License plate recognition (LPR) is an important component of intelligent transportation systems. Compared with letters and numbers, Chinese characters contain more information, making automatic difficult. Accurate LPR (CLPR) determined by three factors: training dataset, feature extractor, classifier. Most license plates benchmark dataset only numbers; thus, the authors build a large for CLPR. Convolutional neural networks (CNNs) can be used to extract inherent image features, on all levels...

This letter adopts level-set methods in order to seek a novel classification strategy which is free of segmentation. A region-driven multiple-level-set (MLS) framework used perform very high resolution image classification. Two specific unsupervised models are presented. First, from the point view feature fusion, an MLS model suggested by fusing texture features and spectral information (TSMLS model). The combines information, extracted image, geometrical characteristics closed curves...

Based on a diversely polarized antenna, we address the problems of adaptive detection and performance enhancement in partially homogeneous environments where test training data samples share same noise covariance matrix up to an unknown scaling factor. The matched subspace detector are employed handle problem for cases known structure, respectively. two detectors is evaluated terms their probabilities false alarm detection. In particular, waveform design algorithm enhance proposed. A...

This letter proposes an associative hierarchical conditional random field (AHCRF) model to improve the classification accuracy of high-resolution remote sensing images. It considers segmentation quality superpixels, avoids a time-consuming selection optimal scale parameters, and alleviates problem sensitive undersegmentation errors that is present in traditional object-oriented methods. The built on graph hierarchy, including pixel layer as base multiple superpixel layers derived from mean...

Precise vegetation maps of mountainous areas are great significance to grasp the situation an ecological environment and forest resources. In this paper, while multi-source geospatial data can generally be quickly obtained at present, realize effective mapping in when samples difficult collect due their perilous terrain inaccessible deep forest, we propose a novel intelligent method sample collection for machine-learning (ML)-based mapping. First, employ geo-objects (i.e., polygons) from...

Abstract. For the purpose of extracting productions some specific branching plants effectively and realizing its 3D reconstruction, Terrestrial LiDAR data was used as extraction source production, a reconstruction method based on technologies combined with L-system proposed in this article. The topology structure plant architectures extracted using point cloud target space level segmentation mechanism. Subsequently, were obtained structural parameters production rules branches, which fit...

This paper studies an new approach to creating a variational level set model for buildings detection by combining LiDAR point clouds and Aerial image data. The introduces object-based analysis technique. Firstly, fundamental framework is built neighbor of remote sensing image. Then, several derived products directly or indirectly from raw cloud data, like nDSM absolute roughness are used construct novel energy term in relation height non-terrain objects, order make up the disadvantages...

Recently, active contour models based on local information have emerged in image segmentation. These are more robust to variations of region interest. But it also brought some new problems, such as minimum, higher computational cost. To effectively alleviate these this paper presents a novel fast model driven by global-local statistical energy. Firstly, is constructed capture the border object accurately. Secondly, global features integrated with so create an improved which can avoid minimum...

Because hyperspectral remote sensing images contain a lot of redundant information and the data structure is highly non-linear, leading to low classification accuracy traditional machine learning methods. The latest research shows that image based on deep convolutional neural network has high accuracy. However, when small amount used for training, methods greatly reduced. In order solve problem existing algorithms samples images, multi-scale residual proposed. extraction fusion spatial...

The combination of UAV-LiDAR and LiDAR-SLAM (Simultaneous Localization Mapping) technology can overcome the scanning limitations different platforms obtain comprehensive 3D structural information forest stands. To address challenges traditional registration algorithms, such as high initial value requirements susceptibility to local optima, in this paper, we propose a high-precision, robust, NDT-VGICP method that integrates voxel features register point clouds at stand scale. First, are...

Abstract Snowmelt floods are highly hazardous meteorological disasters that can potentially threaten human lives and property. Hence, snowmelt susceptibility mapping (SSM) plays an important role in flood prevention systems aids emergency responders risk managers. In this paper, a method of identifying hazards is proposed, large‐scale hazard zonation scheme based on historical recordings multisource remote sensing data established. To assess the quality our approach, proposed model was...

