Map Generation Process Research Articles

A Review on End-to-End High-Definition Map Generation

Abstract. Autonomous driving offers benefits such as congestion mitigation, increased productivity through the reallocation of driving time, and decreased energy waste. However, achieving Level 4 and 5 autonomous driving remains a significant challenge for both academia and industry. Among the various modules of autonomous driving, High-Definition (HD) maps have become a crucial component due to their high precision in map elements, enabling accurate localization, scene interpretation, navigation, vehicle control and motion forecasting of trajectory of surrounding objects. Several map providers, including TomTom, HERE, Waymo, and NVIDIA, create HD maps for their specific purposes. However, most HD map datasets are not publicly available for individual researchers and companies to investigate the current trends in HD map generation. Furthermore, recent survey papers on HD map generation have tended to focus only on specific aspects, such as road topology or boundary extraction, rather than considering the overall end-to-end HD map generation process. Therefore, we begin with a brief definition, standards, and functionality of HD maps, followed by an exploration of different types of HD maps, including offline and online variants, highlighting their respective advantages and disadvantages. Finally, we will discuss the most recent end-to-end HD map generation architectures, along with various types of open-source HD map datasets and compare their performances.

Sand Cat Swarm Optimizer with Deep Wavelet Autoencoder-based Sign Language Recognition for Hearing- and Speech-impaired Persons

Sign language is commonly used to interact with people who have speech and hearing disorders. Sign language was exploited for interacting with people having developmental impairments who have some or no communication skills. Communication using Sign language has become a fruitful means of interaction for speech- and hearing-impaired people. The hand gesture recognition technique is useful for dumb and deaf people by using convolutional neural networks (CNNs) and human–computer interface for recognizing the static indication of sign language. Therefore, this study presents a new Sand Cat Swarm Optimizer with Deep Wavelet Autoencoder-based Intelligent Sign Language Recognition (SCSO-DWAESLR) technique for hearing- and speech-impaired persons. In the presented SCSO-DWAESLR technique, computer vision and CNN concepts are utilized for identifying sign languages to aid the interaction of hearing- and speech-impaired persons. The SCSO-DWAESLR method makes use of the Inception v3 model for the feature map generation process. In addition, the DWAE classifier is utilized for the recognition and classification of different kinds of signs posed by hearing- and speech-impaired persons. Finally, the hyperparameters related to the DWAE classifier are optimally chosen by using the SCSO algorithm. For exhibiting the effectual recognition outcomes of the SCSO-DWAESLR technique, a detailed experimental analysis was performed. The comparative outcome highlights the superior recognition performance of the SCSO-DWAESLR method over existing techniques under several evaluation metrics.

Multiscale Traffic Sign Detection Method in Complex Environment Based on YOLOv4.

Traffic sign detection is a challenging problem in the field of unmanned driving, particularly important in complex environments. We propose a method, based on the improved You only look once (YOLO) v4, to detect and recognize multiscale traffic signs in complex environments. This method employs an image preprocessing module that can classify and denoize images of complex environments and then input the images into the improved YOLOv4. We also design an improved feature pyramid structure to replace the original feature pyramid of YOLOv4. This structure uses an adaptive feature fusion module and a multiscale feature transfer mechanism to reduce putative information loss in the feature map generation process and improve the information transfer between deep and shallow features, enhancing the representation ability of feature pyramids. Finally, we use EIOU LOSS and Cluster-NMS to further improve the model performance. The experimental results on the fusion of Tsinghua-Tencent 100 K and our collected dataset show that the proposed method achieves an mAP of 81.78%. Compared to existing methods, our method demonstrates its superiority with regard to traffic sign detection.

Collection-CAM: A Faster Region-Based Saliency Method Using Collection-Wise Mask Over Pyramidal Features

Due to the black-box nature of deep networks, making explanations of their decision-making is extremely challenging. A solution is using post-hoc attention mechanisms with the deep network to verify the decision basis. However, those methods have problems such as gradient noise and false confidence. In addition, existing saliency methods either have limited performance by using only the last convolution layer or suffer from large computational overhead. In this work, we propose the Collection-CAM, which generates an attention map with low computational overhead while utilizing multi-level feature maps. First, the Collection-CAM searches for the most appropriate form of the partition through bottom-up clustering and clustering validation process. Then the Collection-CAM applies different pre-processing procedures on the shallow feature map and final feature map to overcome the false positiveness when applied without distinction. By combining collection-wise masks according to their contribution to the confidence score, the Collection-CAM completes the attention map generation process. Experimental results on ImageNet1k, UC Merced, and CUB dataset and various deep network models demonstrate that the Collection-CAM not only can synthesize a saliency map with a better visual explanation but also requires significantly lower computational overhead compared to those of region-based saliency methods.

