Road Vector Research Articles

Automatic Registration of Wide Area Motion Imagery to Vector Road Maps by Exploiting Vehicle Detections.

To enrich large-scale visual analytics applications enabled by aerial wide area motion imagery (WAMI), we propose a novel methodology for accurately registering a geo-referenced vector roadmap to WAMI by using the locations of detected vehicles and determining a parametric transform that aligns these locations with the network of roads in the roadmap. Specifically, the problem is formulated in a probabilistic framework, explicitly allowing for spurious detections that do not correspond to on-road vehicles. The registration is estimated via the expectation-maximization (EM) algorithm as the planar homography that minimizes the sum of weighted squared distances between the homography-mapped detection locations and the corresponding closest point on the road network, where the weights are estimated posterior probabilities of detections being on-road vehicles. The weighted distance minimization is efficiently performed using the distance transform with the Levenberg-Marquardt nonlinear least-squares minimization procedure, and the fraction of spurious detections is estimated within the EM framework. The proposed method effectively sidesteps the challenges of feature correspondence estimation, applies directly to different imaging modalities, is robust to spurious detections, and is also more appropriate than feature matching for a planar homography. Results over three WAMI data sets captured by both visual and infrared sensors indicate the effectiveness of the proposed methodology: both visual comparison and numerical metrics for the registration accuracy are significantly better for the proposed method as compared with the existing alternatives.

Blending zone determination for aerial orthimage mosaicking

Creating a composed image from a set of aerial images is a fundamental step in orthomosaic generation. One of the processes involved in this technique is determining an optimal seamline in an overlapping region to stitch image patches seamlessly. Most previous studies have solved this optimization problem by searching for a one-pixel-wide seamline with an objective function. This strategy significantly reduced pixel mismatches on the seamline caused by geometric distortions of images but did not fully consider color discontinuity and mismatch problems that occur around the seamline, which sometimes cause mosaicking artifacts. This study proposes a blending zone determination scheme with a novel path finding algorithm to reduce the occurrence of unwanted artifacts. Instead of searching for a one-pixel-wide seamline, a blending zone, which is a k-pixel-wide seamline that passes through high-similarity pixels in the overlapping region, is determined using a hierarchical structure. This strategy allows for not only seamless stitching but also smooth color blending of neighboring image patches. Moreover, the proposed method searches for a blending zone without the pre-process of highly mismatched pixel removal and additional geographic data of road vectors and digital surface/elevation models, which increases the usability of the approach. Qualitative and quantitative analyses of aerial images demonstrate the superiority of the proposed method to related methods in terms of avoidance of passing highly mismatched pixels.

VECTORIZATION OF ROAD DATA EXTRACTED FROM AERIAL AND UAV IMAGERY

Abstract. Road databases are essential instances of urban infrastructure. Therefore, automatic road detection from sensor data has been an important research activity during many decades. Given aerial images in a sufficient resolution, dense 3D reconstruction can be performed. Starting at a classification result of road pixels from combined elevation and optical data, we present in this paper a fivestep procedure for creating vectorized road networks. These main steps of the algorithm are: preprocessing, thinning, polygonization, filtering, and generalization. In particular, for the generalization step, which represents the principal area of innovation, two strategies are presented. The first strategy corresponds to a modification of the Douglas-Peucker-algorithm in order to reduce the number of vertices while the second strategy allows a smoother representation of street windings by Bezir curves, which results in reduction – to a decimal power – of the total curvature defined for the dataset. We tested our approach on three datasets with different complexity. The quantitative assessment of the results was performed by means of shapefiles from OpenStreetMap data. For a threshold of 6 m, completeness and correctness values of up to 85% were achieved.

VECTORIZATION OF ROAD DATA EXTRACTED FROM AERIAL AND UAV IMAGERY

Road databases are essential instances of urban infrastructure. Therefore, automatic road detection from sensor data has been an important research activity during many decades. Given aerial images in a sufficient resolution, dense 3D reconstruction can be performed. Starting at a classification result of road pixels from combined elevation and optical data, we present in this paper a fivestep procedure for creating vectorized road networks. These main steps of the algorithm are: preprocessing, thinning, polygonization, filtering, and generalization. In particular, for the generalization step, which represents the principal area of innovation, two strategies are presented. The first strategy corresponds to a modification of the Douglas-Peucker-algorithm in order to reduce the number of vertices while the second strategy allows a smoother representation of street windings by Bezir curves, which results in reduction – to a decimal power – of the total curvature defined for the dataset. We tested our approach on three datasets with different complexity. The quantitative assessment of the results was performed by means of shapefiles from OpenStreetMap data. For a threshold of 6 m, completeness and correctness values of up to 85% were achieved.

