George Mitri

A performance evaluation of a burned area object-based classification model when applied to topographically and non-topographically corrected TM imagery

Operational use of remote sensing as a tool for post-fire Mediterranean forest management has been limited by problems of classification accuracy arising from confusion between burned and non-burned land, especially within shaded areas. Object-oriented image analysis has been developed to overcome the limitations and weaknesses of traditional image processing methods for feature extraction from high spatial resolution images. The aim of this work was to evaluate the performance of an object-based classification model developed for burned area mapping, when applied to topographically and non-topographically corrected Landsat Thematic Mapper (TM) imagery for a site on the Greek island of Thasos. The image was atmospherically and geometrically corrected before object-based classification. The results were compared with the forest perimeter map generated by the Forest Service. The accuracy assessment using an error matrix indicated that the removal of topographic effects from the image before applying the object-based classification model resulted in only slightly more accurate mapping of the burned area (1.16% increase in accuracy). It was concluded that topographic correction is not essential prior to object-based classification of a burned Mediterranean landscape using TM data.

A semi-automated object-oriented model for burned area mapping in the Mediterranean region using Landsat-TM imagery

Pixel-based classification methods that make use of the spectral information derived from satellite images have been repeatedly reported to create confusion between burned areas and non-vegetation categories, especially water bodies and shaded areas. As a result of the aforementioned, these methods cannot be used on an operational basis for mapping burned areas using satellite images. On the other hand, object-based image classification allows the integration of a broad spectrum of different object features, such as spectral values, shape and texture. Sophisticated classification, incorporating contextual and semantic information, can be performed by using not only image object attributes, but also the relationship between networked image objects. In this study, the synergy of all these features allowed us to address image analysis tasks that, up until now, have not been possible. The aim of this work was to develop an object-based classification model for burned area mapping in the Mediterranean using Landsat-TM imagery. The object-oriented model developed to map a burned area on the Greek island of Thasos was then used to map other burned areas in the Mediterranean region after the Landsat-TM images had been radiometrically, geometrically and topographically corrected. The results of the research showed that the developed object-oriented model was transferable and that it could be effectively used as an operative tool for identifying and mapping the three different burned areas (~98% overall accuracy).

Fire type mapping using object-based classification of Ikonos imagery

Distinguishing and mapping areas of surface and crown fire spread has significant applications in the study of fire behaviour, fire suppression, and fire effects. Satellite remote sensing has supplied a suitable alternative to conventional techniques for mapping the extent of burned areas, as well as for providing post-fire related information (such as the type and severity of burn). The aim of the present study was to develop an object-based classification model for mapping the type of fire using very high spatial resolution imagery (Ikonos). The specific objectives were: (i) to distinguish between surface burn and canopy burn; and (ii) to assess the accuracy of the classification results by employing field survey data. The methodology involved two consecutive steps, namely image segmentation and image classification. First, image objects were extracted at different scales using multi-resolution segmentation procedures, and then both spectral and contextual object information was employed to classify the objects. The accuracy assessment revealed very promising results (approximately 87% overall accuracy with a Kappa Index of Agreement of 0.74). Classification accuracy was mainly affected by the density of the canopy. This could be attributed to the inability of the optical sensors to penetrate dense canopy to detect fire-affected areas. The main conclusion drawn in the present study is that object-oriented classification can be used to accurately distinguish and map areas of surface and crown fire spread, especially those occurring in open Mediterranean forests.

Mapping post-fire forest regeneration and vegetation recovery using a combination of very high spatial resolution and hyperspectral satellite imagery

Careful evaluation of forest regeneration and vegetation recovery after a fire event provides vital information useful in land management. The use of remotely sensed data is considered to be especially suitable for monitoring ecosystem dynamics after fire. The aim of this work was to map post-fire forest regeneration and vegetation recovery on the Mediterranean island of Thasos by using a combination of very high spatial (VHS) resolution (QuickBird) and hyperspectral (EO-1 Hyperion) imagery and by employing object-based image analysis. More specifically, the work focused on (1) the separation and mapping of three major post-fire classes (forest regeneration, other vegetation recovery, unburned vegetation) existing within the fire perimeter, and (2) the differentiation and mapping of the two main forest regeneration classes, namely, Pinus brutia regeneration, and Pinus nigra regeneration. The data used in this study consisted of satellite images and field observations of homogeneous regenerated and revegetated areas. The methodology followed two main steps: a three-level image segmentation, and, a classification of the segmented images. The process resulted in the separation of classes related to the aforementioned objectives. The overall accuracy assessment revealed very promising results (approximately 83.7% overall accuracy, with a Kappa Index of Agreement of 0.79). The achieved accuracy was 8% higher when compared to the results reported in a previous work in which only the EO-1 Hyperion image was employed in order to map the same classes. Some classification confusions involving the classes of P. brutia regeneration and P. nigra regeneration were observed. This could be attributed to the absence of large and dense homogeneous areas of regenerated pine trees in the study area.

