Electronic Image Stabilization Method Based On Block Matching And Random Ferns

Compared with traditional image stabilization method, the electronic imagestabilization method has many advantages, such as small-size, easy-operation, highaccuracy, reasonable price and low power consumption. It is widely used in military andcivilian and the study of the electronic image stabilization is more and more importantwith image stabilization products come out.At the first the paper describes the significance and necessity of the study ofelectronic image stabilization and the research situation at home and abroad. Thenelaborate on the basic principle of the electronic image stabilization. And motion model ofthe camera and the image motion model is introduced and analyzed. Focus on the analysisand the applicable conditions of several classic core algorithms (motion vector estimation,motion decided, motion compensation) of the electronic image stabilization.There isn’t any electronic image stabilization method can be applied to any scene’video, so the paper designs two electronic image stabilization systems for the two situationwhat are the fixed and moving cameras. Use a suitable motion estimation and motioncompensation algorithm to the different video specifically. Select a large viewing angleimage as the reference frame and adopt the random ferns classification matching andblock matching for the fixed camera to compensate motion vector directly. Design adynamic cache as reference frame and use block matching and feature matching to getmotion vector for the motion decision and compensation using Kalman filtering anddamping. Finally, get the stable video, analysis and evaluate the results of imagestabilization objectively.At last, use the electronic image stabilization system to process and show the videowhich is obtained from an industrial camera connected with the computer. It is to analysisthe real-time performance in the practical application.