The main goal of this study is to evaluate different models for further improvement of the accuracy of land use and land cover (LULC) classification on Google Earth Engine using random forest (RF) and support vector machine (SVM) learning algorithms. Ten indices, namely normalized difference vegetation index, normalized difference soil index, index-based built-up index, biophysical composition index, built-up area extraction index (BAEI), urban index, new built-up index, band ratio for built-up area, bare soil index, and normalized built up area index, were used as input parameters for the machine learning algorithms to improve classification accuracy. The combinatorial analysis of the Sentinel-2 bands and the aforementioned indices allowed us to create four combinations based on surface reflectance characteristics. The study includes data from April 2020 to September 2021 and April 2022 to June 2022. The multitemporal Sentinel-2 data with spatial resolutions of 10 m were used to determine the LULC classification. The major land use classes such as water, forest, grassland, urban areas, and other lands were obtained. Generally, the RF algorithm showed higher accuracy than the SVM. The overall accuracy for RF and SVM was 86.56% and 84.48%, respectively, and the mean Kappa was 0.82 and 0.79, respectively. Using the combination 2 with the RF algorithm and combination 4 with the SVM algorithm for LULC classification was more accurate. The additional use of vegetation indices allowed to increase in the accuracy of LULC classification and separate classes with similar reflection spectra.
Кандидат технічних наук, молодший науковий співробітник відділу геоінформаційних технологій в дистанційному зондуванні Землі (ГІТ в ДЗЗ), orcid.org/0000-0002-7284-6502 Науковий Центр аерокосмічних досліджень Землі ІГН НАН України, Київ НОВИЙ ПІДХІД ДО ЗАСТОСУВАННЯ ПРАВИЛА ДИСКОНТУВАННЯ ПРИ КЛАСИФІКУВАННІ ГІПЕРСПЕКТРАЛЬНИХ КОСМІЧНИХ ЗОБРАЖЕНЬ Анотація. На сьогодні об'єднання інформації є однією із найбільш важливих процедур при класифікуванні гіперспектральних космічних зображень. Метою об'єднання інформації є спрощення даних, отриманих із різних джерел. Багато відомих методів об'єднання включають знаходження середнього арифметичного, середнього геометричного, максимального та мінімального значень. Правила комбінування є особливим типом методів об'єднання даних, отриманих із різних джерел. Ці джерела надають різні оцінки одним і тим же гіпотезам. Вимога щодо незалежності всіх джерел інформації є дуже важливим питанням. Опрацювання суперечливої інформації та комбінування суперечливих даних є дуже складною проблемою в задачах класифікування. Але багато відомих правил комбінування дають неправильні результати за наявності досить суперечливих частин свідчення. Відомі правила комбінування більше акцентують увагу на узгоджених джерелах інформації та ігнорують усі суперечливі частини свідчення. Ці правила не працюють за наявності досить суперечливих даних. Ось чому комбінування суперечливих частин свідчення є найбільш важливим питанням у дистанційному зондуванні Землі. У статті пропонується правило дисконтування для роботи із суперечливими джерелами інформації. Застосовуючи правило дисконтування, спочатку можна дисконтувати джерела, а потім скомбінувати результуючі базові маси за допомогою будь-якого відомого правила комбінування, використовуючи коефіцієнт дисконтування. Цей коефіцієнт дисконтування враховує абсолютну надійність джерел. Абсолютна надійність припускає, що ми можемо розрізняти джерела даних за надійністю і можемо виразити математично відмінності між різними джерелами. Також було зазначено, що правило дисконтування надає ненульову базову масу фрейму розрізнення. Ця процедура не змінює початкової інформації. Також розглянуто приклад застосування правила дисконтування для класифікування космічних зображень. Описане правило дисконтування може бути застосоване при класифікуванні лісів, при пошуку корисних копалин та розв'язку різноманітних екологічних і тематичних завдань. Ключові слова: гіперспектральне космічне зображення; правило дисконтування; класифікування зображень; базова маса Alpert Sofiia PhD (Eng.
