In this paper, we propose a depth-map merging based multiple view stereo method for large-scale scenes which takes both accuracy and efficiency into account. the proposed method, an efficient patch-based matching process is used to generate at each image with acceptable errors, followed by refinement enforce consistency over neighboring views. Compared state-of-the-art methods, can reconstruct quite accurate dense point clouds high computational efficiency. Besides, could be easily...

Batteries have the potential to significantly reduce greenhouse gas emissions from on-road transportation. However, environmental and social impacts of producing lithium-ion batteries, particularly cathode materials, concerns over material criticality are frequently highlighted as barriers widespread electric vehicle adoption. Circular economy strategies, like reuse recycling, can secure regional supplies. To understand for circularity, we undertake a dynamic global flow analysis pack-level...

Structure-from-Motion (SfM) methods can be broadly categorized as incremental or global according to their ways estimate initial camera poses. While system has advanced in robustness and accuracy, the efficiency remains its key challenge. To solve this problem, reconstruction simultaneously estimates all poses from epipolar geometry graph, but it is usually sensitive outliers. In work, we propose a new hybrid SfM method tackle issues of efficiency, accuracy unified framework. More...

Intelligent transportation systems (ITS) and navigation usually demand good timeliness on their used vector road network maps. Unfortunately, existing methods for map generation, such as the surveying method satellite image digitizing method, are difficult to generate up-to-date This paper presents a automatic generation of from massive GPS vehicle trajectories, which can trajectory data that be widely available. Given set data, input first is processed construct bitmap depicts network. Then...

Purpose The purpose of this paper is to propose a screw axis identification (SAI) method based on the product exponentials (POE) model, which concerned with calibrating serial robot m joints equipped stereo‐camera vision system. Design/methodology/approach Different from conventional approaches, like circle point analysis (CPA) or system theoretic must collect great deal data, joint parameters for proposed only needs measure m+1 times n (n≥3) target points mounted manipulator end‐effector....

One of the potentially effective means for large-scale 3D scene reconstruction is to reconstruct in a global manner, rather than incrementally, by fully exploiting available auxiliary information on imaging condition, such as camera location Global Positioning System (GPS), orientation inertial measurement unit (or compass), focal length from EXIF, and so on. However, information, though informative valuable, usually too noisy be directly usable. In this paper, we present an approach taking...

As an intermediate step between raw remote sensing data and digital maps, classification has been a challenging long-standing problem in the research community. In this work, automated effective supervised framework is presented for classifying high-resolution data. Specifically, method proceeds three main stages: feature extraction, classification, classified result refinement. extraction stage, both multispectral images 3D geometry are used, which utilizes complementary information from...

Depth completion aims to recover pixelwise depth from incomplete and noisy measurements with or without the guidance of a reference RGB image. This task attracted considerable research interest due its importance in various computer vision-based applications, such as scene understanding, autonomous driving, 3-D reconstruction, object detection, pose estimation, trajectory prediction, so on. As system input, an map is usually generated by projecting points collected ranging sensors, LiDAR...

Image based modeling and laser scanning are two commonly used approaches in large-scale architectural scene reconstruction nowadays. In order to generate a complete reconstruction, an effective way is completely cover the using ground aerial images, supplemented by on certain regions with low textures complicated structures. Thus, key issue accurately calibrate cameras register scans unified framework. To this end, we proposed three-step pipeline for merging images scans. First, captured...

Semantic segmentation of 3D scenes is one the most important tasks in field computer vision and has attracted much attention. In this paper, we propose a novel framework for semantic aerial photogrammetry models, which uses orthographic projection to improve efficiency while still ensuring high precision, can also be applied multiple types models (i.e., textured mesh or colored point cloud). our pipeline, first obtain RGB images elevation from scene through projection, then use image network...

In this paper, we put forward a new method for surface reconstruction from image-based point clouds. particular, introduce visibility model each line of sight to preserve scene details without decreasing the noise filtering ability. To make proposed suitable clouds with heavy noise, likelihood energy term total binary labeling problem Delaunay tetrahedra, and give its s-t graph implementation. Besides, further improve performance dense technique, which helps keep object edge sharp. The...

Urban scene modeling is a challenging task for the photogrammetry and computer vision community due to its large scale, structural complexity, topological delicacy. This paper presents an efficient multistep framework large-scale urban scenes from aerial images. It takes images textured 3D mesh model generated by image-based system as input outputs compact polygon models with semantics at different levels of detail (LODs). Based on key observation that buildings usually have piecewise planar...

With the popularization of smartphones, larger collection videos with high quality is available, which makes scale scene reconstruction increase dramatically. However, high-resolution video produces more match outliers, and frame rate brings redundant images. To solve these problems, a tailor-made framework proposed to realize an accurate robust structure-from-motion based on monocular videos. The key ideas include two points: one use spatial temporal continuity sequences improve accuracy...

Visual localization, also known as camera pose estimation, is a crucial component of many applications, such robotics, autonomous driving, and augmented reality. Traditional visual localization algorithms typically run on point cloud maps generated by Structure-from-Motion (SfM) or Simultaneous Localization Mapping (SLAM). However, features are sensitive to weak textures illumination changes. In addition, the 3D often contain millions points, which puts higher demands device storage...

We present a method for recovering the 3-D shape of an inextensible deformable surface from monocular image sequence. State-of-the-art methods on this problem, utilize <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">L</i> <sub xmlns:xlink="http://www.w3.org/1999/xlink">¿</sub> -norm reprojection residual vectors and formulate tracking problem as Second-Order Cone Programming (SOCP) problem. Instead using which is sensitive to outliers, we use...

The incremental Structure-from-Motion (SfM) technique has advanced in both robustness and accuracy, but the efficiency scalability remain its key challenges. In this paper, we propose a novel batched SfM to tackle these problems unified framework, where two iteration loops are contained. inner loop is tracks triangulation loop, selection method proposed find compact subset of for bundle adjustment (BA). outer camera registration batch cameras simultaneously added alleviate drifting risk...