JP4484183B2 - Forest information processing system - Google Patents

Description

[0001]
[Field of the Invention]
The present invention relates to a proposed method for evaluating the forest area on the basis of the color image obtained by photographing a forest area from above, in particular, determine the crown shape of trees that make up the forests the color image subjected to image processing by a computer, the tree The present invention relates to a forest information processing system suitable for evaluating a forest area by grasping a distribution state and a degree of mixing.
[0002]
In this specification, “crown” refers to a rounded part composed of the upper branches and leaves of each tree that constitutes the forest, and “crown shape” refers to a forest tree directly above (or abbreviated). This is the crown shape seen from above.
[0003]
[Prior art]
In the conventional evaluation of forest areas, a comparatively narrow area of a vast forest area is set as a survey area in advance, and actually visited the forest area, the distribution state of different trees, the vegetation density of trees, the diameter of the trunk of the tree And the height of trees was measured, and the entire forest area was evaluated based on the field survey in this survey area. Since the entire forest area is estimated by estimation based on such partial survey results, it is extremely difficult to accurately evaluate the entire forest area.
[0004]
The procedure for assessing traditional forest areas is:
(a) A part of the forest area is defined as a survey area with reference to a forest plan with a scale of 1/5000 or aerial photographs.
(b) Use the appropriate means of transportation to go near the survey area and walk into the survey area.
(C) If GPS (global positioning system) is available, use GPS to determine the location of the survey point. If GPS is not available, the location of the survey point is specified from the surrounding terrain.
(d) Enter the survey results on a map and survey notes.
[0005]
[Problems to be solved by the invention]
As described above, since there are many mountainous areas where vehicles cannot be used in the survey area, movement within the survey area has to be relied upon on foot, and there has been a problem that on-site surveys are very expensive and labor intensive. Furthermore, even if it is assumed that the tree species and age in the survey area are the same as the tree species and tree age in other areas, the growth state of the tree differs depending on the location conditions of the tree. Based on it, it was extremely difficult to accurately evaluate the whole forest area.
[0006]
Furthermore, in many areas where trees are prosperous, position information data from GPS artificial satellites cannot be received in many cases, and it is often difficult to accurately determine the position of a survey point.
[0007]
OBJECT OF THE INVENTION
Therefore, according to the present invention, an aerial color photograph or artificial satellite color image obtained by photographing a forest area is input to a computer, a crown shape image of the forest area is obtained by computer image processing, and the crown shape image and the original color image are obtained. The purpose is to investigate and evaluate tree vegetation in the whole forest area from the color and brightness of trees.
[0008]
[Means for Solving the Problems]
According to the present invention, a color image obtained by photographing a forest area from the sky is converted into a black and white image, image processing is performed on the black and white image to obtain a crown shape image, spectrum analysis of the color image is performed, and the spectrum analysis result The forest area is evaluated by using a computer from the canopy shape image.
[0009]
The color image is taken using an aircraft or an artificial satellite, and the evaluation of the forest area includes specifying the type of forest tree.
[0010]
Further, before the black-and-white image conversion, the color image may be subjected to a subpixel conversion process to increase the resolution of the color image.
[0011]
Further, the image processing for the black and white image includes image smoothing and tree border boundary enhancement processing.
[0012]
Furthermore, the present invention converts a color image obtained by photographing a forest area from the sky into a black and white image, performs image processing on the black and white image to obtain a crown shape image, performs spectral analysis of the color image, and results of the spectral analysis And a recording medium recording a program for evaluating a forest area from the canopy shape image.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments according to the present invention will be described with reference to FIGS.
[0014]
FIG. 1 is a diagram for explaining a tree crown and a crown shape. As mentioned above, “crown” refers to the rounded part composed of the upper branches and leaves of each tree that constitutes the forest, and “crown shape” refers to the forest tree directly above (or approximately true). The crown shape seen from above. FIG. 1 (a) shows a state where a forest area is photographed from an aircraft or an artificial satellite and a crown portion. FIG.1 (b) shows a mode that the crown was seen from right above (or substantially right above).