STATE-WIDE CALCULATION OF TERRAIN-VISUALISATIONS AND AUTOMATIC MAP GENERATION FOR ARCHAEOLOGICAL OBJECTS

Abstract. Airborne laser scanning (ALS) became very popular in the last two decades for archaeological prospection. With the state-wide availability of ALS-data in Lower Saxony, Germany, about 48,000 km², we needed flexible and scalable approaches to process the data. First, we produced a state-wide digital terrain model (DTM) and some visualisations of it to use it in standard GIS software. Some of these visualisations are available as web maps and used for prospection also by volunteers. In a second approach, we automatically generate maps for all known archaeological objects. This is mainly used for the documentation of the 130,000 known objects in Lower Saxony, but also for object-by-object revision of the database. These Maps will also be presented in the web portal “Denkmalatlas Niedersachsen”, an open data imitative of the state Lower Saxony.In the first part of this paper, we show how the state-wide DTM and its visualisations can be calculated using tiles. In the second part, we describe the automatic map generation process. All implementations were done with ArcGIS and its scripting interface ArcPy.

Aircraft detection in remote sensing images based on deconvolution and position attention

ABSTRACT Motivated by the development of deep convolution neural networks (DCNNs), the aircraft detection from remote sensing images has gained tremendous progress. However, due to complex background and multi-scale characteristics, it remains a challenge in remote sensing detection. In this paper, we propose a two-stage aircraft detection method based on deep neural networks, which integrates Deconvolution operation with Position Attention mechanism (DPANet). Specifically, considering that remote sensing images are taken from the top-down perspective, which leads to significant external structural characteristic, we introduce a deconvolution module to capture the external structural feature representation of aircraft during the feature map generation process. Moreover, aiming at reducing the error detections caused by complex background in remote sensing, we propose a position attention module in the second stage. By calculating the feature similarity between any two pixels of the target feature map, DPANet can extract the complicated internal structure feature representation of aircraft, which improve the ability to distinguish background and aircraft. By integrating the deconvolution and position attention modules, DPANet can provide better representation ability for the structural characteristic of aircraft in remote sensing images. Experimental results show that the proposed method can effectively reduce the error detections and improve the accuracy of the aircraft detection.

Generating axial compressor maps to zero speed

Gas turbine performance models typically rely on component maps to characterize engine component performance throughout the operational regime. For the sub-idle case, the lack of reliable rig test data or inability to run design codes far from design conditions entails that component maps have to be generated from the extrapolation of existing data at higher speeds. This undermines the accuracy of whole-engine sub-idle performance models, at times impacting engine development and certification of aviation engines and the accuracy of start-up performance prediction in industrial gas turbines. One of the main components driving this issue is the core compression system, which can present operability concerns during light-up and which also sets the combustor airflow required for ignition. This paper presents, discusses, and draws on previous approaches to describe a method enabling the creation of sub-idle compressor maps from analytical and physical grounds. The method relies on the calculation of zero-speed and torque-free lines to generate a map down to zero speed along with analytical interpolation. A method for the interpolation process is described. A sensitivity study is carried out to assess the effects that different elements of the map generation process may have on the accuracy of the resulting performance calculation. Overall, a method for the generation of accurate, consistent maps from limited geometry data is identified.

A Survey on Transit Map Layout – from Design, Machine, and Human Perspectives

Transit maps are designed to present information for using public transportation systems, such as urban railways. Creating a transit map is a time‐consuming process, which requires iterative information selection, layout design, and usability validation, and thus maps cannot easily be customised or updated frequently. To improve this, scientists investigate fully‐ or semi‐automatic techniques in order to produce high quality transit maps using computers and further examine their corresponding usability. Nonetheless, the quality gap between manually‐drawn maps and machine‐generated maps is still large. To elaborate the current research status, this state‐of‐the‐art report provides an overview of the transit map generation process, primarily from Design, Machine, and Human perspectives. A systematic categorisation is introduced to describe the design pipeline, and an extensive analysis of perspectives is conducted to support the proposed taxonomy. We conclude this survey with a discussion on the current research status, open challenges, and future directions.

