Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 3

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: building detection

Sortuj według:

Ogranicz wyniki do:

Building Semantic Segmentation Using UNet Convolutional Network on SpaceNet Public Data Sets for Monitoring Surrounding Area of Chan Chan (Peru)

Chicchon Miguel, Malinverni Eva Savina, Sanità Marsia, Pierdicca Roberto, Colosi Francesca, Trujillo Francisco James León

Geomatics and Environmental Engineering

2024

Vol. 18, no. 3

25--43

The amount of damage to cultural heritage sites is increasing rapidly every year. This is due to inadequate heritage management and uncontrolled urban growth as well as unpredictable seismic and atmospheric events that manifest themselves in a continuously deteriorating ecosystem. Thus, applications of artificial intelligence (AI) in remote-sensing (RS) techniques (machine-learning and deep-learning algorithms) for monitoring archaeological sites have increased in recent years. This research involves the surrounding area of the archaeological site of Chan Chan in Peru in particular. An approach that is based on the use of AI algorithms for building footprint segmentation and changedetection analysis by means of RS images is proposed. It involves a UNet convolutional network based on an EfficientNet B0 to B7 encoder. The network was trained on two public data sets from SpaceNet that were based on WV2 and WV3 satellite images: SpaceNet V1 (Rio), and SpaceNet V2 (Shanghai). In the pre-processing phase, the images from the two data sets have been equalized in order to improve their quality and avoid overfitting. The building segmentation has been performed on HRV images of the study area that were downloaded from Google Earth Pro. The value that was achieved in the IoU metric was around 70% in both experiments. The purpose of this proposed methodology is to assist scientists in drafting monitoring and conservation protocols based on already-recorded data in order to prevent future disasters and hazards.

Conditions and limitations of digital satellite image pre-processing for the further 3D modeling

Hnatushenko V. V., Kavats O. O., Kubanek M., Kibukevych Y. O.

Journal of Applied Mathematics and Computational Mechanics

2016

Vol. 15, nr 3

57-65

Today, 3D models of complex urban buildings are one of the most popular applications of 3D modeling. 3D models of complex urban buildings provide high data interpretation that accurately transfers information about objects or area changes and allows one to solve a number of applied tasks. The quality of the 3D models depends on the quality of the initial images and the method of the object recognition. First of all, the 3D-model building requires identification and classification building borders, which requires determination of the building roof form. The article reviews the existing classification and recognition methods for the 3D further modeling.

Wykrywanie budynków na podstawie lotniczego skanowania laserowego

Bucior M., Borowiec N.;, Jędrychowski I., Pyka K.

Roczniki Geomatyki

2006

T. 4, z. 3

57-70

This paper discusses automatic detection of buildings from airborne laser scanner data. Beside introduction and conclusions there are three main parts in this paper. In part one basic technical parameters of airborne laser scanning are reminded. Part two presents literature review of various methods that have been applied in the detection and modeling of buildings. Part three describes a research experiment carried out by the authors. This part includes a comparison between two methods of detection: the one offered by specialist software and the alternative method proposed by the authors of this paper. The technique of laser scanning, often referred to as LIDAR, continues to develop very dynamically. It is characterized by a high level of efficiency and accuracy. It is most often used to create 3D models of cities. Until now, LIDAR was mostly used in national studies to determine digital terrain models (DTM), which is done by separating certain points (those which result from laser reflections of trees, buildings and other above-ground surfaces) from disorganized .clouds of points.. Meanwhile, the most useful contribution of this technique is that it enables numeric calculation of the digital surface model (DSM). The authors. experiment attempted to analyze the effectiveness of automatic detection of buildings using two different methods. The first method used original data and applied specialist software which detects and models buildings. In the second, the .cloud of points. was replaced by a regular grid, which had been determined through interpolation. Then, using the typical tool of GIS, the authors carried out a series of experiments. In this paper, the authors present their concept of detection of buildings. This concept is based on an analysis of three surface layers: map of heights, map of slopes and map of texture. The final stage consisted of spatial analysis which showed all the places which meet certain conditions that are adequate for buildings, such as heights, slopes and texture. The methods were implemented on two test areas. One area contained independently standing apartment buildings in which the sides and rooftops of buildings were perpendicular and at right angles to each other. The second test area was made up of various buildings of differentiated heights with steep, multidirectional roofs. For both these areas, reference data was obtained through the vectorization of photogrammetric stereoscopic models. Both methods of detection showed comparable effectiveness. The method using .cloud of points. and specialist software showed slightly straighter roof edges, however a slightly worse balance of surface in relation to the reference data, than the method based on GIS analyses which presents the authors. concepts of detections of buildings. However, the differences were negligible and both methods had a similar level of effectiveness in the detection of buildings: approximately 90% for the easy area and about 60% for difficult area. These results are similar to those presented in literature. During the study, all cases in which detection of buildings was ineffective were also analyzed. Tall trees rising above rooftops often presented a significant obstacle. Moreover, the scanning data contained several places, where LIDAR provided measurements with very low density, much smaller than the average density of 1,5 points per m2. These .holes. lowered the effectiveness of the first method. However, the weakness of the raster method was weak representation of the grid in places where trees were located as the applied interpolation smoothed out the original data. The results of this research lead to the conclusion that an optimal method would entail a .combined. approach. First, the raster analysis should be applied to determine the probable location of buildings. Then, for certain atypical spaces one should return to the source data (cloud of points) and vertically assign cross sections in predefined directions. What is still needed is a method of automatic recognition of buildings on the basis of cross sections as well as dimensions of buildings which aim to obtain a 3D model. This paper confirms a huge potential of the laser scanning technique to create 3D models. The proposed method of detection of buildings proved promising and it can be applied even without expensive specialized software.