TY - THES

T1 - Investigation of two colour photometric stereo for the surface inspection of steel bars

AU - Chen, Yuan Fei

PY - 2005/10/14

Y1 - 2005/10/14

N2 - This thesis presents the optical non-contact surface measurement method, termed "Two Colour Photometric Stereo", along with its calibration techniques. The principle is to take two images of an object from the same camera view, while illuminating from two known directions respectively with a blue and a green light source. Based on the intensity values in the two different images, one can estimate the surface normal vectors and surface albedo. The relative position of the light source relative to the camera are determined from a calibration process. During the calibration a target is observed which enables the determination of the calibration matrix. With the calibration matrix, one can compute a measured pixel intensity. Whereby the intensity is corrected to eliminate the effects of nonuniform illumination. From the corrected images, one can obtain the reconstructed albedo and surface normals. The technique has applications in automatic inspection for quality control of a wide range of manufactured and processed surfaces, it uses a low-cost fixed lighting configuration, requires no initial training, and is largely pose invariant. Experimental results are presented in the thesis.

AB - This thesis presents the optical non-contact surface measurement method, termed "Two Colour Photometric Stereo", along with its calibration techniques. The principle is to take two images of an object from the same camera view, while illuminating from two known directions respectively with a blue and a green light source. Based on the intensity values in the two different images, one can estimate the surface normal vectors and surface albedo. The relative position of the light source relative to the camera are determined from a calibration process. During the calibration a target is observed which enables the determination of the calibration matrix. With the calibration matrix, one can compute a measured pixel intensity. Whereby the intensity is corrected to eliminate the effects of nonuniform illumination. From the corrected images, one can obtain the reconstructed albedo and surface normals. The technique has applications in automatic inspection for quality control of a wide range of manufactured and processed surfaces, it uses a low-cost fixed lighting configuration, requires no initial training, and is largely pose invariant. Experimental results are presented in the thesis.

KW - Zwei Farben Photometric Stereo

KW - Oberflächennormale

KW - Rückstrahlverhalten

KW - Two colour Photometric Stereo

KW - Surface Normal

KW - Albedo

M3 - Diploma Thesis

ER -