论文部分内容阅读
The objective of this paper is to compare the four methods in extracting individual tree crown (ITC) from the high-resolution Quickbird image.Many researchers have tried to extract the tree crown using high-resolution remote sensing image,especially aerial images in the past years.And some of them have obtained many achievements on it recently.Among these,they studied more on not the satellite remote sensing with lower revolution,but the aerial images,owing to their high revolution.But on the other hand,the aerial images haven’t been well widely utilized due to its expensive cost.With the high-revolution satellite remote sensing having been made more widespread in commercial use,some researchers have realized it with both easy accessibility and high revolution,and have used the satellite remote sensing images in detecting the individual tree crown to certain precision.As a whole,they have approximately developed four types of algorithms to detect the tree crown.There are local maxima/minima、template matching、region growing and the multi-scale edge detection approach.These automatic identification methods of ITC provided a direct way to delineate the number of tree crowns in the region and some spatial information in the image.Nonetheless,they have their own advantages and disadvantages.So this paper used all the four approaches to delineate the individual tree crown from the Quickbird image for China fir plantations,and then compared,discussed the applicability for each at last.