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. Stereophotogrammetry 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.

No abstract

Image orientation requires ground control as a source of information for both indirect estimation and quality assessment to guarantee the accuracy of the photogrammetric processes. However, the orientation still depends on interactive measurements to locate the control entities over the images. This paper presents an automatic technique used to generate 3D control points from vertical panoramic terrestrial images. The technique uses a special target attached to a GPS receiver and panoramic images acquired in nadir view from different heights. The reference target is used as ground control to determine the exterior orientation parameters (EOPs) of the vertical images. These acquired multi-scale images overlap in the central region and can be used to compute ground coordinates using photogrammetric intersection. Experiments were conducted in a terrestrial calibration field to assess the geometry provided by the reference target and the quality of the reconstructed object coordinates. The analysis was based on the checkpoints, and the resulting discrepancies in the object space were less than 2 cm in the studied cases. As a result, small models and ortho-images can be produced as well as georeferenced image chips that can be used as high-quality control information.

Forest variables are typically surveyed using sample plots, from which parameters for large areas are estimated. The diameter at breast height (DBH) is one of the main variables collected in the field and can be used with other forest measures. This study presents an automatic technique for the mapping and measurement of individual tree stems using vertical terrestrial images collected with a fisheye camera. Distinguishable points from the stem surface are automatically extracted in the images, and their 3D ground coordinates are determined by bundle adjustment. The XY coordinates of each stem define an arc shape, and these points are used as observations in a circle fitting by least squares. The circle centre determines the tree position in a local reference system, and the estimated radius is used to calculate the DBH. Experiments were performed in a sample plot to assess the approach and compare it with a technique based on terrestrial laser scanning. In the validation with measurements collected on the stems using a measuring tape, the discrepancies had an average error of 1.46 cm with a standard deviation of 1.09 cm. These results were comparable with the manual measurements and with the values generated from laser point clouds.

The objective of the proposed approach is to locate and measure ground control points in aerial images when large image search spaces are defined due to the use of inaccurate initial exterior orientation parameters, such as those provided by consumer-grade navigation systems. Vertical terrestrial image patches covering control point areas are generated and compared with aerial patches using feature-and area-based matching algorithms to automatically determine their corresponding positions in aerial images with sub-pixel precision. The approach is based on techniques for both image search space reduction and adaptive least squares matching. Experiments with real data were performed with bundle block triangulation and the results were analysed using control and check points in both object and image spaces. The proposed technique enabled a significant reduction in the search space within which it was feasible to successfully locate control points. Compared with manual measurements, the results obtained by the automatic technique were more accurate, achieving one-fifth of the ground sample distance in planimetric check point discrepancies.