The Research Of Highlight Inhibition Methods Based On Linear Diffuser

3D reconstruction technique which is based on machine vision has a bright application prospect in fields like production automation,military defense,medical diagnosis,and digitization of cultural relics.As one of the approaches for machine vision measurement,the encoded structured light is able to obtain the 3D information of the measured object fast and accurately.Also,it is able to reconstruct the measured object according to the obtained 3D information.However,in the process of the 3D measurement,it will cause information loss because of the highlight generated by the strongly reflected surface of the measured object,which makes it hard to have an accurate reconstruction result.Therefore,it is of great scientific significance for structured light 3D measurement field to study on effective highlight inhibition approaches to accurately reconstruct the strongly reflected surface.In this paper,we study on the existing highlight inhibiting approaches,and focused on the problems of the existing approaches,we proposed a highlight inhibiting approach based on the linear diffuser.The research of the paper can be summarized as follows:1.The diffuser highlight inhibiting model based on the radiation is established.Primarily,we analyze the difference between the traditional diffuser and the microstructured diffuser according to the light path.Moreover,we analyze the light transmittance and uniformity of difference types of micro-structured diffusers,and choose the diffuser with excellent property.Then,we study on the relevant theories of optical radiometry such as the solid angle,the radiance,and the bidirectional reflectance distribution function.According to the theories,we build the diffuser highlight inhibiting model based on the radiation,and obtain the radiation values of a certain highlight point with and without the linear diffuser system.2.Based on the theory of structured light 3D measurement,we design the structured light measurement system for the strongly reflected surface.The design of the system includes the establishment of the experimental platform,the selection of the encoded pattern,and the calibration of the camera and projector.3.With the proposed approach,we implement the 3D reconstruction experiment on 3 strongly reflected objects with different shapes and smoothness.Then,we compare the 3D reconstructed image with the image captured by the camera,and give a quantitative analysis on the number of the highlight pixels in the reconstructed image.According to the experimental results,we can get that the highlight region in the reconstructed image is obviously narrowed when the diffuser is equipped.The highlight pixels of the 3 objects are respectively reduced by 37.4 times,3.83 times and 2.98 times,which proves that the proposed approach is able to restore the texture information of the object effectively.