Research On Computational Imaging And 3D Scanning Measurement Based On Digital Micromirror Device

Structured light 3D scanning measurement technology has been widely used in advanced manufacturing fields.However,the application of the structured light 3D scanning measurement technology is also limited or imperfect in some occasions,such as the larger curvature of surface,convex and complex shape and strong reflection surface and so on.Therefore,structured light 3D scanning measurement technology needs to be further developed and provided new means.In recent years,the emerging computational imaging techniques have creatively introduced computations into all aspects of the imaging process,and optimized the coupling and collecting reconstructed of optical signals at maximum freedom to gain more ideal visual images.According to the characteristics of digital micromirror device(DMD)with all-digital modulation of incident light intensity and high-speed state deflection,this paper introduces the computational imaging methods into 3D scanning measurement,and designs a novel computational imaging and measurement system based on DMD that is used to improve the dynamic range and time resolution of the measurement system.This new measurement method extends the application range of structured light 3D scanning measurement technology,especially for solving the visual imaging and 3D measurement problems of strong reflection surface.The major content and innovations in this papers are shown as follows:(1)The first generation and second generation computational imaging system based on DMD are designed and built.The imaging system can flexibly modulate the spatial information and temproal information of the incident light,which will provide the hardware foundation for subsequent research.(2)The key common technologies of computational imaging system based on DMD are studied.The pixel-level matching between DMD micromirrors and image detector pixels is realized by moiré fringe method.Direct linear transformation mapping algorithm,nonlinear polynomial distortion fitting mapping algorithm and new mapping algorithm based on BP neural network are designed to realize the coordinate mappings between DMD micormirror array and the image sensor pixel array.The response curve of DMD modulation and CMOS image detector are calibrated by using mathematical analysis method,then the response curve calibration method of computational imaging system based on DMD is also proposed,which provides an important basis for the adaptive adjustment of the DMD mask.(3)The adaptive high dynamic range imaging methods based on DMD are studied.According to the structural characteristics of the imaging system,an optical model of imaging system based on DMD is established,and the theory of per-pixel coded exposure is put forward.By using the DMD modulation characteristics with spatial information of incident light,two different adaptive high dynamic range imaging methods are designed,and the methods achieve adaptive high dynamic range imaging of strong reflection surfaces with different materials.(4)A high-speed and high-resolution imaging method based on time-domain pixel multiplexing is designed.Based on the modulated characteristics of DMD with temproal information of incident light,the principle of time-domain pixel multiplexing and the clock control algorithm of imaging system based on DMD are proposed,and a three-element median sorting algorithm is designed to realize the sorting and combination of exposure elements.The multiple high-resolution subframes are orderly extracted from the encoded image,which greatly improves the temporal resolution of the imaging system.(5)A prototype of computational imaging and 3D scanning measurement system based on DMD is developed.Combined with adaptive high dynamic range imaging methods based on DMD and 3D scanning measurement technique used grating projection,the novel prototype of scanning measurement has been realized the 3D measurement of strong reflection surface.