In human conversations, individuals can indicate relevant regions within a scene while addressing others. turn, the other person then respond by referring to specific if necessary. This natural referential ability in dialogue remains absent current Multimodal Large Language Models (MLLMs). To fill this gap, paper proposes an MLLM called Shikra, which handle spatial coordinate inputs and outputs language. Its architecture consists of vision encoder, alignment layer, LLM. It is designed be...

Current state-of-the-art trajectory methods do not perform well in the terminal airspace that surrounds an airport due to its complex structure and frequently changing flight postures of aircraft. Since aircraft takes off or lands must follow a specified procedure, this paper will learn data-driven prediction model from many historical trajectories improve accuracy robustness airspace. A regularization method is utilized reconstruct each obtain high-quality with equal time intervals no...

Aircraft trajectory prediction is the basis of approach and departure sequencing, conflict detection resolution other air traffic management technologies. Accurate can help increase airspace capacity ensure safe orderly operation aircraft. Current research focuses on single aircraft without considering interaction between Therefore, this paper proposes a model based Social Long Short-Term Memory (S-LSTM) network to realize multi-aircraft collaborative prediction. This establishes an LSTM for...

The aircraft trajectory clustering analysis in the terminal airspace is conducive to determining representative route structure of arrival and departure extracting their typical patterns, which important for air traffic management such as optimization, planning, prediction. However, current methods perform poorly due large flight traffic, high density, complex airspace. In recent years, continuous development Deep Learning has demonstrated its powerful ability extract internal potential...

Air traffic control (ATC) hazard feature extraction is a key information retrieval task for air records. While text-based ranks term importance based solely on statistical results, we aim to use external knowledge extract features that meet the definition of hazards. This paper proposes method expert define and construct analysis framework. We illustrate model training process using communication navigation surveillance (CNS) data, which includes candidate generation, vectorization,...

In light of the rapid expansion civil aviation, addressing delays and congestion phenomena in vicinity metroplex caused by imbalance between air traffic flow capacity is crucial. This paper first proposes a bi-level optimization model for collaborative flight sequencing arrival departure flights with multiple airports, considering both runway systems TMA (Terminal Control Area) entry/exit fixes. Besides, adaptive to various scenarios. The genetic algorithm employed solve proposed model....

The non-local means (NLM) provides a useful tool for image denoising and many variations of the NLM method have been proposed. However, few works tried to tackle task adaptively choosing patch size according region characteristics. Presented is region-based noise removal. proposed first analyses classifies into several types. According type, local window adjusted match property region. Experimental results show effectiveness demonstrate its superiority state-of-the-art methods.

Driven by the spectacular success of deep learning, several advanced models based on neural networks have recently been proposed for single-image super-resolution, incrementally revealing their superiority over alternatives. In this paper, we pursue latest line research and present an improved network structure taking advantage component learning. The core idea difference learning strategy are to use residual extracted from input predict its counterpart in corresponding output. To end, a...

In this paper, we propose a novel superresolution (SR) reconstruction algorithm to handle license plate texts in real traffic videos. To make numbers more legible, generalized discontinuity-adaptive Markov random field (DAMRF) model is proposed based on the recently reported bilateral filtering, which not only preserves edges but robust noise as well. Moreover, instead of looking for fixed value regularization parameter, method automatically estimating it applied input images. Information...

Learning long-term dependences (LTDs) with recurrent neural networks (RNNs) is challenging due to their limited internal memories. In this paper, we propose a new external memory architecture for RNNs called an addressable and working (EALWM)-augmented RNN. This has two distinct advantages over existing architectures, namely the division of into parts-long-term memory-with both capability learn LTDs without suffering from vanishing gradients necessary assumptions. The experimental results on...

Due to the strong propagation causality of delays between airports, this paper proposes a delay prediction model based on deep graph neural network study from perspective an airport network. We regard airports as nodes and use directed construct airports’ relationship. For adjacent weights edges are measured by spherical distance them, while number flight pairs them is utilized for connected flights. On basis, diffusion convolution kernel constructed capture characteristics it further...

Multi frame super-resolution (SR) reconstruction algorithms make use of complimentary information among low-resolution (LR) images to yield a high-resolution (HR) image. Inspired by recent development on the video denoising problem, we propose robust variational approach for SR-based constrained model that uses nonlocal total variation (TV) as regularisation term. In our method, weighted fidelity term is proposed take into account inaccurate estimates registration parameters and point spread...