No abstract

The requirements for advanced knowledge on forest resources have led researchers to develop efficient methods to provide detailed information about trees. Since 1999, orbital remote sensing has been providing very high resolution (VHR) image data. The new generation of satellite allows individual tree crowns to be visually identifiable. The increase in spatial resolution has also had a profound effect in image processing techniques and has motivated the development of new object-based procedures to extract information. Tree crown detection has become a major area of research in image analysis considering the complex nature of trees in an uncontrolled environment. This chapter is subdivided into two parts. Part I offers an overview of the state of the art in computer detection of individual tree crowns in VHR images. Part II presents a new hybrid approach developed by the authors that integrates geometrical-optical modeling (GOM), marked point processes (MPP), and template matching (TM) to individually detect tree crowns in VHR images. The method is presented for two different applications: isolated tree detection in an urban environment and automatic tree counting in orchards with an average performance rate of 82% for tree detection and above 90% for tree counting in orchards.

Cerrado is a savannah biome covering about 60% of the State of Minas Gerais, Brazil, but with a rate of conversion that surpasses that of all other biomes in Brazil. Remote sensing is the only practical means of monitoring the conversion and regeneration of this vegetation formation. The objective of this article is to model the process of regeneration of Cerrado vegetation and to estimate its age using a RapidEye image mosaic and textural features derived from the grey-level co-occurrence matrix. The study area is Parque Estadual Veredas do Peruaçu, a State Park in Minas Gerais, which was a plantation of eucalyptus until 1994 with a broad range of regeneration ages between 15 and 37 years. A total of 47 plots were surveyed for which the exact ages of regeneration are known and other structural variables were measured. Multiple regression and stepwise feature selection were used to create models that explain over 80% of the age, 58% of the crown closure, and 48% of the height of the vegetation. Texture proved essential for capturing the patterns of light and shadow generated by the trees of varying widths and heights.

With the availability of high-resolution satellite data, much research has been focused on the automatic detection and classification of individual tree crowns. Most of these studies were applied to temperate climates of the northern hemisphere, especially for forests of coniferous. Very few studies have been applied to the detection of trees in the tropical regions, least of all in the urban environment. Urban trees play a major role in maintaining or even improving the quality of life in cities by their contribution to the quality of the air, by absorbing rain water, by refreshing the air through transpiration and providing shadow. In this study we explored the potential of high-resolution WorldView-2 satellite data for the identification of urban individual tree crowns in the city of Belo Horizonte, Minas Gerais, Brazil, through an object-oriented approach. Irrelevant areas were masked (e.g. buildings, asphalt, shadows, exposed soil) using a threshold of NDVI. Three different approaches were tested to isolate and delineate individual tree crowns: region growing, watershed and template matching. For the first two approaches several parameters were tested to find the best result for the isolation of the individual tree crowns. An in-house program has been developed for template matching using a set of seven different templates of different species. A set of 300 individual tree crowns were visually interpreted in the WorldView-2 image to serve as validation and to compare the performance of the three different approaches. Then, the comparison was performed between the visual interpretation and the results of each approach by calculating the difference between the areas as a ratio of the validated area. Our results show that the region growing approach provided the best results, with an accuracy of over 80%.

ABSTRACT:This article presents an original algorithm created to detect and count trees in orchards using very high resolution images. The algorithm is based on an adaptation of the "template matching" image processing approach, in which the template is based on a "geometricaloptical" model created from a series of parameters, such as illumination angles, maximum and ambient radiance, and tree size specifications. The algorithm is tested on four images from different regions of the world and different crop types. These images all have < 1 meter spatial resolution and were downloaded from the GoogleEarth application. Results show that the algorithm is very efficient at detecting and counting trees as long as their spectral and spatial characteristics are relatively constant. For walnut, mango and orange trees, the overall accuracy was clearly above 90%. However, the overall success rate for apple trees fell under 75%. It appears that the openness of the apple tree crown is most probably responsible for this poorer result. The algorithm is fully explained with a step-by-step description. At this stage, the algorithm still requires quite a bit of user interaction. The automatic determination of most of the required parameters is under development.