Research On The System Of Collimation Measuring Between Laser Head And Disk For Optical Storage Device By Means Of Facula Image Analysis

Technologies to ensure the correct recording and reading the optical information, to guarantee the integrality and reliability of the recording and reading data, and to duplicate optic disks are three key technologies for optical storage systems. One of the preconditions to achieve technologies mentioned above is to make the laser head-disk in the state of kilter perpendicularity. To develop a technical frame of laser head-disk obliquity measuring for optical storage system is the main target of this thesis. Facula images are sampled when with glass disk to modulate the laser head light path and without the glass disk's modulation to get two kinds of facula as circle and diffraction shapes. By means of measuring the displacement of the center points of these faculae the obliquity would be calculated. The systematic frame is designed as well some key technologies for it are studied further.Principal component analysis as the branch of multi-variant statistics analysis, digital morphological hit-and-miss transformation (HTM), thin processing, Lee's image enhancement operator, Hough transformation, et al are main tools used in the study work. Main work formulated the thesis including aspects as following.Laser head-disk obliquity measuring model was developed by means of the calculation of the displacement of the faculae. The laser head and its facula of optic storage system were reviewed and analyzed. Formulations for the displacement calculation of laser head's center lights and the center points of faculae were reduced based on geometry light tracing. Comparative analysis showed that the latter would be more rational and feasible for laser head-disk obliquity measuring.The concept of abnormal pixel was defined while the identification model for abnormal pixels in images frames was developed. The residual matrix was obtained by subtracting the DC component from each frame; the eigen-images for the original frames were developed using principal component analysis to the residual matrix. Mahalanobis distance was defined for pixels and hence the Mahalanobis distance of the original pixels could be expressed by the eigen-images. Finally the identification model for abnormal pixels was developed by the eigen-images and Mahalanobis distance. Blind Gaussian noise estimation method was developed. By experiments the neighbor widow scale for Lee's image denoising enhancement operator was determined.The scale of the neighbor widow should be determined to complete Lee's denoising algorithm. Two criterions as Signal-Noise Ratio and Edge Strength Ratio were developed and used to determine the scale of the neighbor window.Hybrid algorithm for center point of circle facula image was developed. The main contain of the algorithm is briefed as following. Segmenting the circle facula image with a special grayscale threshold. And the segmented image will be worked as the input image for the next step. Digital morphological algorithms can detect the central region and edge of the circle facula. With the basic idea of Hough transformation as the duality between parameter space and geometry space, the center pixel can be fitted in the center region with facula's edge detected.Technologies mentioned and described above were combined to formula a frame for the collimation measuring between the laser head and optic disk. The whole system should including facula image processing subsystem, laser head-CCD camera collimating subsystem, and laser head-disk collimating subsystem. The technical process diagrams were given of the above subsystemsm, while the systematic error be likely to introduce of the subsystems was described.