Wavefront Coding System Based On The DM642Digital Signal Processor

Wavefront coding system is a two-step imaging system including optical encoding and digital decoding, can achieve deep depth of field that is impossible for traditional optical imaging systems and improve the imaging quality. A specific designed phase mask is positioned on the pupil plane, in the image plane a defocus-insentitive but blurred image is acquired, then the final sharp image can be given by using the digital image restoration algorithms. This technology will be widely used in optical imaging systems for either military or civil use.We realized the wavefront coding system on the embedded processing platform for the first time. Based on the process of the image convolution in spatial domain, combined anti-reflective boundary conditions and Bi-CGSTAB(BiConjugate Gradient Square Stable Methods) image restoration algorithm, we realized wavefront coding system on the platform of DM642as the core of system. The experiment shows the effect of system based on the platform of DM642is sharp, fast and stable. This gives the possibility to realization of the truly portable and practical of the wavefront coding system.This thesis mainly discussed in circuit design and algorithm realization.Firstly, we design and realize the circuit of wavfront coding system. The circuit system includes DSP core module, peripheral storage circuit module, video capture circuit module, video display circuit module, power management and timing circuit module. The important works we completed are list below. We complete the circuit design of core of system with media digital signal processor from Texas Instruments; We extend the meomory space by FLASH and SDRAM; We complete the video capture circuit design with CMOS sensor; We complete the video display circuit design with DVI transmitter from Texas Instruments; We complete the power supply circuit by DC-DC; We complete the power management and timing circuit by using supervisor and timing management chips; We complete the6layer high speed PCB layout design.Secondly, we discussed the Bi-CGSTAB algorithm with anti-reflective boundary conditions in detail. As an algorithm based on the krylov subspace, Bi-CGSTAB is suitable for solving non-symmetric matrix equation. Compared with traditional image restoration algorithm, Bi-CGSTAB could reduce the computation and have better restoration results. Effectiveness and speediness of the proposed method was demonstrated by experiment results, which can give excellent deconvolution and noise suppression restored image and reduce the boundary ringing effect under the anti-reflective boundary conditions.