Identification of Successional Stages and Cover Changes of Tropical Forest Based on Digital Surface Model Analysis

Forests are in a permanent state of change due to natural and anthropogenic processes. Long-term time series analysis makes it possible to reconstruct the forest history and perform a multitemporal analysis on the cause and effect of changes. This paper describes an approach for successional stage classification in a tropical forest based on vertical structure variations. Stereo-photogrammetry and novel image matching methods are used to produce dense digital surface models (DSMs) from optical images (historical and contemporary). An approach was developed to classify the successional stages of trees using local height variations provided by a DSM and image intensity values. Experiments were performed in a semi-deciduous tropical forest fragment located in the West of São Paulo State, Brazil. Six test sample plots and a line transect were established and field surveys were conducted to collect forest variables. These variables were used to characterize and validate five successional classes based on secondary tree species that stratify the forest canopy. The current status of the entire forest fragment was characterized using recent photogrammetric imagery, and a map of historical successional stages was established by analyzing the historical photogrammetric imagery. The investigation demonstrated that the proposed technique can be used to reconstruct the geometric structure of a forest canopy from aerial images. The successional stages can be identified and compared over time using multitemporal photogrammetric imagery and DSMs, which enables an analysis of forest cover changes. The results indicated that the successional stage has changed dramatically during the 50 years period of time.