Benefit of airborne full waveform LIDAR for 3D segmentation and classification of single trees

The paper demonstrates the advantage of full waveform LIDAR data for segmentation and classification of single trees. First, a new 3D segmentation technique is highlighted that detects single trees with an improved accuracy. The novel method uses the normalized cut segmentation and is combined with a special stem detection method. A subsequent classification identifies tree species using salient features that utilize the additional information the waveform decomposition extracts from the reflected laser signal. Experiments were conducted in the Bavarian Forest National Park with conventional first/last pulse and full waveform LIDAR data. The first/last pulse data result from a flight with the Falcon II system from TopoSys in leaf-on situation at a point density of 10 points/m ². Full waveform data were captured with the Riegl LMS-Q560 system at a point density of 25 points/m ² (leaf-off and leaf-on) and at a point density of 10 points/m ² (leaf-on). The study results prove that the new 3D segmentation approach is capable of detecting small trees in the lower forest layer. This was practically impossible so far if tree segmentation techniques based on the canopy height model (CHM) were applied to LIDAR data. Compared to the standard watershed segmentation the combination of the stem detection method and the normalized cut segmentation performs better by 12%. In the lower forest layers the improvement is even more than 16%. Moreover, the experiments show clearly that the usage of full waveform data is superior to first/last pulse data. The unsupervised classification of deciduous and coniferous trees is in the best case 93%. If a supervised classification is applied the accuracy is slightly increased with 95%. Classification with first/last pulse data ends up with only 80% overall accuracy. Interestingly, it turns out that the point density has practical no impact on the segmentation and classification results.