Traditional trackers are easily affected by uncertain changes in tracking targets, such as occlusion, deformation, and background clutter. To solve these problems, we propose a tracking method, namely location-matching tracking under a convolutional neural network (CNN), which consists of a process of localization, recognition, and model updating. In the location subprocess, the target’s locations of the previous (first) frame and the current frame are utilized to estimate a series of specific regions by the average displacement, and these locations are proven to be useful to improve the probability of a successful tracking. In the recognition subprocess, a CNN is adopted to classify the estimated regions, and we calculate the confidence score maps of these regions to estimate the final target region. To improve the accuracy of the tracking, we propose an optimal similarity matching to verify the final target region and make a confidence decision to update the network. Compared with the state-of-the-art trackers on challenging object tracking benchmark benchmarks, the proposed method can achieve the same or even higher tracking accuracy.
To solve the problems of fuzzy details, color distortion, low brightness of the image obtained by the dark channel prior defog algorithm, an image defog algorithm based on open close filter and gradient domain recursive bilateral filter, referred to as OCRBF, was put forward. The algorithm named OCRBF firstly makes use of weighted quad tree to obtain more accurate the global atmospheric value, then exploits multiple-structure element morphological open and close filter towards the minimum channel map to obtain a rough scattering map by dark channel prior, makes use of variogram to correct the transmittance map,and uses gradient domain recursive bilateral filter for the smooth operation, finally gets recovery images by image degradation model, and makes contrast adjustment to get bright, clear and no fog image. A large number of experimental results show that the proposed defog method in this paper can be good to remove the fog , recover color and definition of the fog image containing close range image, image perspective, the image including the bright areas very well, compared with other image defog algorithms,obtain more clear and natural fog free images with details of higher visibility, what’s more, the relationship between the time complexity of SIDA algorithm and the number of image pixels is a linear correlation.