Method and device for visually identifying artificial intelligent machine under common and special environments

Abstract

The invention relates to a method and a device for visually identifying an artificial intelligent machine under common and special environments. The invention aims to solve the problem that the prior art is not suitable for multi-field measurement. A light output system is composed of a laser (1), a beam shaping system (2), a rectangular raster (3), a monochrome Fourier transform lens (4), a frequency spectrum selector (5), a wave plate (6), a monochrome multiplying power lens (7) and a high-speed scan mirror (9) which are successively arranged on a precision lead screw (8-1) and a precision guide rail (8-2) in turn along the light path; the light output system is utilized to project sine structured light; a visual collecting identifying system is used for visually collecting and identifying a to-be-identified target (10); the visual collecting identifying system is composed of a computer (14), a telecentric lens connected with the computer (14), a high speed CCD or CMOS (11), a laser ranging system (12) and an internal environment control system (13); a protecting shell (15), in which the light output system and the visual collecting identifying system are arranged, is provided with a radiation protection window used for performing radiation protection on the light output system and the visual collecting identifying system. The device is specifically fit for the continuous output of sine structured light under large-field range measurement from the near to the distant and has the advantage of mode recognition for static, low-speed and high-speed moving objects at different distances and with different characteristics.

Description

technical field [0001] The invention relates to a machine vision recognition method, in particular to an artificial intelligence machine vision recognition method and device for common and special environments. Background technique [0002] At present, artificial intelligence machine vision recognition methods mainly realize pattern recognition by directly shooting images and combining graphic image recognition algorithms. For specific objects, specific software algorithms are used for identification; such methods and technologies are also mainly applied to specific and single identification objects. For example: integrity recognition of tablets and capsules in medicine; liquid level recognition of plastic bottles, beverages, alcohol, etc.; edge detection and recognition fields such as image edge integrity recognition in the printing and dyeing industry; others such as: ports, airports, railway stations, Fields such as import and export goods detection generally require man...

AI Technical Summary

Problems solved by technology

This type of method is slow because of the existence of the scanning structure; or, because of the use of general-purpose equipment, such as a projector, it is not suitable for special environments such as high radiation, high temperature, and high pressure; or, the computational load of dual cameras is relatively large. ; or, because of the use of holographic interferometry, it is only suitable for laboratories, limiting the use of the environment

Images