Deep learning (DP) based road extraction methods often produce fragmented results. Direct optimization of end-to-end DP-based methods requires the design of more complex network structures to enhance the model’s adaptability in complex scenarios. To tackle this challenge, this paper adopts a novel approach that treats the discrepancy (the road breakage part) between road prediction and ground-truth as the extraction target, thereby constructing a highly efficient and lightweight semantic segmentation network, termed the Road Patch Extraction Network (RPE-Net). RPE-Net includes multi-directional striped residual (MDSR) encoder, multi-directional striped pooling (MDSP) units, and multi-directional striped decoder (MDSD). The structure is similar to LinkNet34, but the overall size of the network parameters is just 1.56MB, which is 1/50 of LinkNet34. The post-processing datasets used for training can be semi-automatically generated by the algorithm, and only a small amount of manual intervention can be used for training. A large number of experiments showt hat the post-processing method proposed in this paper has extremely high speed and generalization ability.

This paper investigates the problem of building extraction from very high resolution (VHR) satellite imagery. Deep learning methods are deemed as emerging trends for solving this problem due to their increasingly prominent extraction effects. However, most state-of-the-art deep learning-based building segmentation methods produce pixel-level segmentation masks rather than accurate polygon-level extraction results required in real-world applications. This paper introduces a Harris function based active contour network (HACNet) that extracts polygon-level masks directly from satellite imagery. Our proposed HACNet not only leverages the full potential of active contour methods, but also successfully integrates the Harris function into the polygon contour evolution of the active contour models. The incorporation of the Harris function concentrates the attention of the neural network more on the vicinity of corners, thereby significantly improving the performance of building contour extraction, particularly in scenes with high curvature and noisy boundaries. Experiments on two public datasets (namely Vaihingen, Bing Huts) demonstrate the effectiveness and superiority of our model over other outstanding deep learning methods for the building extraction.

Using deep learning to extract roads from satellite images is one of the most popular methods. However, the existing encoder-decoder-based deep networks usually produce fragmented roads, due to the complex spatial and color characteristics of the road. In this paper, motivated by the road multi-scale information, we proposed a multi-scale and multi-direction feature fusion network (MSMDFF-Net) to reduce the fragmentation of road extraction results. The proposed method mainly consists of three processes: 1) In the initial stage, the image details from different directions were transmitted; 2) At different encoding stages, the multi-scale information of the image was fused; 3) In the decoding process, the matching modules of road characteristics were used to up-sample the feature map. Extensive experiments on the popular datasets (LSVD and Deep-Globe datasets) demonstrate that the MSMDFF-Net has higher accuracy and generalization performance with less fragmentary road results.

In order to make the best use of the available data in the remote sensing database, this paper focuses on an important topic of using the complementary information of multi-source remote sensing data, that is, road network extraction based on fusion technology with synthetic aperture radar (SAR) and optical images. Starting with the line segments achieved from the road segmentation maps, a decision-level fusion method which mainly includes two stages is proposed in this paper. The first stage is fusing based on the geometric overlapping rules. In the second stage, a road network extraction approach that takes into account both the contextual information and evidence theory is presented. The experiments on TerraSAR-X and WorldView-4 images showed that our proposed method had an excellent performance in terms of the completeness and quality of the road extraction.

In order to speed up the update of existing road maps, it is crucial to develop a more efficient road extraction method from remote sensing images. In recent years, deep learning techniques have been widely used for road extraction applications. Among current CNN-based deep networks for road extraction, few works study the shape of the convolution kernel, and the remote contextual features dependency relationship is not fully utilized. In view of these problems, an improved DLinkNet is proposed in this paper. Firstly, a convolutional layer which fuses information of multiple scales is used to replace the InitBlock in the front of the network. Secondly, instead of using the D-Block, the DenseRe-Block is applied to the center structure of DLinkNet. The experimental results show that the improved network has a higher IoU score than DLinkNet when extracting roads from optical images in both city and mountain town areas.

Atmospheric fine particulate matter (PM2.5) pollution has brought a strong focus on public health and environmental quality in China because of its adverse effects. To evaluate the effect of technological and social development on environmental quality in China, it's in urgent need of understanding the long-term spatiotemporal trend of PM2.5 concentrations. In this paper, satellite-derived annual mean PM2.5 estimates (1998-2016) were validated using ground-based PM2.5 measurements (2015-2016) and then used for spatiotemporal trend analysis. The results indicated that national mean PM2.5 concentrations in China increased primarily before 2008, and then decreased. The spatial distribution of PM2.5 concentration is high level in the east and while low in the west of Heihe-Tengchong Line in China. Our findings provided a profound understanding of PM2.5 variations and affirmed the effectiveness of implemented measures of reducing PM2.5 loadings in China, offering important reference data for relevant policy making of air pollution prevention in the future.

Block sizes of practical vector quantization (VQ) image coders are not large enough to exploit all high-order statistical dependencies among pixels. Therefore, adaptive entropy coding of VQ indexes via statistical context modeling can significantly reduce the bit rate of VQ coders for given distortion. Address VQ was a pioneer work in this direction. In this paper we develop a framework of conditional entropy coding of VQ indexes (CECOVI) based on a simple Bayesian-type method of estimating probabilities conditioned on causal contexts, CECOVI is conceptually cleaner and algorithmically more efficient than address VQ, with address-VQ technique being its special case. It reduces the bit rate of address VQ by more than 20% for the same distortion, and does so at only a tiny fraction of address VQ's computational cost.

Vector quantization (VQ) is a source coding methodology with provable rate-distortion optimality. However, despite more than two decades of intensive research, VQ theoretical promise is yet to be fully realized in image compression practice. Restricted by high VQ complexity in dimensions and due to high-order sample correlations in images, block sizes of practical VQ image coders are hardly large enough to achieve the rate-distortion optimality. Among the large number of VQ variants in the literature, a technique called address VQ (A-VQ) by Nasrabadi and Feng (1990) achieved the best rate-distortion performance so far to the best of our knowledge. The essence of A-VQ is to effectively increase VQ dimensions by a lossless coding of a group of 16-dimensional VQ codewords that are spatially adjacent. From a different perspective, we can consider a signal source that is coded by memoryless basic VQ to be just another signal source whose samples are the indices of the memoryless VQ codewords, and then induce the problem of lossless compression of the VQ-coded source. If the memoryless VQ is not rate-distortion optimal (often the case in practice), then there must exist hidden structures between the samples of VQ-coded source (VQ codewords). Therefore, an alternative way of approaching the rate-distortion optimality is to model and utilize these inter-codewords structures or correlations by context modeling and conditional entropy coding of VQ indexes.

In order to obtain the high-fidelity medical compressed images, a new compression scheme is proposed. Based on the stringent requirements on lossy medical image compression, we refine the context modeling for a given class of medical images and utilize the conditional entropy coding of the VQ index (CECOVI) scheme to code the MR head images. The experimental results show that the image-type-dependent CECOVI can achieve better rate-distortion performance than the state-of-art wavelet image coder SPIHT. This also implies that incorporating the conditional entropy coding strategy into the VQ process is an appropriate way for high-fidelity medical image compression.