The classic super-resolution algorithm via sparse coding has high computational cost during the reconstruction phase. In view of the disadvantages, a predictive sparse coding-based single image super-resolution method was proposed. In the training phase, the proposed method imposed a code prediction error term to the traditional sparse coding error function, and used an alternating minimization procedure to minimize the resultant objective function. In the testing phase, the reconstruction coefficient could be estimated by simply multiplying the low-dimensional image patch with the low-dimensional dictionary, without any need to solve sparse regression problems. The experimental results demonstrate that, compared with the classic single image super-resolution algorithm via sparse coding, the proposed method is able to significantly reduce the reconstruction time while maintaining super-resolution visual effect.

A requirement modeling language called SPARDL was proposed for modeling and analyzing such periodic control systems consisting of periodic behaviors together with the mode transition mechanism. The Event-B interpretation was specified for the SPARDL model. The semantics of SPARDL were presented by Event-B and a refinement framework was introduced to develop the Event-B models based on the features of the SPARDL model. Finally, a case study was analyzed to show the effectiveness of our proposed approach to modeling and validation of the SPARDL model by Event-B.

An effective registration approach was proposed for multiple-view images captured on a turntable. In combination with binocular stereo, a complete system for 3D reconstruction was proposed. The calibration target was mounted on a step-motor-controlled turntable and it was acquired under multiple views. The coordinate transformation between the turntable and the camera coordinate systems was computed by the angular point information on the calibration target. The coordinate transformation among multiple views was used for data registration at the fixed viewpoint. The experimental results show that the proposed registration method achieves high precision and can be effectively used in 3D reconstruction.