Industrial automation has promoted the rapid development of machine vision technology. Because the structure of the machine vision system is relatively complex, most people do not know it well, but the basic model architecture is the same. According to the flow sequence of signals, it mainly includes the following modules:
The modules of the machine vision system can be divided into two parts: lighting system planning and the planning of lens optical system.
Lighting planning is to reasonably plan the intensity, color, uniformity, structure, and size of the light source by studying the optical characteristics, distance, object size, and scenery characteristics of the measured object, and plan a reasonable light path to reach the intention of obtaining guidance-related structural information.
The lens is the main component that projects the space information of object to the image space. The adjustment of the lens is mainly to select the focal length and aperture scale of the lens according to the detected lighting conditions and characteristics of guidance. After determining the type of lens, adjust the rear fixed structure of the lens.
The image sensor module of the machine vision system is mainly responsible for the photoelectric conversion of information and is located on the image plane at the back of the lens. At present, the main image sensor can be divided into CCD (Charge-coupled Device) and CMOS image sensor. Smooth and excellent circuit drive is the key to this module.
The image processing module of the machine vision system is mainly responsible for the processing of images and the proposal of information parameters, which can be divided into two levels: hardware structure and software algorithm.
The hardware is generally a circuit system centered on the CPU. The PC-based machine vision makes use of the PC's CPU and related peripherals; the machine vision based on the embedded system, such as the smart camera with independent data processing ability depends on the information processing chip on the board such as DSP, ARM, FPGA.
The software part includes a complete image processing plan and decision plan, including a series of algorithms. In the high-end image system, the data algorithm library will be integrated to facilitate the transplantation and reuse of the system. When the algorithm library is large, the algorithm library can be called through the graphical interface.
The IO module is a module that outputs the calculation results and data of the telecentric lens machine vision system. According to the PC's machine vision system, the interface can be divided into internal interface and external interface. The internal interface only needs to serve as the high-speed communication port of the system to transmit signals to the PC. The external interface completes the function of communication and information communication between the system and other systems or users. Smart cameras generally use general-purpose IO and high-speed Ethernet to complete all the corresponding functions.
The flash module can be regarded as a special user IO, and it can use a more intuitive detection system in the work process. In terms of the PC-based machine vision system, the data information of the system can be directly transmitted to the graphics card through the PCI bus, and to the computer screen through the VGA interface. The independent processing smart camera generally completes the visualization of the image by expanding the LCD screen and the image flashing control chip.