Search

A two-stage denoising algorithm based on local similarity is proposed to process lowly and moderate corrupted images with random-valued impulse noise in this paper. In the noise detection stage, the pixel to be detected is centered and the local similarity between the pixel and each pixel in its neighborhood is calculated, which can be used as the probability that the pixel is noise. By obtaining the local similarity of each pixel in the image and setting an appropriate threshold, the noise pixels and clean pixels in the damaged image can be detected. In the image restoration stage, an improved bilateral filter based on local similarity and geometric distance is designed. The pixel detected as noise in the first stage is filtered and the new intensity value is the weighted average of all pixel intensities in its neighborhood. A large number of experiments have been conducted on different test images and the results show that compared with the mainstream denoising algorithms, the proposed method can detect and filter out the random-value impulse noise in the image more effectively and faster, while better retaining the edges and other details of the image.

Remote sensing images have been widely applied in various industries; nevertheless, the resolution of such images is relatively low. Panchromatic sharpening (pan-sharpening) is a research focus in the image fusion domain of remote sensing. Pan-sharpening is used to generate high-resolution multispectral (HRMS) images making full use of low-resolution multispectral (LRMS) images and panchromatic (PAN) images. Traditional pan-sharpening has the problems of spectral distortion, ringing effect, and low resolution. The convolutional neural network (CNN) is gradually applied to pan-sharpening. Aiming at the aforementioned problems, we propose a distributed fusion framework based on residual CNN (RCNN), namely, RDFNet, which realizes the data fusion of three channels. It can make the most of the spectral information and spatial information of LRMS and PAN images. The proposed fusion network employs a distributed fusion architecture to make the best of the fusion outcome of the previous step in the fusion channel, so that the subsequent fusion acquires much more spectral and spatial information. Moreover, two feature extraction channels are used to extract the features of MS and PAN images respectively, using the residual module, and features of different scales are used for the fusion channel. In this way, spectral distortion and spatial information loss are reduced. Employing data from four different satellites to compare the proposed RDFNet, the results of the experiment show that the proposed RDFNet has superior performance in improving spatial resolution and preserving spectral information, and has good robustness and generalization in improving the fusion quality.

Nowadays, intelligent surveillance has received extensive attention from academia, business, and industry. Deep learning algorithms are widely used in the field of intelligent surveillance. Recently, most deep learning models are limited to a short-term behavior recognition in the entire video. In order to better identify human behavior in the video, we combined a Two-stream network and a Temporal Relation network (TRN) and added a time pyramid pooling operation. In this way, the Two-Stream Temporal Relation-Time Pyramid Pooling Network (TTR-TPPN) can be constructed. The relational pyramid pool network integrated the frame-level features in the video into video-level features. We applied the TTR-TPPN to the Internet public standard data set UCF101 and the self-made DW20 data set. It is found through experiments that this network has a higher recognition rate than other behavior recognition methods on both data sets, and it has better performance in long-term behavior recognition. Therefore, the TTR-TPPN enables it to recognize long-time sequence behavior and improves the accuracy of human behavior recognition.

In the process of image generation or transmission, the image quality is often degraded due to the interference and influence of speckle noise, which will adversely affect the subsequent image processing. However, most of the existing Optical Coherence Tomography (OCT) image denoising methods usually only use part of the prior information of the OCT image, and ignore the changes in the texture and structure of the OCT image. There is often a problem that the network is too deep and the calculation complexity is too large. Aiming at this shortcoming, this paper proposes an improved deep convolutional neural network image denoising algorithm based on the original DenseNet algorithm idea. First, by constructing a speckle noise image data set, a series of preprocessing is performed on the input image data set, then the visual statistical feature map of the image is extracted, and finally the DenseNet network structure is improved for network training. The experimental results show that the improved DenseNet has better performance no matter compared with BM3D, which is recognized as the best denoising algorithm in the field of image denoising, or compared with DnCNN, an advanced image denoising algorithm in the field of deep learning.