Integrating Volcanic Hazard Data in a Systematic Approach to Develop Volcanic Hazard Maps in the Lesser Antilles

We report on the process of generating the first suite of integrated volcanic hazard zonation maps for the islands of Dominica, Grenada (including Kick ‘em Jenny and Ronde/Caille), Nevis, Saba, St. Eustatius, St. Kitts, Saint Lucia and St Vincent in the Lesser Antilles. We developed a systematic approach that accommodated the range in prior knowledge of the volcanoes in the region. A first-order hazard assessment for each island was used to develop one or more scenario(s) of likely future activity, for which scenario-based hazard maps were generated. For the most-likely scenario on each island we also produced a poster-sized integrated volcanic hazard zonation map, which combined the individual hazardous phenomena depicted in the scenario-based hazard maps into integrated hazard zones. We document the philosophy behind the generation of this suite of maps, and the method by which hazard information was combined to create integrated hazard zonation maps, and illustrate our approach through a case study of St. Vincent. We also outline some of the challenges we faced using this approach, and the lessons we have learned by observing how stakeholders have interacted with the maps over the past ~10 years. Based on our experience, we recommend that future map makers involve stakeholders in the entire map generation process, especially when making design choices such as type of base map, use of colour and gradational boundaries, and indeed what to depict on the map. We also recommend careful consideration of how to evaluate and depict offshore hazard of island volcanoes, and recommend computer-assisted modelling of all phenomena to generate more realistic hazard footprints. Finally, although our systematic approach to integrating individual hazard data into zones generally worked well, we suggest that a better approach might be to treat the integration of hazards on a case-by-case basis to ensure the final product meets map users’ needs. We hope that the documentation of our experience might be useful for other map makers to take into account when creating new or updating existing maps.

Visual–inertial navigation for pinpoint planetary landing using scale-based landmark matching

Landing an autonomous spacecraft within 100 m of a mapped target is a navigation challenge in planetary exploration. Vision-based approaches attempt to pair 2D features detected in camera images with 3D mapped landmarks to reach the required precision. This paper presents a vision-aided inertial navigation system for pinpoint planetary landing called LION (Landing Inertial and Optical Navigation). It can fly over any type of terrain, whatever the topography. LION uses measurements from a novel image-to-map matcher in order to update through a tight data fusion scheme the state of an extended Kalman filter propagated with inertial data. The image processing uses the state and covariance predictions from the filter to determine the regions and extraction scales in which to search for non-ambiguous landmarks in the image. The image scale management process operates per landmark and greatly improves the repeatability rate between the map and descent images. A lunar-representative optical test bench called Visilab was also designed in order to test LION. The observability of absolute navigation performances in Visilab is evaluated with a model developed specifically for this purpose. Finally, the system performances are evaluated at a number of altitudes along with its robustness to off-nadir camera angle, illumination changes, a different map generation process and non-planar topography. The error converges to a mean of 4 m and a 3-RMS dispersion of 47 m at 3 km of altitude on the test setup at scale.

Study of Hybrid DNA Physical Mapping Based on Approximation Algorithm with Errors

A human chromosome is a DNA molecule with approximately 10 8 base pairs. The techniques developed to date for sequencing are restricted to pieces of DNA with up to tens of thousands of base pairs. This means that when a piece is sequenced, only an extremely small part of a chromosome can be seen. Molecular biologists use special techniques to deal with DNA molecules comparable in size to a chromosome. These techniques enable them to create maps of an entire chromosome or of significant fractions of chromosomes. Computational techniques were studied that could potentially aid biologists in the map-generation process. An algorithm that solves the consecutive 1s problem was studied. Such a problem is a good model of hybridization mapping when there are no errors and when probes are unique. If errors are present, another approach is needed, and the approximation algorithm is a prospective problem solver for hybridization physical mapping of DNA with errors.

