Study On HDR Adaptive Imaging Method For High-speed Scanning3D Measurement System

Structured light3D laser scanning has been studied over the past several decadesand is widely used in many industrial fields for dimensional analysis, real-timeinspection, component quality control and solid modelling. However, themeasurement of a3D profile of a high dynamic range (HDR) brightness surface suchas a partially highlighted object or a partial specular reflection remains one of themost challenging problems. This difficulty has limited the adoption of such scannersystems. The measuring speed, accuracy and resolution are a contradiction cube forthe traditional3D vision measurement method. In order to improve the speed, accuacyand resolution of the3D vision measurement system to enhance environmentaladaptability, a novel measuring priciple and mathematical model with high dynamicrange imaging have been proposed in this paper. The dynamic range of imagingsubsystem is expended by combination of optical attenuator (LCoS) and solid-stateimage sensor (CCD or CMOS) in optical space. The brightness adaptive imaging wasimplemented by an auto-adjust algorithm making the brightness to change in responseto the environment. An HDR image, reconstructed by the LCoS reflectivity mask, wasused to calibrate the3D laser scanner system for improving the accuracy.The major contribution and innovations of this dissertation are as follows:1. Because the image quality will affect the accuracy of vision measurementsystem, it is necessary to analysis the dynamic range of the imaging system how toinfluence the image quality. The principle and method of enhancing the dynamicerange of the imaging in optical space was studied. A method using optical attenuatorto extend the dynamic range of the imaging system for the3D vision measurement isproposed.2. The imaging subsystem with HDR enhancement scheme was designed bycombination of the optical attenuator and the image sensor. The dynamic rangeextension of imaging subsystem was developed by selecting the high-resolutionreflective Liquid Crystal on Silicon (LCoS) as an optical attenuator according to theproposed HDR extension theory in the optical space.3. The Pixel-level brightness value adjusted algorithm of the LCoS pixel mappingthe CCD pixel was designed through simplifying Direct Linear Transformation (DLT) camera model to obtain the transformation matrix. The calibration method betweenLCoS attenuation values and the CCD image brightness values of the high dynamicimage restoration was designed. Finally, the3D measurement system HDR adaptiveimaging algorithm was developed to automatically ensure the image sensor not oversaturation owing to the highlight of the surface material, specular reflection or thechanges of the measured environment.4. Using brightness Saturation Region Adjustement(SRA) mask instead of LCoSmask to simple the mathematical transformation, the mathematical thorey to restore3D HDR image using the SRA mask value and the CCD grayscale value was founded.The measurement accuracy was improved by using HDR image as the input of the3Dmeasuring image process system.5. Using Kalman-PID control algorithm to adaptive the HDR imaging process,the brightness automatic tracking of the capture image was preliminary implemented.