Abstract
Region growing is frequently applied in automated individual tree crown delineation (ITCD) studies. Researchers have developed various rules for initial seed selection and stop criteria when applying the algorithm. However, research has rarely focused on the impact of tree-oriented growth order. This study implemented a marker-controlled region growing (MCRG) algorithm that considers homogeneity, crown size, and shape using airborne laser scanning (ALS) data, and investigated the impact of three growth orders (i.e., sequential, independent, and simultaneous) on tree crown delineation. The study also investigated the benefit of combining ALS data and orthoimagery in treetop detection at both plot and individual tree levels. The results showed that complementary data from the orthoimagery reduced omission error associated with small trees in the treetop detection procedure and improved treetop detection percentage on a plot level by 2%–5% compared to ALS alone. For tree crown delineation, the growth order applied in the MCRG algorithm influenced accuracy. Simultaneous growth yielded slightly higher accuracy (about 2% improvement for producer’s and user’s accuracy) than sequential growth. Independent growth provided comparable accuracy to simultaneous growth in this study by dealing with overlapping pixels among trees according to crown shape. This study provides several recommendations for applying region growing in future ITCD research.
Highlights
Individual tree measurement is an important component of forest inventory that was historically dependent on highly trained personnel in expensive and time-consuming field surveys [1]
Trees omitted in the airborne laser scanning (ALS) processing were identified using the orthoimagery when they had a local maximum in the green band and were higher than a specific percentile of height within a corresponding crown-size window on the smoothed canopy maximum model (CMM)
In the extreme case, when the 50th percentile was used in Plot 2, the number of detected treetops (885) was greater than the number of reference treetops (858); the 103.1% detection percentage (DP) failed to represent the performance of the treetop detection procedure
Summary
Individual tree measurement is an important component of forest inventory that was historically dependent on highly trained personnel in expensive and time-consuming field surveys [1]. Since the application of aerial photographs to forest inventory in the mid-20th century, the efficiency of field work has been complemented by image analysis. Obtaining individual tree-based information through visual interpretation of imagery is still a cost- and labor-intensive process [2,3]. Researchers have devoted significant effort to developing techniques to obtain individual tree- and stand-level information by automatically delineating individual tree crowns in remotely sensed data, (e.g., [4,5,6]). LiDAR (light detection and ranging) data supports these modern approaches to forest inventory [7]. Delineated tree crowns are an essential component of precision forestry to support estimation of crown size and tree diameter [8], crown closure [9], canopy structure [10], height and biomass [11], and improve tree species classification [7] and tree growth evaluation [12]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
