At present, with the advance of satellite image processing technology, remote sensing images are becoming more widely used in real scenes. However, due to the limitations of current remote sensing imaging technology and the influence of the external environment, the resolution of remote sensing images often struggles to meet application requirements. In order to obtain high-resolution remote sensing images, image super-resolution methods are gradually being applied to the recovery and reconstruction of remote sensing images. The use of image super-resolution methods can overcome the current limitations of remote sensing image acquisition systems and acquisition environments, solving the problems of poor-quality remote sensing images, blurred regions of interest, and the requirement for high-efficiency image reconstruction, a research topic that is of significant relevance to image processing. In recent years, there has been tremendous progress made in image super-resolution methods, driven by the continuous development of deep learning algorithms. In this paper, we provide a comprehensive overview and analysis of deep-learning-based image super-resolution methods. Specifically, we first introduce the research background and details of image super-resolution techniques. Second, we present some important works on remote sensing image super-resolution, such as training and testing datasets, image quality and model performance evaluation methods, model design principles, related applications, etc. Finally, we point out some existing problems and future directions in the field of remote sensing image super-resolution.

Traffic signs detection and recognition is an essential and challenging task for driverless cars. However, the detection of traffic signs in most scenarios belongs to small target detection, and most existing object detection methods show poor performance in these cases, which increases the difficulty of detection. To further improve the accuracy of small object detection for traffic signs, this paper proposed an optimization strategy based on the YOLOv4 network. Firstly, an improved triplet attention mechanism was added to the backbone network. It was combined with optimized weights to make the network focus more on the acquisition of channel and spatial features. Secondly, a bidirectional feature pyramid network (BiFPN) was used in the neck network to enhance feature fusion, which can effectively improve the feature perception field of small objects. The improved model and some state-of-the-art (SOTA) methods were compared on the joint dataset TT100K-COCO. Experimental results show that the enhanced network can achieve 60.4% mAP(Mean Average Precision), surpassing the YOLOv4 by 8% with the same input size. With a larger input size, it can achieve a best performance capability of 66.4% mAP. This work provides a reference for research on obtaining higher accuracy for traffic sign detection in autonomous driving.

With the accelerated development of artificial intelligence, remote-sensing image technologies have gained widespread attention in smart cities. In recent years, remote sensing object detection research has focused on detecting and counting small dense objects in large remote sensing scenes. Small object detection, as a branch of object detection, remains a significant challenge in research due to the image resolution, size, number, and orientation of objects, among other factors. This paper examines object detection based on deep learning and its applications for small object detection in remote sensing. This paper aims to provide readers with a thorough comprehension of the research objectives. Specifically, we aggregate the principal datasets and evaluation methods extensively employed in recent remote sensing object detection techniques. We also discuss the irregularity problem of remote sensing image object detection and overview the small object detection methods in remote sensing images. In addition, we select small target detection methods with excellent performance in recent years for experiments and analysis. Finally, the challenges and future work related to small object detection in remote sensing are highlighted.

Surveillance video has been widely used in business, security, search, and other fields. Identifying and locating specific pedestrians in surveillance video has an important application value in criminal investigation, search and rescue, etc. However, the requirements for real-time capturing and accuracy are high for these applications. It is essential to build a complete and smooth system to combine pedestrian detection, tracking and re-identification to achieve the goal of maximizing efficiency by balancing real-time capture and accuracy. This paper combined the detector and Re-ID models into a single end-to-end network by introducing a new track branch to YOLOv5 architecture for tracking. For pedestrian detection, we employed the weighted bi-directional feature pyramid network (BiFPN) to enhance the network structure based on the YOLOv5-Lite, which is able to further improve the ability of feature extraction. For tracking, based on Deepsort, this paper enhanced the tracker, which uses the Noise Scale Adaptive (NSA) Kalman filter to track, and adds adaptive noise to strengthen the anti-interference of the tracking model. In addition, the matching strategy is further updated. For pedestrian re-identification, the network structure of Fastreid was modified, which can increase the feature extraction speed of the improved algorithm by leaps and bounds. Using the proposed unified network, the parameters of the entire model can be trained in an end-to-end method with the multi-loss function, which has been demonstrated to be quite valuable in some other recent works. Experimental results demonstrate that pedestrians detection can obtain a 97% mean Average Precision (mAP) and that it can track the pedestrians well with a 98.3% MOTA and a 99.8% MOTP on the MOT16 dataset; furthermore, high pedestrian re-identification performance can be achieved on the VERI-Wild dataset with a 77.3% mAP. The overall framework proposed in this paper has remarkable performance in terms of the precise localization and real-time detection of specific pedestrians across time, regions, and cameras.

Lane line detection is a fundamental and critical task for geographic information perception of driverless and advanced assisted driving. However, the traditional lane line detection method relies on manual adjustment of parameters, and has poor universality, a heavy workload, and poor robustness. Most deep learning-based methods make it difficult to effectively balance accuracy and efficiency. To improve the comprehensive perception ability of lane line geographic information in a natural traffic environment, a lane line detection algorithm based on a mixed-attention mechanism residual network (ResNet) and row anchor classification is proposed. A mixed-attention mechanism is added after the backbone network convolution, normalization and activation layers, respectively, so that the model can focus more on important lane line features to improve the pertinence and efficiency of feature extraction. In addition, to achieve faster detection speed and solve the problem of no vision, the method of lane line location selection and classification based on the row direction is used to detect whether there are lane lines in each candidate point according to the row anchor, reducing the high computational complexity caused by segmentation on a pixel-by-pixel basis of traditional semantic segmentation. Based on TuSimple and CurveLane datasets, multi-scene, multi-environment, multi-linear road image datasets and video sequences are integrated and self-built, and several experiments are designed and tested to verify the effectiveness of the proposed method. The test accuracy of the mixed-attention mechanism network model reached 95.96%, and the average time efficiency is nearly 180 FPS, which can achieve a high level of accuracy and real-time detection process. Therefore, the proposed method can meet the safety perception effect of lane line geographic information in natural traffic environments, and achieve an effective balance between the accuracy and efficiency of actual road application scenarios.

Geological models are essential components in various applications. To generate reliable realizations, the geostatistical method focuses on reproducing spatial structures from training images (TIs). Moreover, uncertainty plays an important role in Earth systems. It is beneficial for creating an ensemble of stochastic realizations with high diversity. In this work, we applied a pattern classification distribution (PCD) method to quantitatively evaluate geostatistical modeling. First, we proposed a correlation-driven template method to capture geological patterns. According to the spatial dependency of the TI, region growing and elbow-point detection were launched to create an adaptive template. Second, a combination of clustering and classification was suggested to characterize geological realizations. Aiming at simplifying parameter specification, the program employed hierarchical clustering and decision tree to categorize geological structures. Third, we designed a stacking framework to develop the multi-grid analysis. The contribution of each grid was calculated based on the morphological characteristics of TI. Our program was extensively examined by a channel model, a 2D nonstationary flume system, 2D subglacial bed topographic models in Antarctica, and 3D sandstone models. We activated various geostatistical programs to produce realizations. The experimental results indicated that PCD is capable of addressing multiple geological categories, continuous variables, and high-dimensional structures.

No abstract