Automatic Geometric Processing for Very High Resolution Optical Satellite Data Based on Vector Roads and Orthophotos

In response to the increasing need for fast satellite image processing SPACE-SI developed STORM—a fully automatic image processing chain that performs all processing steps from the input optical images to web-delivered map-ready products for various sensors. This paper focuses on the automatic geometric corrections module and its adaptation to very high resolution (VHR) multispectral images. In the automatic ground control points (GCPs) extraction sub-module a two-step algorithm that utilizes vector roads as a reference layer and delivers GCPs for high resolution RapidEye images with near pixel accuracy was initially implemented. Super-fine positioning of individual GCPs onto an aerial orthophoto was introduced for VHR images. The enhanced algorithm is capable of achieving accuracy of approximately 1.5 pixels on WorldView-2 data. In the case of RapidEye images the accuracies of the physical sensor model reach sub-pixel values at independent check points. When compared to the reference national aerial orthophoto the accuracies of WorldView-2 orthoimages automatically produced with the rational function model reach near-pixel values. On a heterogeneous set of 41 RapidEye images the rate of automatic processing reached 97.6%. Image processing times remained under one hour for standard-size images of both sensor types.

Vector Road Map Registration to Oblique Wide Area Motion Imagery by Exploiting Vehicle Movements

Vector Road Map Registration to Oblique Wide Area Motion Imagery by Exploiting Vehicle Movements

Automatic vehicle detection based on automatic histogram-based fuzzy C-means algorithm and perceptual grouping using very high-resolution aerial imagery and road vector data

This study presents an approach for the automatic detection of vehicles using very high-resolution images and road vector data. Initially, road vector data and aerial images are integrated to extract road regions. Then, the extracted road/street region is clustered using an automatic histogram-based fuzzy C-means algorithm, and edge pixels are detected using the Canny edge detector. In order to automatically detect vehicles, we developed a local perceptual grouping approach based on fusion of edge detection and clustering outputs. To provide the locality, an ellipse is generated using characteristics of the candidate clusters individually. Then, ratio of edge pixels to nonedge pixels in the corresponding ellipse is computed to distinguish the vehicles. Finally, a point-merging rule is conducted to merge the points that satisfy a predefined threshold and are supposed to denote the same vehicles. The experimental validation of the proposed method was carried out on six very high-resolution aerial images that illustrate two highways, two shadowed roads, a crowded narrow street, and a street in a dense urban area with crowded parked vehicles. The evaluation of the results shows that our proposed method performed 86% and 83% in overall correctness and completeness, respectively.

Multi-Criteria, Graph-Based Road Centerline Vectorization Using Ordered Weighted Averaging Operators

In this paper a novel road vectorization methodology based on image space clustering technique and weighted graph theory is presented. The proposed methodology describes a road as a set of optimized points on the centerline which should be connected by defining a number of appropriate criteria. The main contribution of this paper is to design a weighting scheme for combining a small number of road identities using Ordered Weighted Averaging (owa) operators by defining appropriate decision strategy. In this regard, a novel geometric criterion is introduced. Result of the OWA aggregation specifies weight of each edge in the road network graph. Comparing the proposed approach with two state-of-the-art image space clustering-based road vectorization methods proves its efficiency to deal with roads with different widths, parallel roads with different distances, different types of intersections, and also noise clusters. Obtaining improved quality measures for several high-resolution images, demonstrates the successfulness of the vectorization approach.