Arctic sea route path planning based on an uncertain ice prediction model

As the global temperature has been increasing, the ice-covered area in the Arctic region has been decreasing. One of the opportunities to take advantage of the situation is the Arctic shipping between Asia and Europe. For economical and safe route selection, an ice navigation system is required for ships operating in the region, since the ice condition is changing hour after hour, and the ships sail long distances for weeks. Therefore, route selection by hand is challenging. The ice model in an ice navigation system simulates dynamic sea ice behavior, and its result is delivered to an optimization model for the path planning domain. During the map generation process for the path planning, both the ice model and the input data for the ice model have uncertainties, which cause prediction uncertainties in the sea ice behavior. For safety, the uncertainties should be considered carefully. Otherwise, accidents in the extremely severe conditions can happen because of unexpected events caused by the uncertainties of the prediction. Notwithstanding the safety issues, uncertainties in the ice navigation system have not been considered in previous research. In this work, the path planning problem in ice-covered waters is presented as a dynamic stochastic path planning problem by generating a map through the ensemble simulation of an ice model. And an uncertainty-based path planning model is proposed to find an optimal route under time-varying stochastic conditions. A heuristic path planning algorithm is employed, and different path planning models (stochastic vs. deterministic) are compared through simulations in two different scenarios for validation of the model.

Geo-Multi-Agent System Based Webmapping Approach Using Multiple Representation and Generalisation Driven by Domain Ontology

Geographic Information (GI) is currently available at any time, from anywhere on the surface of the earth, for any person connected to internet. Some applications of design, implementation, generation and dissemination of maps on the web are recognized as “Webmapping” application, geographic web services or more generally on demand-map making tools. All these web applications aims the satisfaction of user needs by providing personalized maps in a fast response time with a good quality. However, the complexity and diversity of aspects taking into account have lead researchers to focus on one aspect at the expense of others. Consequently, few works have addressed all these aspects simultaneously. The authors propose in this paper, a Webmapping approach organized into two main tasks: (1) query analysis driven by domain ontology in analyzing a query launched by a user on a web browser and (2) map generation process. The first step allows extracting and formalizing user needs through two map determinants factors: the Level of Detail (LoD) and Point of View (PoV) and the second, exploit an hybrid approach “Multi Representation and Generalization” in storing and generating geographical data with integrating Multi-Agent technology in all steps of processing. To evaluate the effectiveness of our proposal, a first tool prototype implementing our approach is so developed using a geographic vector dataset provided by national cartographic agency.

Intuitive approach for necrotic tissue detection

Event Abstract Back to Event Intuitive approach for necrotic tissue detection Petr Maule1*, Vladimir Rohan2 and Jana Kleckova1 1 University of West Bohemia, Department of Computer Science and Engineering, Czechia 2 University Hospital in Pilsen, Czechia Acute stroke is one of the most frequented cause of death and disability in population over 60 years old. We can see high increase in number of studies concerning in acute stroke diagnostics. Diagnostic methods are based on postprocessing of data from acquired examinations, mostly Computed Tomography (CT) and Magnetic Resonance (MR). The highest information value can be gained from perfusion maps which are based on perfusion examination. This examination observes temporal changes of brain tissue densities evocated by contrast material bolus throughpass. The same area of the brain is repeatedly scanned for about 1 minute . Further processing is required resulting in perfusion maps like Cerebral Blood Flow (CBF), Cerebral Blood Volume (CBV), Mean Time to Transmit (MTT), etc.An available treatment method of acute stroke is trombolytical treatment. Article [1] points that mentioned maps may be insufficient for trombolytical treatment application decision. Further it describes method of whole brain perfusion maps generation process using Non Contrast CT examination (NCCT) and Computed Tomography Angiography (CTA) examination. Generated perfusion maps display intensity of necrotic tissue. Necrotic tissue volume of whole brain is the crucial factor for the trombolytical treatment application decision.This arcticle describes new way of necrotic tissue evaluation method. Figure 1 depicts block diagram of the map generation process. The difference against [1] is that analysis (described further) is performed immediately after registra tion process resulting directly in perfusion maps.Necrotic tissue can be characterized by low density increase (or even decrease) between corresponding places of CTA and NCCT. For all points x of resulting maps there are evaluated coefficients c which are proportional to intensity of necrotic tissue at the selected area. It can be described by following formula: c = dCTA(x)-dNCCT(x) where de(x) corresponds to average density of x neighbourhood in the e examination. Considered neighbourhood remains circular area with x as a center and radius of 12 pixels. For averaging purposes there are taken into account only those places where density is between 20 HU and 80 HU. Evaluated coefficient c for place x represents average density difference. These differences are simply depicted using color scale. Only values from -5 HU to +20 HU of the average density difference are displayed, other differences are trimmed to these border values.This new approach for necrotic tissue evaluation is intuitive and gives to doctors clear information about density differences between corresponding places of input examinations. This method is usable for all kinds of input examination's slice thickness. Generated maps are clearly calibrated and border color scale values are adjusted for wider range of necrotic tissue details presentation. INCF-09-31-edit 1 INCF-09-31-3 INCF-09-31-2 INCF-09-31-1