[0015]
FIG. 2 is a flowchart for explaining an image processing step (process) according to the present embodiment. In FIG. 2, an analog color photograph of a forest area taken from an aircraft or an artificial satellite is input to a computer and converted into a digital image (step 10). However, if a color image taken by a digital camera from an aircraft or an artificial satellite is input to the computer, the digitization of the analog color image in step 10 can be omitted.
[0016]
FIG. 3 is a diagram showing an example of a color image captured from an aircraft and input to a computer as a black and white image. This is because the drawings attached to the specification cannot be displayed in color.
[0017]
Next, when the digital color image obtained in step 10 does not have a sufficient resolution for extracting the crown shape, in step 12, a subpixel conversion process is performed to accurately obtain the crown shape. Add However, this step 12 can be omitted if the digital color image has a sufficient resolution.
[0018]
In step 14, the color digital image is converted into a multi-value black and white digital image (monochrome digital image) (means for obtaining a black and white image from the color image is referred to as image conversion means) . In the next step 16, the image is smoothed. To prepare for subsequent image processing. The purpose of this smoothing process is to eliminate the problem that a single crown shape is divided into a plurality of shapes when obtaining a crown shape in step 20 to be described later if there are fine pixel value variations in the image. It is. This image smoothing process is known as a method using a Gaussian filter.
[0019]
Next, in step 18, the boundary of the adjacent tree crown is emphasized. The purpose of this process is to make it easier to determine the crown shape in subsequent image processing. This boundary line emphasis can be performed by high-pass filter processing, for example. Of course, this processing step can be omitted if the image is such that the portion considered to be the boundary of the crown shape is clearly shown. However, it seems that the boundary line emphasis process of step 18 is usually necessary.
[0020]
Next, in step 20, the crown shape is obtained. The means for obtaining the crown shape is referred to as a crown shape image creating means. A method for obtaining the crown shape will be described with reference to FIG. When a crown is photographed from an aircraft or an artificial satellite, the brightness increases according to the rise of the crown. That is, the luminance is highest at the apex portion of the tree crown, and the luminance change is smallest at the apex portion. The brightness decreases toward the periphery of the tree canopy, and the brightness is lowest at the border between adjacent tree canopies. In the embodiment of the present invention, the crown shape is obtained by a so-called watershed algorithm using the change in luminance accompanying the roundness of the tree crown.
[0021]
In the watershed algorithm, a marker is set at the center of the portion where the change in luminance (gradient) is the smallest. That is, the marker is set at the apex of the respective crown or near the apex (in FIG. 4, the mark * is added to the initial setting position of the marker). Next, markers that are initially set at the vertices of the respective crowns are grown in four directions along the luminance change (gradient) (indicated by a plurality of arrows in FIG. 4). Subsequently, a portion that collides with a marker of an adjacent tree crown is determined as a boundary line of the tree crown. In this way, in the watershed algorithm, the crown shape is extracted by growing a region along the luminance change with the crown top having a small luminance change (gradient) as a starting point.
[0022]
FIG. 5 is a diagram showing a crown shape image obtained by the crown shape image creation means using the above-described watershed algorithm from the color image of FIG. 3 (which is expressed in black and white in the figure).
[0023]
Subsequently, in step 22 of FIG. 2, the tree species is obtained using the crown shape of the forest area and the wavelength spectrum (spectrum analysis result) of the original color image input to the computer in step 10. The means for obtaining the tree species is referred to as tree species specifying means. Specifically, for example, the tree species may be obtained by superposing the crown shape image of FIG. 5 and the original color image and performing spectrum analysis.
[0024]
An example of this tree species determination will be described with reference to FIG. Fig. 6 shows the light wavelengths (horizontal axis) of four types of trees: Japanese cedar (cedar), Japanese cypress (japanese cypress), Japanese oak (small cocoon), and oak (japanese cypress), taken in color in November. It is the figure which showed the relationship (spectrum) of the brightness | luminance (vertical axis | shaft) with respect to.
