No abstract

Acquiring clear images is a crucial precondition in many image-related applications, such as wireless sensor network, industrial inspection, and machine vision. In this paper, a multi-scale image adaptive enhancement algorithm for image sensors in wireless sensor networks based on non-subsampled shearlet transform is presented. The images are decomposed into different scales of coefficients. Then the coefficients are enhanced by a non-linear enhancement function. We set two thresholds for this function. One is used to classify the coefficients between the set to be denoised and the set to be enhanced; the other is to control the enhanced intensity of the coefficients. The thresholds are selected adaptively according to the decomposition scale and the standard deviation of the coefficients. The performance of the proposed algorithm is evaluated both objectively and subjectively. And the results show that the visibility of the images is enhanced significantly.

This paper describes a new type of image segmentation method based on deep convolutional neural networks (DCNN) in the actual autonomous driving scene. The spatial pyramid pooling model is used to identify and segment the actual scene to complete the machine-aware task. In order to improve the information aggregation of the whole image, we use atrous convolution for multi-scale feature extraction based on the pyramid structure of image cascade network (ICNet), removing a residual module in the fifth stage of the network, in order to reduce the scale of the convolutional layer. The feature map is preprocessed by padding and atrous convolution before the four-level spatial pyramid model. Then, conventional convolutions are introduced to compose the atrous spatial pyramid pooling (ASPP) structure. Finally, the four feature maps in the pyramid are merged with the feature maps before input into the pyramid. This paper analyzes the spatial pyramid model, receptive field, and dilation convolution in detail and propose atrous image cascade network (AtICNet). Experiment results in the cityscape dataset have shown that AtICNet has some improvements over ICNet, by improving the accuracy of the segmentation.

Heterogeneous image fusion is a technique of fusing images captured by different sensors into one image, then the fused image will present more information than the original images. This paper studies the compressive sensing image fusion algorithm and applies shearlet and wavelet transforms to represent the image sparsely. By compressing the sampled coefficients of the original images, the computational complexity in the image fusion process is reduced and the fusion efficiency is improved. We focus on the image fusion rules of compressive domain. Image coefficients of different frequencies are compressed by various sampling rates and fused according to different fusion rules. So an ideal fusion results can be obtained under a low sampling rate.

No abstract

No abstract

Acquiring clear images is a requisite in visual sensor networks. Image enhancement is an effective way to improve image quality. In this paper, non-sub-sampled shearlet transform (NSST) multi-scale analysis is combined with phase stretch transform (PST) to nonlinearly enhance the images captured by visual sensors. The components of different scales after NSST multi-scale decomposition are processed by nonlinear models with different thresholds. The thresholds of the enhanced model are determined by the local standard deviation of PST feature map. The noise is well suppressed, and the detail features are enhanced obviously. Experiments show that the proposed algorithm can improve image distortion, clear details, and enhance image contrast effectively.

Underwater acoustic modeling in shallow water environment is difficult since sound waves reflect several times between the surface and the water bottom. This article discusses an underwater acoustic characteristics analysis method based on self-similarity. It is found that acoustic signal has good self-similarity in shallow water. The actual towed hydrophone linear array was established and it was used for underwater acoustic signal acquisition experiment in Qilihai Reservoir which is located in the suburb of Tianjin, China. It can be derived that the signals acquired by hydrophones have self-similarity by the analysis of the variance of m-aggregated time series. It is proved that the characteristics of self-similarity can be used for the sound pulse propagation in shallow water.