in addition to the manufacture and assembly accuracy to ensure strict, its dynamic characteristics, especially the rapidity and stability is very important also, must pay enough attention when the design.

 

(6) integrated machine vision components at present, the machine vision system based on PC is not modularized structure, inconvenient installation, poor portability, particularly with PLC interface is more troublesome. From two aspects of software and hardware development, require a more integrated machine vision components for industrial demand. Currently COGNEX company has developed a modular component called vision sensors. This visual sensor integrated light source, camera, image processor, control and communication interface standard, since becoming an intelligent image acquisition and processing unit, the internal program memory can store image processing algorithms, and can use computer, use special configuration software of all kinds of algorithm, and then downloaded into the program memory of visual sensor. Vision sensors will be the flexibility of the PC, PLC reliability, distributed network technology together, with such a vision sensor and PLC can be more easily constitute a machine vision system.

 

2 the application of machine vision technology instance machine vision is an involves artificial intelligence, neurobiology, psychology, physics, computer science, image processing, pattern recognition and other fields of interdisciplinary. It is not only an extension of the human eye, more important is to have part of the brain function. There will be no one eye fatigue, machine vision has higher resolution than human eyes precision and speed, with the help of infrared, ultraviolet, X-ray, ultrasound and other advanced detection technology, it in detecting non visual object and high risk scenario, more has its prominent advantages [4]. Here are a few instances of machine vision applications.

 

(1) in the application of industrial and technical analysis of case 1: electronic plug parts stamping quality testing [5]

 

Electronic connectors in the stamping process quality inspection is the application of machine vision detection technology is the most typical example. The following chip pins will be stamping quality testing, for example, explain the principle of machine vision detection system, structure, operation, and analyzes the problems existing in the actual use. As shown in figure 4. Machine vision detection system:

 

(1) a pair of adjustable opening chute, to guide below the metal belt carrying the chip pins through the camera; (2) one or several digital cameras, installed above the metal belt; (3) a lamp lighting and a diffuse plexiglass, installation under the metal belt; (4) a set of photosensitive detector to detect the system to provide for the workpiece to the trigger signal.