Automated Road Extraction From High Resolution Satellite Images

The importance of road extraction from satellite images arises from the fact that it greatly enhances the efficiency of map generation and thus can be a big help in car navigations systems or any emergency (rescue) system that needs instant maps. Therefore, increasing research is being dedicated and focused on the development of efficient methods to extract topographical meaningful features (like roads) from digital remote sensed images. The work deals with extraction of roads from satellite images. This is a challenging domain compared to extraction from aerial images as satellite images are noisy and of lower resolution. In this method, a Vectorization Approach for the automatic method of road extraction is being used where the image is segmented to identify the road network regions followed by a decision making and continuity procedure to correctly detect the roads and the Vectorization step to identify the line segments or curved segments which represents the road. This method may be employed for obtaining information for feeding large-scale Geographic Information System. In the automatic method of road extraction the extracted roads are converted into road vectors in order to use these vector road maps in GIS. A semi-automated scheme is used for scenarios where fully automated system fails. A combination of both methods can be devised for a full fledged real business scenario

Automatic Orthorectification of High-Resolution Optical Satellite Images Using Vector Roads

This paper presents a completely automatic processing chain for orthorectification of optical pushbroom sensors. The procedure is robust and works without manual intervention from raw satellite image to orthoimage. It is modularly divided in four main steps: metadata extraction, automatic ground control point (GCP) extraction, geometric modeling, and orthorectification. The GCP extraction step uses georeferenced vector roads as a reference and produces a file with a list of points and their accuracy estimation. The physical geometric model is based on collinearity equations and works with sensor-corrected (level 1) optical satellite images. It models the sensor position and attitude with second-order piecewise polynomials depending on the acquisition time. The exterior orientation parameters are estimated in a least squares adjustment, employing random sample consensus and robust estimation algorithms for the removal of erroneous points and fine-tuning of the results. The images are finally orthorectified using a digital elevation model and positioned in a national coordinate system. The usability of the method is presented by testing three RapidEye images of regions with different terrain configurations. Several tests were carried out to verify the efficiency of the procedure and to make it more robust. Using the geometric model, subpixel accuracy on independent check points was achieved, and positional accuracy of orthoimages was around one pixel. The proposed procedure is general and can be easily adapted to various sensors.

Road network extraction: a neural-dynamic framework based on deep learning and a finite state machine

Extracting road networks from very-high-resolution (VHR) aerial and satellite imagery has been a long-standing problem. In this article, a neural-dynamic tracking framework is proposed to extract road networks based on deep convolutional neural networks (DNN) and a finite state machine (FSM). Inspired by autonomous mobile systems, the authors train a DNN to recognize the pattern of input data, which is an image patch extracted in a detection window centred at the current location of the tracker. The pattern is predefined according to the environment and associated with the states in the FSM. A vector-guided sampling method is proposed to generate the training data set for the DNN, which extracts massive image-direction pairs from the imagery and existing vector road maps. In the tracking procedure, the size of the detection window is determined by a fusion strategy and the extracted image patches represent the orientation features of the road (local environment) that can be recognized by the trained DNN. The reactive unit in FSM associates states with behaviours of the tracker while continually modifying the orientation to follow the road and generating a sequence of states and locations. In this way, our framework combines the DNN and FSM. DNN acts as a key component to recognize patterns from a complex and changing environment; FSM translates the recognized patterns to states and controls the behaviour of the tracker. The results illustrate that our approach is more accurate and efficient than the traditional ones.

Automatic Near-Real-Time Image Processing Chain for Very High Resolution Optical Satellite Data

Abstract. In response to the increasing need for automatic and fast satellite image processing SPACE-SI has developed and implemented a fully automatic image processing chain STORM that performs all processing steps from sensor-corrected optical images (level 1) to web-delivered map-ready images and products without operator's intervention. Initial development was tailored to high resolution RapidEye images, and all crucial and most challenging parts of the planned full processing chain were developed: module for automatic image orthorectification based on a physical sensor model and supported by the algorithm for automatic detection of ground control points (GCPs); atmospheric correction module, topographic corrections module that combines physical approach with Minnaert method and utilizing anisotropic illumination model; and modules for high level products generation. Various parts of the chain were implemented also for WorldView-2, THEOS, Pleiades, SPOT 6, Landsat 5-8, and PROBA-V. Support of full-frame sensor currently in development by SPACE-SI is in plan. The proposed paper focuses on the adaptation of the STORM processing chain to very high resolution multispectral images. The development concentrated on the sub-module for automatic detection of GCPs. The initially implemented two-step algorithm that worked only with rasterized vector roads and delivered GCPs with sub-pixel accuracy for the RapidEye images, was improved with the introduction of a third step: super-fine positioning of each GCP based on a reference raster chip. The added step exploits the high spatial resolution of the reference raster to improve the final matching results and to achieve pixel accuracy also on very high resolution optical satellite data.