Nowadays unmanned aerial vehicles (drones) are applied for solution numerous remote sensing tasks. They give a new opportunites for conducting environmental monitoring and give images with a very high resolution. Unmanned aerial vehicles are applied for solution numerous agricultural problems. They give a detail picture of fields. Unmanned aerial vehicles are used to help increase crop production. With technology constantly improving, imaging of the crops will need to improve as well.Digital images obtained by unmanned aerial vehicles (drones) can be used in forestry, they are used for environmental monitoring, plant health assessment and analysis of natural disasters. Unmanned aerial vehicles are also used for mining, they are applied for mapping deposit sites, exploring for oil and gas, surveying mines.Laser gyroscope is an essential component of a drones flight control system. Laser gyroscopes provides orientation control of drone and essential navigation information to the central flight control systems. Laser gyroscopes provide navigation information to the flight controller, which make drones easier and safer to fly. Laser gyroscope is one of the most important components, allowing the drone to fly smooth even in strong winds. The smooth flight capabilities allows us to get images with high precision.Nowadays the main function of gyroscope technologies is to improve the unmanned aerial vehicles flight capabilities. It was described a structure and main characteristics of laser gyroscopes. It was noted, that laser gyroscope is operated on the principle of the Sagnac effect. Sagnac effect is a phenomenon encountered in interferometry that is elicited by rotation. It were described main advantages and disadvantages of laser gyroscopes. A comparative analysis of mechanical and laser gyroscopes was carried out too.It also was noted, that laser gyroscopes are applied in different areas, such as: inertial navigation systems, aircraft, ships, unmanned aerial vehicles (drones) and satellites. Nowadays laser technology is developed further. There are all prerequisites for improving the precision and other technical characteristics of laser gyroscopes.
Due to modern microcomputers and platforms based on microprocessors such as, for example, Raspberry Pi, Orange Pi, Nano Pi, Rock Pi, Banana Pi, Asus Tinker Board – the development of prototypes of em-bedded systems is possible in a «design» mode. The software part is implemented on the basis of operat-ing systems and standard technologies based on well-known programming languages such as C / C++, Python, C#, Java, etc. In such case the control channel for the embedded system can be either imple-mented via a web service separated by a communication channel or controlled independently. It is im-portant to understand that creating an embedded system on a standard platform is much more expensive than buying a ready-made mass-produced device with the same functionality. Therefore, it makes sense to use platforms like the Raspberry Pi mainly for individual artificial devices. If it is necessary to build a project of embedded systems and there is a problem with choosing a hardware platform for the client side, then currently there is a wide range of boards and solutions for building an efficient and inexpen-sive system using ready-made modules. The number of expansion cards and various sensors, video cam-eras, internet connection via Ethernet, Wi-Fi and Bluetooth provides a wide range of opportunities for building almost any solution based on this component base. The foundation can be made within a small budget, with minimal time spent, using large blocks and ready-made libraries for programming embed-ded systems. This article presents the results of research and development work on the creation of a software and hardware infrastructure of a terrestrial platform with the elements of artificial intelligence. Based on the actual results of the research, a deployment diagram and a component diagram of such an infrastructure have been constructed.
