Detecting changes in land use and land cover (LULC) from space has long been the main goal of satellite remote sensing (RS), yet the existing and available algorithms for cloud classification are not reliable enough to attain this goal in an automated fashion. Clouds are very strong optical signals that dominate the results of change detection if they are not removed completely from imagery. As various architectures of deep learning (DL) have been proposed and advanced quickly, their potential in perceptual tasks has been widely accepted and successfully applied to many fields. A comprehensive survey of DL in RS has been reviewed, and the RS community has been suggested to be leading researchers in DL. Based on deep residual learning, semantic image segmentation, and the concept of atrous convolution, we propose a new DL architecture, named CloudNet, with an enhanced capability of feature extraction for classifying cloud and haze from Sentinel-2 imagery, with the intention of supporting automatic change detection in LULC. To ensure the quality of the training dataset, scene classification maps of Taiwan processed by Sen2cor were visually examined and edited, resulting in a total of 12,769 sub-images with a standard size of 224 × 224 pixels, cut from the Sen2cor-corrected images and compiled in a trainset. The data augmentation technique enabled CloudNet to have stable cirrus identification capability without extensive training data. Compared to the traditional method and other DL methods, CloudNet had higher accuracy in cloud and haze classification, as well as better performance in cirrus cloud recognition. CloudNet will be incorporated into the Open Access Satellite Image Service to facilitate change detection by using Sentinel-2 imagery on a regular and automatic basis.
16 December 2013 in Journal of Visual Communication and Image Representation
Defogging is an essential preprocessing technique for object detection in computer vision-based systems and has been widely used in outdoor surveillance system applications. This paper proposes an efficient defogging algorithm for both static images and videos. Considering real-time applications, the proposed defogging algorithm involves a low-cost hardware oriented design that is based on an atmospheric scattering model and dark channel prior. Compared with previous low-complexity techniques, simulation results indicated that the proposed design demonstrated superior performance in terms of quantitative and qualitative evaluations. A weighting technique and a contour preserving estimation approach are adopted alternately to refine the factors in the defogging process. Furthermore, in the atmospheric light estimation, an adjuster is applied to the video for preventing “flicker” which means that the brightness changes dramatically between two neighbor frames in the video. Hence, the proposed algorithm is suitable for video defogging applications, which have not been dealt with in previous approaches. To achieve the requirement of real-time applications for both static and dynamic images, an implementation of seven-stage very-large-scale integration (VLSI) architecture for the proposed algorithm is presented. By using TSMC 0.13-um technology, the design yielded a processing rate of approximately 200 Mpixels/s.
This work aims at optimizing the hardware implementation of the SubBytes and inverse SubBytes operations in the advanced encryption standard (AES). To this, the composite field arithmetic (CFA) is employed to optimize all building blocks in S-box (and inverse S-box) of SubBytes (and inverse SubBytes) transformation. A joint design of S-box and inverse S-box is also proposed to further enhance the area efficiency. Specifically, the area of multiplier in the Galois composite field, GF $((2^{2})^{2})$ , is reduced. The squaring and multiplication with constant $\lambda $ in GF $((2^{2})^{2})$ are combined and optimized as well. Moreover, the multiplicative inversion in GF $((2^{2})^{2})$ is manually optimized. Furthermore, the S-box and inverse S-box are combined and optimized using the pre_processing and post_processing modules. To increase the throughput, a balanced and pipelined architecture is derived. Using the proposed architecture, a throughput of 5.79 Gbps for the S-box can be achieved on Virtex-6 XC6VLX240T and 10% better than the conventional work. According to the ASIC implementation result, the proposed design can still achieve the highest area efficiency and approximately 30% better than conventional works using TSMC 90nm process.
In the real-time outdoor application of computer vision-based systems, haze removal, a pre-processing technique that can recover clear images from foggy ones, is necessary for object detection. There-fore, in this paper, an efficient haze removal method suitable for hardware design is proposed to obtain high quality fog-free images. Based on the atmosphere scattering model and the dark channel prior method, the atmospheric light of the whole image and the transmission map could be extracted. These are important and necessary parameters of the recovery model. Instead of using single global atmospheric light to restore foggy image, a local atmospheric light estimation method is applied in the proposed design to achieve optimal results. To ensure that the overall image is consistent without block artifacts, dynamic adjustment of local atmospheric light is made based on global atmospheric light. Additionally, to obtain a transmission map, a refined estimation method is performed to ease the halo effect. In terms of both quantitative and qualitative evaluations, the simulation results indicate that the proposed design exhibits superior performance without color oversaturation and distortion. To meet the requirements of real-time applications, a six-stage very-large-scale integration (VLSI) architecture for the proposed algorithm is implemented by using TSMC 0.13-um technology. The synthesis results show that the design yields a processing rate of approximately 200 Mpix-els/s, which is rapid enough to facilitate Full HD resolution at 30 fps in real time.
14 January 2020 in IEEE Transactions on Broadcasting