The development history and development trend of machine vision inspection technology

Machine vision is a new technology developed in recent years. It uses the means of opto-mechanical integration to make the machine have the function of vision. Introducing machine vision into the field of inspection can realize online high-precision and high-speed measurement in many occasions. At the same time, the theory of machine vision detection technology has also grown and developed step by step, which can be divided into the following development processes and trends.

1. Primary Vision Theory: It is mainly aimed at the inverse problem of optical imaging. It consists of a series of processing procedures that can recover the physical properties of a three-dimensional visible surface from a two-dimensional light intensity array. The input data and calculation purpose of each process here can be clearly described, such as edge detection, stereo matching, and structure recovery from motion. In the process of projecting a three-dimensional object into a two-dimensional image, there is a lot of loss of three-dimensional information, which leads to ill-conditioned problems. Therefore, it is particularly important to strengthen the research on the primary visual process and its constraints, which are mainly aimed at 3D reconstruction and binocular vision.

2. Active vision theory: Active vision refers to the technical method in which the observer moves to track the target and perceive the object in a determined or indefinite manner. In active vision, the observer and the target object can also move at the same time. The movement of the observer provides additional conditions for studying the shape, distance and movement of the target. The important research directions are target tracking and missile interception.

3. Visual information fusion: Integrating multiple visual information with each other may break through the limitations of single visual information acquisition, achieve the acquisition of static and instantaneous visual information in an ideal environment, and meet the requirements of understanding the complex objective world. The main research areas are Image information fusion.

4. 3D scene reconstruction: At present, the recovery theory and algorithm of 3D scene are limited to the "visible" part of the scene, which belongs to the 2.5-dimensional information expression, and only provides the 3D information within the visible outline of the object. Recovering the complete information of the visible and invisible parts of the scene surface is a complex but urgent theoretical problem.

5. Algorithm performance evaluation: Machine vision research focuses on whether the task can be carried out or can be completed, and lacks the engraving and evaluation of the performance quality of algorithms and system methods. In practical applications, efficiency and performance are very important, otherwise the algorithms and systems cannot go out of the laboratory. Therefore, the establishment of machine vision algorithm performance evaluation is essential.

6. Visual parallel computing: There are still many theoretical, algorithmic and technical problems in visual real-time computing. The development trend of visual parallel computing structure is to use smaller and smaller processing units in larger and larger structures, and its development direction is to form a huge processing network system by basic logic operation processing units.

7. General visual information system: a general visual information system that can complete various visual tasks, that is, to establish a machine vision system that is analogous to the functions of the human visual system, and gradually develop into a complete general visual system by establishing a dedicated visual system platform, such as vision platform, highly intelligent vision robot, etc.