Mapping Postfire Vegetation Recovery Using EO-1 Hyperion Imagery

The aim of this paper is to investigate whether it is possible to accurately map postfire vegetation recovery on the Mediterranean island of Thasos by employing Earth Observing-1 (EO-1) Hyperion imagery and object-based classification. Specific objectives include the following: 1) locating and mapping areas of forest regeneration and other vegetation recovery and distinguishing among them; 2) distinguishing between Pinus brutia regeneration and Pinus nigra regeneration within the area of forest regeneration; and 3) examining whether it is possible to distinguish between areas of forest regeneration (Pinus brutia, Pinus nigra) and mature forest. The data used in this study consist of satellite images, field-spectroradiometry measurements, and field observations of the homogenous revegetated areas. The methodology comprised four consecutive steps. The first step involved preprocessing of the Hyperion image and field data. Subsequently, an object-oriented model was developed, which involved three steps, namely, image segmentation, object training, and object classification. The process resulted in the separation of five classes (¿brutia mature,¿ ¿ nigra mature,¿ ¿brutia regeneration,¿ ¿nigra regeneration,¿ and ¿other vegetation¿). The accuracy assessment revealed very promising results (approximately 75.81% overall accuracy, with a Kappa Index of Agreement of 0.689). Some classification confusion involving the classes of Pinus brutia regeneration and Pinus nigra regeneration was recorded. This could be attributed to the absence of large homogenous areas of regenerated pine trees. The main conclusion drawn in this paper was that object-based classification can be used to accurately map postfire vegetation recovery using EO-1 Hyperion imagery.

Mapping the severity of fire using object-based classification of IKONOS imagery

Mapping fire severity is necessary in order (1) to locate areas in need of special or intense post-fire management; (2) to allow the study of fire impact and vegetation recovery; and (3) to validate fire risk and fire behaviour models. The present study aimed to develop a method to map the severity of fire by employing post-fire IKONOS imagery. The objective was to develop an object-oriented model that would distinguish between different degrees of fire severity and to assess the accuracy of the model by employing field-collected data. The work comprised five main consecutive steps, namely, field data collection, data preprocessing, image segmentation, image classification and accuracy assessment. An adapted version of the FIREMON (fire effects monitoring and inventory protocol) landscape assessment method was employed to quantitatively record fire severity in the field. As a result, two different datasets were used: one for training and developing the classification rules, and another one for assessing the accuracy of the classification. Separate and independent data were used for training and for accuracy assessment. Overall accuracy was estimated to be 83%, while the Kappa Index of Agreement obtained was 0.74. The main source of inaccuracy was the inability of IKONOS to penetrate the dense canopy of unburned vegetation. The main conclusion drawn from the present work was that object-based classification applied to IKONOS imagery has the potential to produce accurate maps of fire severity, especially in the case of the open Mediterranean forest.

Advances in Remote Sensing of Post-Fire Vegetation Recovery Monitoring - A Review

Ioannis Gitas1, George Mitri2, Sander Veraverbeke3,4 and Anastasia Polychronaki1 1Laboratory of Forest Management and Remote Sensing, Aristotle University of Thessaloniki, Thessaloniki, 2Biodiversity Program, Institute of the Environment, University of Balamand and Department of Environmental Sciences, Faculty of Science, University of Balamand, 3Department of Geography, Ghent University, Ghent 4Jet Propulsion Laboratory, California Institute of Technology,Pasadena, CA, 1Greece 2Lebanon 3Belgium 4USA

Efficient forest fire occurrence prediction for developing countries using two weather parameters

Forest fire occurrence prediction plays a major role in resource allocation, mitigation and recovery efforts. This paper compares two artificial intelligence based methods, artificial neural networks (ANN) and support vector machines (SVM), utilizing a reduced set of weather parameters. Using a reduced set of parameters results in an efficient and reduced cost prediction system especially for developing countries. In this paper the aim is to predict forest fire occurrence by reducing the number of monitored features, and eliminating the need for weather prediction mechanisms. The reason is to reduce errors due to inaccuracies in weather prediction. The challenge is to choose a limited number of easily measurable features in the aim of reducing the cost of the system and its maintenance. At the same time, the chosen features must have a high correlation with the risk of fire occurrence. A literature review of forest fire prediction methods divided into systems/indices, and artificial intelligence is provided. The two fire danger prediction algorithms utilize relative humidity and cumulative precipitation to output a risk estimate. The assessment of these algorithms, using data from Lebanon, demonstrated their ability to accurately predict the risk of fire occurrence on a scale of four levels.

Artificial intelligence for forest fire prediction

Forest fire prediction constitutes a significant component of forest fire management. It plays a major role in resource allocation, mitigation and recovery efforts. This paper presents a description and analysis of forest fire prediction methods based on artificial intelligence. A novel forest fire risk prediction algorithm, based on support vector machines, is presented. The algorithm depends on previous weather conditions in order to predict the fire hazard level of a day. The implementation of the algorithm using data from Lebanon demonstrated its ability to accurately predict the hazard of fire occurrence.

Remote Sensing of Burn Severity

Accurate information relating to the impact of fire on the environment and the way it is distributed throughout the burned area is critical for short-term mitigation and rehabilitation treatments. The aim of this chapter is to review the role of Remote Sensing (RS) in estimating burn severity. Initially, the terminology was clarified since two different terms are usually used in the literature: fire and burn severity. Methods and techniques that have been employed to estimate burn severity on the ground and by RS were reviewed. Special attention was paid to the different type of sensors that are used in previously conducted studies. Finally, the future trends in researches related to remote sensing of burn severity were identified.