Object Oriented Image Analysis in Remote Sensing of Forest and Vineyard Areas

Abstract

The study of vegetation cover, forests, orchards or vineyards and crops through satellite techniques is increasingly promoted as a result of facilities they offer. Since 2010 until today, it is launched over 50 satellite platforms, delivering images of the Earth's surface with different spectral, spatial and radiometric characteristics. New satellite images such as RapidEye, WorldView2 or WorldView3, with its high spatial and radiometric resolution, prevents the use of the standard image analysis techniques (such as supervised or unsupervised), and involves the use of modern methods in image analysis, such as object base image analysis. A number of methods and techniques for processing and analysis of satellite images are developed to increase the precision of the working, given the diversity of vegetation structure analysis and expected results. This study aimed to analyze the capabilities of object-oriented image analysis (OBIA) for recognition forest and vineyard areas. OBIA is automated process of object extraction by modelling of human visual system for image interpretation. The basis for classification process is object, which is created according to the set of characteristics. In object-oriented approach classification description is based on classification rules including spectral characteristics, size, shape, as well as content and texture information. Analysis is done on multispectral imagery of high and very high spatial resolution. Represented results show the usefulness of RapidEye and WorldView2 images as well as importance of classification based on OBIA. Object-oriented image analysis (OBIA) method based on satellite imagery has facilitated the recognition forest and vineyard areas with high accuracy.

Mapping Floods in Lowland Forest Using Sentinel-1 and Sentinel-2 Data and an Object-Based Approach

“…It is the process of dividing imagery into segments or objects with a certain number of pixels. Each object created in the segmentation process is characterized by shape, size, color, and a topological relationship with neighboring objects [64]. Each object also contains attribute features based on spectral, geometric, textural, and contextual properties [14].…”

Evaluation the Microelements and Heavy Metals Content in the Conditions of Some Agricultural and Forest Lands

“…For the assessment of soils and lands, in agricultural or natural ecosystems, both classical methods, based on soil samples and laboratory analyses, as well as alternative methods, based on imaging, were used (Syso et al, 2014;Govedarica et al, 2015;Popescu et al, 2020).…”

Model for Estimating Rapeseed Production Based on Remote Sensing

“…( Within agricultural ecosystems, the need to estimate agricultural production is an important aspect, in relation to cultivated plants and genotypes (Constantinescu et al, 2018;Shook et al, 2021), culture technologies (Chlingaryan et al, 2018;Najafi et al, 2018), fertilization (Sala and Boldea, 2011;Sala et al, 2016;Szulc et al, 2021), category of agricultural products (Kuehne et al, 2017;Kurumatani, 2020), storage and capitalization of production (Tașkiner and Bilgen, 2021), product market (Jiang et al, 2020;Kamath et al, 2021), the profitability of farmers (Sangeeta, 2020;Kamath et al, 2021) Different methods and models of production prediction have been proposed and developed to estimate agricultural production and yields in relation to crops, production factors, environmental factors and their interaction (Beres et al, 2020; Meng et al, 2021). Methods and models based on remote sensing are very useful and used for the analysis and characterization of natural areas of agroecosystems, with a high level of estimation of agricultural production (Govedarica et al, 2015;Awad, 2019;Popescu et al, 2020;Khaki et al, 2021).…”