A method to generate near real time UV-Index maps of Austria

Abstract. A method is presented that combines individual ground based ultraviolet (UV) measurements and satellite data from MSG (Meteosat Second Generation) to generate a UV-Index map all over the region of Austria. Cloud modification factors (CMFs) gathered from satellite images give an area wide information of the cloud situation but compared to ground based measurements significant discrepancies have been found. These discrepancies are especially large for high mountain stations but also for low altitude sites differences between satellite derived and ground based data in the order of 15% have been found. To overcome these discrepancies a correction procedure based on the correlation of both data sets at the pixel of the measurement stations is developed. The uncertainty of the final UV-Index map is evaluated exemplarily for the pixel of the measurement sites Bad Vöslau and München by eliminating the information of both sites from the UV-Index map generation process and comparing afterwards calculated and measured results. More than 5000 maps for all weather conditions have been evaluated leading to a mean agreement of 1.02 (standard deviation ±0.31, Bad Vöslau) and 1.01 (standard deviation ±0.34, München) between calculated and measured UV-Indexes.

Focusing on feature-based differences in map comparison

Differences between two maps can be quantified using any cartometric summary; of these, feature-based statistics are potentially more relevant than pixel-based statistics. Distributions of feature areas and shape indices from the workshop maps were compared using a classic information theory index. To place differences in a significance context, uncertainty about cartometrics that arise from both the form of data model and mapping function are considered. Data model characteristics such as raster cell size, raster origin and attribute membership functions can generate variation without the existence of feature differences in the actual surfaces. Stochastic simulation can represent uncertainty due to the mapping function. An understanding of the end-to-end map generation process, absent in the workshop example, can help build a description of its space of uncertainty so that cartometric quantities can be meaningfully compared.

Incremental document map formation: multi-stage approach

The paper presents methodology for the incremental map formation in a multi-stage process of a search engine with the map based user interface1. The architecture of the experimental system allows for comparative evaluation of different constituent technologies for various stages of the process. The quality of the map generation process has been investigated based on a number of clustering and classification measures. Some conclusions concerning the impact of various technological solutions on map quality are presented.

The Uppsala Electron-Density Server

The Uppsala Electron Density Server (EDS; http://eds.bmc.uu.se/) is a web-based facility that provides access to electron-density maps and statistics concerning the fit of crystal structures and their maps. Maps are available for approximately 87% of the crystallographic Protein Data Bank (PDB) entries for which structure factors have been deposited and for which straightforward map calculations succeed in reproducing the published R value to within five percentage points. Here, an account is provided of the methods that are used to generate the information contained in the server. Some of the problems that are encountered in the map-generation process as well as some spin-offs of the project are also discussed.

Map Generation by Cooperative Low-Cost Robots inStructured Unknown Environments

In this paper we present some results obtained with a troupe of low-cost robots designed to cooperatively explore unknown structured orthogonal environments. In order to improve the covering of the explored zone the robots show different behaviours (routine, normal and anxious) and cooperate by transferring each other the perceived environment when they meet; therefore, not all the information of the non-returning robots is lost provided that they had encountered robots that safely returned. The returning robots deliver to a host their perceived and communicated (by other robots) partial maps and the host incrementally generates the most plausible map of the environment. To perform the map generation, a fusion, completion and alignment process of the partial maps, based on fuzzy techniques, has been developed.

Scientific Applications: BRAD

A multiconsole computer display system has been designed that provides very rich displays at low unit cost. Each BRAD (Brookhaven RAster Display) console can plot tens of thousands of points, or up to 4000 characters at 30 frames per second. After an initial display system investment of $50,000, each display, with teletype, cost less than $3,000. The technique employed is that of programmatically generating a binary image of the desired display in a computer. The image is written on a rotating drum memory. Independent read heads continuously display the picture, which is generated by swept horizontal lines. A standard TV monitor serves as the display device. The technique has two drawbacks. A computer must compute any image to be displayed. Also, the “pointing” interaction is more difficult. This is because the pointing function gives only the coordinates of the point on the screen. The inverse of the map generation process is required to calculate the coordinates at the input space.