The Impact of Stand Identification Through An Object-Oriented Approach For Forest Management Planning*

Stand boundary identification is an important factor in forest management planning that is dependent on the quality and resolution of the imagery available. We identified stand boundaries on an area within the John C. Stennis Space Center using two sources of multispectral imagery, IKONOS and QuickBird. Forest stands were delineated using eCognition v3.0. Segmentation was performed by initially finding an algorithm to produce objects representing forest stands larger than ten acres. Once this parameter for the algorithm was found, additional parameters (color, shape, compactness, smoothness) were iteratively changed to replicate area estimation. Preliminary results show that main differences in images exist in forest boundary locations. Also, the range of forest stand sizes within the each image type exhibit the largest differences between the images. Future work, analysis of stand boundaries in a forest planning model, will need to be completed to assess the impacts of area estimation.