Road Network Matching Method Based on Particle Swarm Optimization Algorithm

Combined the global optimization ability of particle swarm algorithm and memory capacity of tabu algorithm, this paper proposed an automatic vector road network matching method based on the combination of particle swarm opti- mization and tabu strategy. Firstly, the similarity between node entities is evaluated by means of geometric and topologi- cal characteristics. Then, the basic principle of global optimization of particle swarm optimization is introduced and road matching model based on particle swarm optimization algorithm is designed. Meanwhile, the tabu search algorithm is joined, by using the ability of tabu strategy which expanded the search of the neighborhood. The algorithm fully reflects the climbing feature of tabu strategy, in order to find the global optimal solution of the matching relationship of road network entities. Three different forms of road network data of Wuhan are selected to test our method, the result indicates that the matching method based on the combination of particle swarm optimization and tabu strategy is effective and fea- sible, which can provide a new idea to solve the matching problem.

Road Vector Map Change Monitoring Based on High Resolution Remote Sensing Images

Some studies about road vector map change detection were done in this paper. Firstly, on the basis of old road vector data, the original high resolution remote sensing image was cut into segments. Then, gray analysis and edge extraction of those segments were done so that changes of roads could be detected. Finally, according to the vector data and gray information of roads which were not changed, road templates were extracted and saved automatically. This method was performed on the World View high resolution image of certain parts in the country. The detection result shows that detection correctness is 79.56% and completeness can reach 97.72%. Moreover, the extracted road templates are essentials for the template matching method of road extraction.

A Novel Vehicle Detection Method With High Resolution Highway Aerial Image

A robust and efficient vehicle detection method from high resolution aerial image is still challenging. In this paper, a novel and robust method for automatic vehicle detection using aerial images over highway was presented. In the method, a GIS road vector map was used to constrain the vehicle detection system to the highway networks. After the morphological structure element was identified, we utilized the grayscale opening transformation and grayscale top-hat transformation to identify hypothesis vehicles in the light or white background, and used the grayscale closing transformation and grayscale bot-hat transformation to identify the hypothesis vehicles in the black or dark background. Then, targets with large size or covering a large area were sieved from the hypothesis vehicles using an area threshold that is much larger than a typical vehicle. Targets, whose width is narrower than the diameter of structure element utilized in the grayscale morphological transformation, were smoothed out from the hypothesis vehicles using binary morphological opening transformation. Finally, the hypothesis vehicles detected in both cases were overlaid. It should be noted that in the detection system, a vehicle could be detected twice by the two approaches. The two identical hypothesis vehicles should be amalgamated into a single one for accuracy assessment subsequently. We tested our system on seventeen highway scenes of aerial images with a spatial resolution of 0.15 × 0.15 m. The experimental results showed that the correctness, completeness, and quality rates of the proposed vehicle detection method were about 98%, 93%, and 92%, respectively. Thus, our proposed approach is robust and efficient to detect vehicles of highway using high resolution aerial images.

Extraction of roads in mining area based on Canny edge detection operator

As the main transport corridors of the mine,the roads in the mining area are very important parts of the mining activities. Spatially,the roads are the connection links of the other elements in the mining area,and there are important topological relationships between them: The major road usually connects such elements as the mining surface area( or mine adit),the processing pool,the solid waste mineral,the shed( or buildings) and the tailings. The extraction and recognition of the mine road are helpful to building the surface mining system( SMS), making it easy to monitor the mining area and get automatic interpretation of remote sensing. In order to extract the mine road,the authors used high- resolution remote sensing image and proposed an approach based on Canny edge detection operator for automatic extraction of the road in the mining area in this paper. Firstly,the road edges were extracted by using the Canny detector. Furthermore,the edge matching method was used to determine the road edges,which resulted in the quantization and positioning of the road vector in the study area. The methods proposed in this paper were used to extract the road in the mining area successfully. The results show that the proposed method is of very strong practical applicability.

