Data combination method in Remote Sensing tasks in case of conflicting information sources

• Дата: 2021
• Автор: Alpert, Sofiia
• Формат: Стаття
• Мова: English
• Опубліковано: Scientific Centre for Aerospace Research of the Earth Institute of Geological Science National Academy of Sciences of Ukraine, Kyiv, Ukraine 2021
• Теми: evidence theory; Jaccard coefficient; basic probability assignment; combination rules; conflicting evidence; теория свидетельств; коэффициент Жаккара; базовые массы; правила комбинирования; противоречивые свидетельства; теорія свідчень; коефіцієнт Жаккара; базові маси; правила комбінування; суперечливі свідчення
• Онлайн доступ: https://ujrs.org.ua/ujrs/article/view/201
• Назва журналу: Ukrainian Journal of Remote Sensing of the Earth

Nowadays technologies of UAV-based Remote Sensing are used in different areas, such as: ecological monitoring, agriculture tasks, exploring for minerals, oil and gas, forest monitoring and warfare. Drones provide information more rapidly than piloted aerial vehicles and give images of a very high resolution, sufficiently low cost and high precision.Let’s note, that processing of conflicting information is the most important task in remote sensing. Dempster’s rule of data combination is widely used in solution of different remote sensing tasks, because it can processes incomplete and vague information. However, Dempster’s rule has some disadvantage, it can not deal with highly conflicted data. This rule of data combination yields wrong results, when bodies of evidence highly conflict with each other. That’s why it was proposed a data combination method in UAV-based Remote Sensing. This method has several important advantages: simple calculation and high accuracy. In this paper data combination method based on application of Jaccard coefficient and Dempster’s rule of combination is proposed. The described method can deal with conflicting sources of information. This data combination method based on application of evidence theory and Jaccard coefficient takes into consideration the associative relationship of the evidences and can efficiently handle highly conflicting sources of data (spectral bands).The frequency approach to determine basic probability assignment and formula to determine Jaccard coefficient are described in this paper too. Jaccard coefficient is defined as the size of the intersection divided by the size of the union of the sample sets. Jaccard coefficient measures similarity between finite sets. Some numerical examples of calculation of Jaccard coefficient and basic probability assignments are considered in this work too.This data combination method based on application of Jaccard coefficient and Dempster’s rule of combination can be applied in exploring for minerals, different agricultural, practical and ecological tasks.