Ryotaro Komura

Delineation of tree crown in high resolution satellite image using circle expression and watershed algorithm

Forests play an important role as environment for living things. For controlling forests, it is useful to know condition of tree efficiently. Recently, high spatial resolution images are taken using satellite and detail of trees in forest can be identified by visual inspection of satellite image. Shape of tree crown is one of important parameters that can be identified using satellite image. In this study, we developed a method to delineate crown and applied to high resolution satellite image. Then result of crown delineation was compared with result of visual inspection. At first, we assumed that shape of crown resembled circle and calculated circle radiuses showing crown size. Calculated radius was larger as similar brightness region became wide. The radiuses at each pixel position were calculated and the result was used as the radius distribution image (RDI). In the RDI, the radius was larger as the position became close to center part of crown. If radius value is considered as elevation, RDI seems to be DEM (Digital Elevation Model). The value of each pixel of RDI was inverted and inverse radius distribution image (IRDI) was calculated. In the IRDI, it is expected that a cone-shaped geographical feature is on the crown region and each crown region make each watershed. Then crown regions were calculated applying watershed algorithm to IRDI. In the results, shapes of calculated regions were close to crown shape identified with visual inspection

Classification of forest stand considering shapes and sizes of tree crown calculated from high spatial resolution satellite image

Forests play important role as environment for living things. There are many types of forest stand and the identification of forest type is available to conserve and to improve the forest. The field investigation on forest area is a hard work and needs many cost. The field investigation on wide area is almost impossible. The remote-sensing is useful for investigation of wide area of forest, especially on mountain area where field investigations are usually laborious. The spectral characteristics are different by the species. Many classifications of forest type based on spatial characteristics are operated. But, some kinds of forest types are difficult to be classified by using only spatial characteristics. Each type of forest stand has a spatial characteristic depending on the shape of trees. By using not only spatial characteristics but also tree shapes, number of classified types of forest can be increased. We developed the method for the delineation of tree crown with image processing using satellite images with high spatial resolution and aerial photographs. In this method, tree crowns that have similar color are delineated with circles. Densities, size and numbers of trees are different by species, and these information are available to classify the type of forest stand. In this study, forest stands were classified with not only spatial information but also the number and densities of trees. Multi-spectral images and pan image of IKONOS were used in this study. The NDVI (Normalized Difference Vegetation Index) was calculated from red and near-infrared images. Densities and numbers of trees on a forest stand were calculated using pan images with high spatial resolution by using the method for crown delineation with circler expression. The types of forest stand were classified using thresholds of NDVI and numbers of trees. A forest classification method using spatial information, densities and numbers of trees increased the number of classified type of forests than using only spatial information.

Identification of dead tree of Japanese oak wilt (JOW) using high spatial resolution satellite imagery

The incidence of Japanese oak wilt (JOW) has been increasing in Japan since late 1980s. The JOW is caused by the ambrosia fungus Raffaelea quercivorus vectored by an ambrosia beetle Platypus quercivorus. Detection of trees killed by JOW is important to prevent new incidence of JOW but difficult because most of JOW mortality occurs in mountainous area. Remote sensing with high spatial resolution imageries taken by IKONOS® or Quickbird® can be laborsaving technology to identify individual trees killed by JOW over wide areas. A purpose of this study is to develop a laborsaving image processing technique to identify individual tree crowns. The orthophoto imagery was transformed both to HSI (Hue, Saturation, and Intensity) and NDVI (Normalized Difference Vegetation Index) datasets. Individual dead tree crowns killed by JOW were identified by cluster analysis using datasets of H, S, and NDVI.

Representation of tree crown by multiple circles in aerial images

Investigates the distribution of crown size of trees using aerial images. In a previous study, a method was developed to draw a tree as a circle and estimate the crown size from the circle radius. But it was difficult to represent the complex shape of a crown by a circle. In this study the complex crown shape was reconstructed by several circles for estimating the size of the crown. First, the size of a circle was computed using the distribution of the brightness value within the circle area. Next, overlapping circles having similar brightness values were unified. As a result of this method the complex shape of a crown could be represented by several circles, and the size of the crown was analyzed accurately.

Analysis of relationships between image data and spectrum of vegetation measured from helicopter

The spatial resolution of satellite mounted sensors has improved. However, it is not well known how effective the usual method is in using high resolution sensors. In this study, two kinds of sensors which differ in spatial resolution were used to check the effectiveness of the usual method. There were several differences in the results from the two sensors.

Comparison of vegetation measured from three different distances

Remote sensing is widely used for the monitoring of forests and the VIS/NIR reflectance is commonly used for the identification and characterization of the vegetation. The reflectance data obtained at higher altitudes is some kind of average over a certain extension of area, and also the atmosphere that lies between the sensor and the object affects the data. Therefore, in the interpretation of the remote sensing data, knowing the difference between the data obtained at different scales and distances is important. The authors measured the spectra of trees at three different scales: 1) individual leaves, 2) part of a tree seen from a distance of 40 m, 3) mixture of several different trees seen from a helicopter, and investigated what affects the data during the scaling up of the measurements.

A observation of predictor of Japanese oak wilt using high spectral sensor

The incidence of Japanese oak wilt (JOW) has been increasing in Japan since the late 1980s. JOW is caused by the ambrosia fungus Raffaelea quercivorus vectored by the ambrosia beetle Platypus quercivorus. Detection of trees killed by JOW is important to prevent new incidence of JOW but is difficult because most of JOW mortality occurs in forests in mountainous areas. Remote sensing with high spatial resolution imagery can be labor saving technology to detect individual trees killed by JOW over wide areas. The detection of JOW area in early stage is important in the prevention of JOW. In the past method, the detection of JOW area was possible in the stage after outbreak, but the detection in the early stage of JOW was impossible because the symptoms of JOW did not appear in the property on the multi-spatial sensor on the satellite. Recently the number of satellite-borne hyper-spectral sensor is increasing and the performance of the sensor is improved in the spatial and the spectral resolution. If the symptoms of JOW appear in the data from hyper-spectral sensor, the detection of JOW in early stage is realized. In this study, the artificial JOW model trees were set up and the spectral property of the model trees were measured from making of the model trees with a spectrometer. We measured the property of the model trees until the dead of model trees and analyzed the property to find the symptoms of JOW in early stage. As a result, we could find a feature on the property as the symptoms of JOW by calculation of spectral differential value.

Temporal analysis of forest activity using remote sensing and DEM

Vegetation in forest is influenced by elevation and topographical features (e.g. temperature, sunshine, slope direction, etc.). In mountain areas vegetation aspects are different with elevation and slope direction. But it is difficult to study the relationship between vegetation in such areas by field work. The technology which makes it possible is satellite remote sensing. It enables the research of forest vegetation activity by multispectral data from satellite. In combination with a DEM (Digital Elevation Model) one can research the change of forest activity by elevation. Furthermore, it is possible to calculate the direction of each slope and a degree of the slope from the DEM. In this study, the temporal change of forest activity around Mt. Hakusan, Japan, with slope direction was analyzed using Landsat TM data.

Analysis of temporal changes in vegetation in central Japan by remote sensing and ground data

Monitoring the extent and type of vegetation cover and its state of health will always be important. Spatial and temporal changes in localized vegetation were analyzed using Landsat TM data. Vegetation vigor and abundance around Mt. Hakusan in Japan have decreased in 10 years.