[0025]
As shown in FIG. 6, the brightness of the crown of the cedar is generally low over the entire spectrum from blue to near-infrared compared to other tree species, and the crown of the cypress is cedar in the visible light of blue, green and red. The brightness is low in the same manner as in the above, but the brightness is high in the near infrared region. On the other hand, Quercus generally has a higher luminance than cedar and cypress, and Oak has an overall higher luminance like Quercus but has lower green and red spectra than Quercus.
[0026]
In the forest area, evergreen trees, deciduous trees, and the like are mixed, and the spectrum of each tree in the photographed color image varies depending on the season. However, if “seasonal fluctuation data” appearing in color images of such forest areas is accumulated and verified over an appropriate period of time, an accurate tree based on the crown shape and spectral analysis of the color image obtained for the crown shape Identification is possible. Furthermore, data fluctuations due to differences in sunlight at the time of shooting are problems that can be solved by storing and verifying shooting data.
[0027]
7 inputs the image shown in FIG. 3 into a computer, performs the image processing of FIG. 2 to obtain the crown shape image shown in FIG. 5, and obtains the cedar obtained by referring to the spectrum of each tree shown in FIG. It is the figure which evaluated distribution of four types of trees, cypress, Japanese oak, and oak.
[0028]
The above-mentioned evaluation of forest areas explained the distribution of trees. However, in the case of the same kind of tree, it is possible to estimate the diameter, height and age of the trunk based on the size of the crown. Furthermore, the present invention is not limited to a high tree and can be applied to the distribution evaluation of a shrub.
[0029]
【The invention's effect】
As described above, according to the present invention, an aerial color photograph or an artificial satellite color image obtained by capturing a forest area is input to a computer, and a crown shape image of the forest area is obtained by computer image processing. Since we are investigating and evaluating tree vegetation in the entire forest area based on the color and brightness of trees obtained from color images, it is possible to overcome conventional problems associated with forest area evaluation based on field surveys. Has a noticeable effect.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a tree crown of a forest area.
FIG. 2 is a flowchart for explaining computer processing of an image according to an embodiment of the present invention.
3 is a diagram showing a color image of a forest area input to a computer in the computer processing of FIG. 2 as a black and white image.
4 is a diagram for explaining computer processing for obtaining the crown shape of FIG. 2; FIG.
FIG. 5 is a diagram showing a crown shape image obtained by the computer processing of FIG. 2;
6 is a diagram showing a spectrum of a tree used for specifying a tree in the computer processing of FIG. 2; FIG.
FIG. 7 is a view showing a tree distribution obtained by the computer processing of FIG. 2;
[Explanation of symbols]
10: Computer processing for inputting a color image (analog) obtained by photographing a forest area from the sky and converting it into a digital image Step 12: Computer processing for performing subpixel processing Step 16: Computer processing for smoothing the image Step 18: Computer processing step 20 for emphasizing the crown boundary line: Computer processing step 22 for determining the crown shape Computer processing step for determining the model based on the crown shape and spectrum analysis

Claims (4)

A forest information processing system that determines the distribution of tree species in a forest area using color images taken from above .
Image conversion means for converting the color image into a black and white image;
A crown shape image creating means for creating a crown shape image by obtaining a crown shape by applying a watershed algorithm to the black and white image;
Obtaining a wavelength spectrum of a color image portion corresponding to a crown portion of the crown shape image, referring to a wavelength spectrum for each tree species obtained in advance, and a tree species identifying means for identifying a tree species of the crown portion of the crown shape image;
Forest information processing system . The forest information processing system according to claim 1, wherein the color image is taken using an aircraft or an artificial satellite. 3. The forest information processing system according to claim 1, wherein before the image conversion by the image conversion unit , a process for increasing the resolution of the color image by performing a subpixel process on the color image is performed. The forest information processing system according to any one of claims 1 to 3 , wherein after the image processing by the image conversion means, processing of image smoothing and tree border boundary enhancement is performed .

Images