Satellite imagery especially with high spatial resolution often shows spectral variations and details disturbances in a class. These characteristics bring difficulties to people who are working at automatic classification the remote sensing fields. To seek more effective method, this paper presents new multiple level set model implement unsupervised for multispectral images. Firstly, medium filtering technique oriented from image processing is introduced into traditional improve performance...

전라북도 고창군 상하면의 가막도에 대한 지형 조사의 일환으로 암석 반발 강도 조사와 각 부분의 화학적 조성을 분석하였다. 가막도의 암체 최상부는 지중 풍화 기원인 핵석이 일부 나타나고 있으며, 그보다 고도가 낮은 부분엔 지중에서 전선을 형성하던 기반암이 노출되어 있다. 가막도 암체의 강도는 정상부에서 해안 대지 쪽으로 가면서 증가하는 것으로 나타났다. 이는 정상부가 침식에 저항 정도가 약한 볼 수 한편 지수, 풍화각 조성 물질의 변화 등으로 때 하부가 정상부에 비하여 풍화가 덜 진행된 하지만 상부의 표면부에서 나타나는 현상으로 풍화층이 제거될 경우 생경한 암석의 풍화대가 나타날 보인다. 풍화각에 발달한 다각형 균열의 전반적으로 조성과 정도에서 정상부와 하부의 중간정도의 수준이었다. 풍화각의 특성을 파악하기 위한 분석 결과 표면에 균열이 형성되는 경우, 표면 부분에 기저 부분이 지수가 높은 경향성이 이전의 연구와 일치하나, SiO2의 함량 등은 일치하지 않는 나타났으며, 환경과...

The proposed approach in this paper is divided into three steps namely the location of plate, segmentation characters and recognition characters. algorithm firstly consisted two video captures to get high quality images, estimates size vehicle plate these images via parallel binocular stereo vision algorithm. Then method extracts edge based on second generation non-orthogonal Haar wavelet transformation, locates according estimated result first step. Finally, realized Radial Basis Function...

To further improve the accuracy of crop detection and acquire more information for land use investigation agriculture management, this paper proposes a variational level set model by combining airborne LiDAR(Light Detection Range) points cloud aerial image simultaneously acquired LiDAR device. Specifically, normalized digital surface (nDSM) derived from raw are combined with so as to alleviate misclassification caused insufficient only based on remote sensing data. This fusion combines...

Multispectral remotely sensing imagery with high spatial resolution, such as QuickBird, IKONOS satellite or Aerial imagery, especially in urban scenes, often perform spectral variations and rich details within a category, resulting poor accuracy of classification. To seek an efficient solution, this paper presents non-parametric variational multiple level set model by joint use image two products, digital terrain (DTM) surface (DSM), directly indirectly derived from raw LiDAR (Light...

This paper presents a new approach to creating variational level set model for vegetation detection combining 3D irregular point clouds and aerial image simultaneously acquired by LiDAR light scanning imaging device. Firstly, fundamental statistical framework is built which integrates texture information improve the quality of detection. Then, several derived products directly or indirectly from raw cloud data, like DTM(digital terrain model,) nDSM(normalized digital surface model) local...

Objects detection in high resolution (HR) remote sensing images plays an important role modern military, national defense, and commercial field. Because of a variety object types different sizes, it is difficulty to realize the rapid multi-scale objects, provides support for succeeding decision making responses. This paper proposes fast method image objects with deep learning model, named YOLOv <sub xmlns:mml="http://www.w3.org/1998/Math/MathML"...

This paper presents a method for ports detection based on the framework of feature level fusion. Bearing in mind fact that parallel lines and rectangular corners are main features most ports, large scale man-made objects, these firstly extracted from high-to-moderate resolution optical satellite imagery. Taking account balance data acquisition spatial resolution, SPOT panchromatic image is used such extraction. Considering whether conditions coastal area, which characterized by rainy cloudy...

Post-classification analysis is an important way for remotely sensed imagery change detection. In this paper, we propose a novel classification detection using multispectral IKONOS imagery. The called after Enhanced Growing Self-Organization Map (EGSOM). EGSOM designed to solve two limitation of traditional Self Organization Feature (SOM). One the training time SOM endless, other SOM's structure fixed before train. make use and network's weights are initialized hierachical clustering method....