A Photogrammetric Approach for Assessing Positional Accuracy of OpenStreetMap© Roads

As open source volunteered geographic information continues to gain popularity, the user community and data contributions are expected to grow, e.g., CloudMade, Apple, and Ushahidi now provide OpenStreetMap© (OSM) as a base layer for some of their mapping applications. This, coupled with the lack of cartographic standards and the expectation to one day be able to use this vector data for more geopositionally sensitive applications, like GPS navigation, leaves potential users and researchers to question the accuracy of the database. This research takes a photogrammetric approach to determining the positional accuracy of OSM road features using stereo imagery and a vector adjustment model. The method applies rigorous analytical measurement principles to compute accurate real world geolocations of OSM road vectors. The proposed approach was tested on several urban gridded city streets from the OSM database with the results showing that the post adjusted shape points improved positionally by 86%. Furthermore, the vector adjustment was able to recover 95% of the actual positional displacement present in the database. To demonstrate a practical application, a head-to-head positional accuracy assessment between OSM, the USGS National Map (TNM), and United States Census Bureau’s Topologically Integrated Geographic Encoding Referencing (TIGER) 2007 roads was conducted.

Ad hoc matching of vectorial road networks

In integration of road maps modeled as road vector data, the main task is matching pairs of objects that represent, in different maps, the same segment of a real-world road. In an ad hoc integration, the matching is done for a specific need and, thus, is performed in real time, where only a limited preprocessing is possible. Usually, ad hoc integration is performed as part of some interaction with a user and, hence, the matching algorithm is required to complete its task in time that is short enough for human users to provide feedback to the application, that is, in no more than a few seconds. Such interaction is typical of services on the World Wide Web and to applications in car-navigation systems or in handheld devices. Several algorithms were proposed in the past for matching road vector data; however, these algorithms are not efficient enough for ad hoc integration. This article presents algorithms for ad hoc integration of maps in which roads are represented as polylines. The main novelty of these algorithms is in using only the locations of the endpoints of the polylines rather than trying to match whole lines. The efficiency of the algorithms is shown both analytically and experimentally. In particular, these algorithms do not require the existence of a spatial index, and they are more efficient than an alternative approach based on using a grid index. Extensive experiments using various maps of three different cities show that our approach to matching road networks is efficient and accurate (i.e., it provides high recall and precision). General Terms:Algorithms, Experimentation

Automatic determination of seamlines for aerial image mosaicking based on vector roads alone

Image mosaicking is defined as the registration of two or more images that are then combined into a single image. One of the most difficult steps in the automatic mosaicking of orthoimages is deciding where to place seamlines in overlapping regions. Based on millions of image pixels, existing seamline detection methods mainly focus on how to avoid crossing buildings that are higher than the ground, which results in parallax on the overlapping images. However, various data in vector format, such as vector roads plotted manually and precisely, have not been used to aid the selection of seamlines. This paper presents a novel approach using vector roads alone to generate seamlines, and describes its application to the automatic generation of seamlines for image mosaicking of the city of Wuhan, China. A representative seamline is extracted as follows. First, the skeleton of the overlapping region of adjacent images is extracted after the delineation of boundaries of individual images. Second, vector roads in the overlapping region are overlaid with the extracted skeleton to build a weighted graph G (V, E). Finally, the Floyd–Warshall algorithm is applied to find the lowest cost path from I to O, which refer to two intersections of adjacent image polygons, with the lowest-cost path being the seamline. This vector-based approach is typically more efficient than raster-based approaches. Experiments demonstrate the merits of the proposed approach especially when vector road networks are available.

Aerial Image Mosaicking with the Aid of Vector Roads

This article proposes a novel approach for using vector roads for the generation of seamline networks for aerial image mosaicking. A representative seamline is extracted as follows. First, the straight skeleton of the overlapping area of adjacent images is extracted. Second, vector roads in the overlapping regions are overlaid with the extracted skeleton to build a weighted graph G and the lowest-cost path is considered as a seamline candidate. Finally, the seamline candidate is refined by considering its surrounding pixels. The refined seamline is employed as the final seamline. The resultant seamlines commonly follow the centerlines of roads as much as possible and the mosaic image appears more seamless. Moreover, the vector-based approach is typically more efficient than the existing raster-based approaches: the vector-based approach is nearly ten times faster than a raster-based representative.