Задачі підсуNowadays with the rapid development of information technologies, UAV-based Remote Sensing (drone remote sensing) gives a new opportunities for conducting scientific research in a much more detail way. UAVs (unmanned aerial vehicles) give the opportunity to acquire data at sufficiently low cost. They also provide remote data more rapidly than piloted aerial vehicles. Nowadays drones are often used, because application of piloted aerial vehicles can be dangerous, difficult and expensive for some territories. Application of low altitude UAVs give a possibility to achieve images with a very high resolution and sufficient precision. In this article structure and main details of drones were considered. It also was noted, that technologies of UAV-based Remote Sensing are used in different areas.Agricultural drones help to analyze crops, make decisions on how to use the crop information and take the necessary actions to correct the problems. These unmanned aerial vehicles let to see fields from the sky. Agricultural drones are used to help increase crop production and monitor crop growth. Drones and sensors give a detail picture of fields. They can survey the fields periodically. Agricultural drones can reveal many issues such as soil variation, pest infestations and changes in the crops over time. They also show differences between healthy and unhealthy plants. Drones are flied over the crops and help to make decisions on how to proceed given the crop information. Nowadays there is a large capacity for growth in the area of agricultural unmanned aerial vehicles. With technology constantly improving, imaging of the crops will need to improve as well.Drones are used for exploring for minerals and mapping deposit sites, they are used in the oil and gas industry for remote monitoring. Drones can provide information of nature disasters and give help to assess property damage. They help to conduct forest monitoring and to assess plant health. Unmanned aerial vehicles are also used in a military capacity and ecological monitoring. It also was noted, that there is a large capacity for development and improvement of unmanned aerial vehicles.путникового моніторингу в аерокосмічному комплексі
Multispectral remote sensing is one of the most popular techniques in the earth observation, because this technique can provide information of ground objects on Earth’s surface using hundreds of narrow bands. However, multispectral images produces a very large volume of data. Processing the huge volume of information is one of most important and actual problems of remote sensing. The rapid development of the remote sensing demand to develop the data processing algorithms. But at present data processing techniques cannot give accurate results. If we use traditional methods to process multispectral images, the volume of the data increases. The main goal of the band selection is to choose the optimal combination of spectral bands for the solution of the particular remote sensing task. This process is important because different bands are sensitive to different objects. Selecting the right bands can help to optimize the detection of different ground objects. Some spectral bands are more sensitive to minerals, while others are more sensitive to vegetation or water bodies. Under a small number of training samples, the classification accuracy of multispectral images decreases when the volume of multispectral data increases. Usually adjacent bands are highly correlated, and some spectral bands may not carry unique information. That’s why it is necessarily to reduce the dimensionality of multispectral data. It helps to store, process, transmit information more efficiently and to reduce the computational costs while processing images. The different modern methods of multispectral band selection are also considered and analyzed in this work. It also is proposed a new method to select spectral bands, which is based on the concept of criterion function of information capability of spectral bands. In this article some examples using criterion function of information capability are considered too.
Nowadays solution of different scientific problems using satellite images, generally includes a classification procedure. Classification is one of the most important procedures used in remote sensing, because it involves a lot of mathematical operations and data preprocessing. The processing of information and combining of conflicting data is a very difficult problem in classification tasks. Nowadays many classification methods are applied in remote sensing. Classification of conflicting data has been a key problem, both from a theoretical and practical point of view. But a lot of known classification methods can not deal with highly conflicted data and uncertainty. The main purpose of this article is to apply proportional conflict redistribution rule (PRC5) for satellite image classification in conditions of uncertainty, when conflicting sources of evidence give incomplete and vague information. This rule can process conflicting data and combine conflicting bodies of evidence (spectral bands). Proportional conflict redistribution rule can redistribute the partial conflicting mass proportionally on non-empty sets involved in the conflict. It was noticed, that this rule can provide a construction of aggregated estimate under conflict. It calculates all partial conflicting masses separately. It was also shown, that proportional conflict redistribution rule is the most mathematically exact redistribution of conflicting mass to non-empty set. But this rule consists of difficult calculation procedures. The more hypotheses and more masses are involved in the fusion, the more difficult is to implement proportional conflict redistribution rule, therefore special computer software should be used. It was considered an example of practical use of the proposed conflict redistribution rule. It also was noticed, that this new approach to the application of conflict redistribution rule in satellite image classification can be applied for analysis of satellite images, solving practical and ecological tasks, assessment of agricultural lands, classification of forests, in searching for oil and gas.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.