Search

3D reconstruction of scenes using multiple images, relying on robust correspondence search and depth estimation, has been thoroughly studied for the two‐view and multi‐view scenarios in recent years. Multi‐view stereo (MVS) algorithms aim to generate a rich, dense 3D model of the scene in the form of a dense point cloud or a triangulated mesh. In a typical MVS pipeline, the robust estimations for the camera poses along with the sparse points obtained from structure from motion (SfM) are used as input. During this process, the depth of generally every pixel of the scene is to be calculated. Several methods, either conventional or, more recently, learning‐based have been developed for solving the correspondence search problem. A vast amount of research exists in the literature using local, global or semi‐global stereomatching approaches, with the PatchMatch algorithm being among the most popular and efficient conventional ones in the last decade. Yet, and despite the widespread evolution of the algorithms, yielding complete, accurate and aesthetically pleasing 3D representations of a scene remains an open issue in real‐world and large‐scale photogrammetric applications. This work aims to provide a concrete survey on the most widely used MVS methods, investigating underlying concepts and challenges. To this end, the theoretical background and relative literature are discussed for both conventional and learning‐based approaches, with a particular focus on close‐range 3D reconstruction applications. [ABSTRACT FROM AUTHOR]

Image-based 3D reconstruction is widely used in practice due to less hardware constraints,lower cost and higher flexibility.Especially for the problems of sparseness and uneven density of the three-dimensional point cloud data generated by the image due to the occlusion between various parts of the object,it has always been a difficulty and hot issue to deal with.In this paper,a mask-based sparse point cloud filtering algorithm is proposed.Firstly,the bounding box of the point cloud is calculated and the grid is adaptively divided according to the sparseness of the point cloud.Secondly,Depth-first search is used to recursively find all customized connected domains composed of grids generated at the first step.Then adaptively calculating the threshold based on the quantized importance index,selecting the connected domains that should be retained based on the adaptive threshold,and defining the set of all retained connected domains as a mask,which is used to describe the global spatial topology information of the sparse point cloud.Finally,points covered by the mask are retained while points of the uncovered area are removed,so as to filter the outliers.This method can handle the point cloud data generated by occlusion and with great differences in spatial density.It can effectively remove outliers in the original three-dimensional point cloud data,while maintaining the detailed information of the point cloud.

Forest cover change can be detected with high precision using 3D geospatial data and semi-automatic analyses of Remote Sensing data. The aim of our study, performed in Gorce National Park in Poland, was to generate a land use land cover (LULC) map and use it to analyse forest cover change. The study area is a subalpine forest region that has been affected by bark beetle and wind disturbances. The Geographic Object-Based Image Analysis approach was used for classification, with Colour Infrared orthophotos and normalized Digital Surface Models generated using image-matching approach. Gathered results showed that dominating LULC class is coniferous forests (3380 ha; 47% of study area), when second largest class is deciduous forests (2204 ha; 30%). The dead Norway spruce stands (465.5 ha; 6.5%) showed significant increase comparing to 114.1 ha mapped in 1997.

Forest cover change can be detected with high precision using 3D geospatial data and semi-automatic analyses of Remote Sensing data. The aim of our study, performed in Gorce National Park in Poland, was to generate a land use land cover (LULC) map and use it to analyse forest cover change. The study area is a subalpine forest region that has been affected by bark beetle and wind disturbances. The Geographic Object-Based Image Analysis approach was used for classification, with Colour Infrared orthophotos and normalized Digital Surface Models generated using image-matching approach. Gathered results showed that dominating LULC class is coniferous forests (3380 ha; 47% of study area), when second largest class is deciduous forests (2204 ha; 30%). The dead Norway spruce stands (465.5 ha; 6.5%) showed significant increase comparing to 114.1 ha mapped in 1997. [ABSTRACT FROM AUTHOR]

Stereomatching is an effective way of acquiring dense depth information from a scene when active measurements are not possible. So-called lightfield methods take a snapshot from many camera locations along a defined trajectory (usually uniformly linear or on a regular grid--we will assume a linear trajectory) and use this information to compute accurate depth estimates. However, they require the locations for each of the snapshots to be known: the disparity of an object between images is related to both the distance of the camera to the object and the distance between the camera positions for both images. Existing solutions use sparse feature matching for camera location estimation. In this paper, we propose a novel method that uses dense correspondences to do the same, leveraging an existing depth estimation framework to also yield the camera locations along the line. We illustrate the effectiveness of the proposed technique for camera location estimation both visually for the rectification of epipolar plane images and quantitatively with its effect on the resulting depth estimation. Our proposed approach yields a valid alternative for sparse techniques, while still being executed in a reasonable time on a graphics card due to its highly parallelizable nature. [ABSTRACT FROM AUTHOR]

