Hierarchically Adaptive Block Matching Algorithm And Application Under Complicated Scenes

Image block matching is one of active research fields in computer vision,which has been widely used in terget tracking,security monitoring and motion estimation.Due to great disparities in images of the same scene under various illumination and view,block matching under complicated scenes has been a challenging task.To this end,we improved matching criteria and searching strategy.The image structure information and correlation coefficient are combined as the matching criterion,adaptive stratification for the image and different searching strategies for different level.The proposed hierarchical block matching method which is adaptive to the computational complexity and applied to the field of target tracking.Simulation results show that,compared with the classical three steps,four steps searching algorithms and OHBM algorithm,this algorithm can reduce the time complexity efficiently.Moreover,since the algorithm makes use of the structure details of images,it is robust to complicated scenes.the proposed block matching method has better performances in the searching time and accuracy in contrast to the classic Mean-Shift target tracking algorithm and Particle Filter target tracking algorithm under the complex circumstances such as illumination variation and partial occlusion.The main contents as follows:Firstly,in order to reduce searching time,the whole algorithm is established in the framework of pyramid structures.By the coarse matching on the top level and the fine matching on the bottom level,the matching blocks can be accurately located and the speed of the algorithm is accelerated.Adaptive hierarchical image based on the number of image blocks.In addition,at different pyramid level,an adaptive searching strategy is adopted according to the number of candidate features.Secondly,Except the improvement of search strategy,we also develop a fast matching algorithm based on image blocks.The method does not use a single matching criterion such as MAD or MSE.The correlation coefficient and structural functions are adopted to evaluate the similarity between two images.At the same time,the size of the image block is adjusted according to the change of the matching image angle,and the size of the image block is increased when the visual angle is changed,so as to ensure that there are enough image details for matching.Lastly,traditional algorithms often track the object through adjacent frames,whereas our algorithm will search two frames with a variable interval on the basis of matching results.The intervals are obtained by using the bisection method and thus our method has the higher accuracy and the lower time complexity.