An algorithm for delineation of individual tree crowns in high spatial resolution aerial images using curved edge segments at multiple scales

This paper presents an automatic multiple-scale algorithm for delineation o f individual tree crowns in high spatial resolution aerial images. The tree crown contours were identified as convex zero-crossings which were compute d with sub-pixel accuracy on the intensity image for each image scale. The centre of curvature was estimated for every edge segment pixel. For each se gment these centre points formed a path (evolute) or swarm which was modell ed as a primal sketch using an ellipse extended with the mean circle of cur vature. The model described the position or region of the derived tree crow n based on the edge segment at the current scale. The sketch was rescaled w ith a significance value and accumulated for a scale interval. In the accum ulated sketch a tree crown segment was grown, starting at local peaks, unde r the condition that it was inside the area of healthy vegetation in the ae rial image and did not trespass a neighbouring crown segment. The method wa s evaluated by comparison with a manual delineation work and ground truth o n 43 randomly selected sample plots. It was concluded that the performance of the method is almost equivalent to visual interpretation. On the average , seven out of ten tree crowns were the same. Furthermore, ground truth ind icated a large number of hidden trees on the sample plots. The proposed tec hnique could be used as a basic tool in forest surveys.