In recent years the term "precise forestry" has been used more and more often, referring to a modern and sustainable model of forest management. Functioning of such management of wood biomass resources is based, among others, on precisely defined and log-term monitored selected forest taxation parameters of single trees and whole forest stands based on modern geoinformation technologies, including Airborne Laser Scanning (ALS) and digital photogrammetry. The purpose of the work was the analysis of the usefulness of the CHM (Canopy Height Model) generated from the image-based point cloud or ALS technology to define the number of trees using the method of the segmentation of single Scots pine (Pinus sylvestris L.) crowns. The study was carried out in the Scots pine stands located in the Bory Tucholskie National Park (Poland). Due to the intentional lack of certain silviculture treatments, over the recent decades, these forest stands have been characterized by relatively high tree density, compared to managed forests. The CHM was generated from digital airborne photos (CIR composition; GSD 0.15 m) and on the other hand - from the ALS point clouds (4 points/m2; ISOK project). To generate point clouds from airborne photos using stereomatching method, the PhotoScan Professional (Agisoft) software was applied. The CHM coming from the Image-Based Point Cloud (CHM_IPC; GSD: 0.30 m) and ALS data (CHM_ALS; GSD: 0.75 m) were generated using FUSION (USDA Forest Service) software. The segmentation of tree crowns was carried out in eCognition Developer (TRIMBLE GeoSpatial) software. Apart from height models, also spectral information was used (so-called true CIR orthophotomaps; GSD: 0.3 and 0.75 m). To assess the accuracy of the obtained results, the ground truth data from 248 reference areas were used. The carried out analyses showed that in forest stands of younger age classes (< 120 years) better results were achieved applying the method of image matching (CHM_IPC), while in the case of older stands (> 120 years) the accuracy of the detection rate of tree crowns was the highest when CHM_ALS model was applied. The mean percentage error (defined by the number of trees, based on the detection of single pine crowns), calculated based on 248 ground truth areas was 0.89%, which shows a great potential of digital photogrammetry (IPC) and GEOBIA. In case of almost full nationwide cover in Poland of airborne digital images (present IPC models) and ALS point clouds (DTM and DSM), at almost 71% forest stands in the Polish State Forests National Forest Holding (PGL LP), one can assume wide application of geodata (available free of charge) in precise modelling of selected tree stand parameters all over Poland. [ABSTRACT FROM AUTHOR]

Census transform is a non-parametric local transform which is based on the relative ordering of pixels instead of the pixel intensity values within a window; therefore it is insensitive to radiometric variations like exposure changes and illumination variations. This work focuses on an efficient census filtering based local steromatching for images taken under varying radiometric conditions. For matching cost value computation hamming distance is taken for the dissimilarity measure. Finally for cost aggregation guided filtering is used. Guided filter has edge preserving property and runs faster than the bilateral filtering based cost aggregation. The experimental results shows that the proposed algorithm provides better disparities for images taken under varying radiometric conditions than the conventional stereomatching algorithm. [ABSTRACT FROM AUTHOR]

Stereomatching is an effective way of acquiring dense depth information from a scene when active measurements are not possible. So-called lightfield methods take a snapshot from many camera locations along a defined trajectory (usually uniformly linear or on a regular grid—we will assume a linear trajectory) and use this information to compute accurate depth estimates. However, they require the locations for each of the snapshots to be known: the disparity of an object between images is related to both the distance of the camera to the object and the distance between the camera positions for both images. Existing solutions use sparse feature matching for camera location estimation. In this paper, we propose a novel method that uses dense correspondences to do the same, leveraging an existing depth estimation framework to also yield the camera locations along the line. We illustrate the effectiveness of the proposed technique for camera location estimation both visually for the rectification of epipolar plane images and quantitatively with its effect on the resulting depth estimation. Our proposed approach yields a valid alternative for sparse techniques, while still being executed in a reasonable time on a graphics card due to its highly parallelizable nature.

The rapid developments in the field of digital aerial photogrammetry (DAP) in recent years have increased interest in the application of DAP data for extracting three-dimensional (3D) models of forest canopies. This technology, however, still requires further investigation to confirm its reliability in estimating forest attributes in complex forest conditions. The main purpose of this study was to evaluate the accuracy of tree height estimation based on a crown height model (CHM) generated from the difference between a DAP-derived digital surface model (DSM) and an airborne laser scanning (ALS)-derived digital terrain model (DTM). The tree heights determined based on the DAP-CHM were compared with ground-based measurements and heights obtained using ALS data only (ALS-CHM). Moreover, tree- and stand-related factors were examined to evaluate the potential influence on the obtained discrepancies between ALS- and DAP-derived heights. The obtained results indicate that the differences between the means of field-measured heights and DAP-derived heights were statistically significant. The root mean square error (RMSE) calculated in the comparison of field heights and DAP-derived heights was 1.68 m (7.34%). The results obtained for the CHM generated using only ALS data produced slightly lower errors, with RMSE = 1.25 m (5.46%) on average. Both ALS and DAP displayed the tendency to underestimate tree heights compared to those measured in the field; however, DAP produced a higher bias (1.26 m) than ALS (0.88 m). Nevertheless, DAP heights were highly correlated with the heights measured in the field (R2 = 0.95) and ALS-derived heights (R2 = 0.97). Tree species and height difference (the difference between the reference tree height and mean tree height in a sample plot) had the greatest influence on the differences between ALS- and DAP-derived heights. Our study confirms that a CHM computed based on the difference between a DAP-derived DSM and an ALS-derived DTM can be successfully used to measure the height of trees in the upper canopy layer. [ABSTRACT FROM AUTHOR]

Disparity acquisition is beneficial to various 3D multimedia applications, i.e., free viewpoint TV (FTV), 3DTV, and 3D scene model reconstruction. In this paper, we propose an efficient disparity map estimation method with occlusion handling taken into account. In order to detect occlusion, we obtain left and right initial disparity maps via optimization based on modified constant-space belief propagation, a low complexity algorithm. The initial disparity maps provide clues for occlusion detection. From the clues, an energy function is defined and optimized by energy minimization framework. We define two occlusion types and apply suitable occlusion handling processes, accordingly. The proposed occlusion handling method extends disparity values of visible pixels to occluded pixels. Experimental results show that our method provides near-ground truth occlusion and a superior disparity map compared to other stateof- the-art methods with occlusion handling. Our contribution provides solid advantages regarding disparity map estimation, which is useful for various 3D multimedia applications. [ABSTRACT FROM AUTHOR]