JP3816826B2 - Urban area extraction processing method, urban area extraction processing apparatus, urban area extraction processing program, and recording medium for the program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an urban area extraction processing method, an urban area extraction processing apparatus, an urban area, and the like, which automatically extract an urban area from an image taken from above such as a satellite photograph image and an aerial photograph image based on the texture information. The present invention relates to an extraction processing program and a recording medium for the program.
[0002]
[Prior art]
In the conventional urban area extraction technology, if a closed area that satisfies a specific color and brightness condition is focused on and the closed area is close to a predetermined shape, the closed area is assumed to be a building area. , The buildings extracted by such a method were integrated and recognized as an urban area.
[0003]
[Problems to be solved by the invention]
However, this type of conventional method is used when the image is monochrome or grayscale and there is no color information, or when a part of the building is hidden by a tree or another structure and does not take a closed area shape. There is a problem that the urban area cannot be extracted stably when each building does not have a sufficient size due to its low resolution and the shape of the closed area is collapsed.
[0004]
The present invention is problematic in the above-described conventional technique by focusing on the texture feature in the urban area, that is, the characteristic that the edge power is strong and the edge tangent direction tends to be orthogonal in the urban area. Even in such a case, the object is to enable the city area extraction to be performed stably.
[0005]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the present invention performs urban area determination based on a feature quantity that quantitatively represents this tendency because the edge power of edge tangents is strong and the tendency of the tangent directions to be orthogonal is strong in urban areas. The following method is used as an implementation means.
1. Set an arbitrary size block centered on each pixel for the image of the urban area extraction target, calculate the edge difference in the horizontal and vertical directions at each pixel, and calculate the angle parameter of the edge tangent based on the obtained edge difference value A distance parameter is calculated.
2. A Hough transform is performed on the calculated edge tangent parameter, and each edge tangent parameter is voted on a two-dimensional voting space having the angle parameter and the distance parameter as two axes.
3. An appropriate feature amount is obtained from the obtained voting space, and based on this, it is determined whether or not the representative point of the block is an urban area.
[0006]
This method is particularly effective for extracting large buildings such as factories. The block size should be large enough to fit the building. As described above, it is possible to extract a stable region based on the texture characteristics of the urban area.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described in detail with reference to the drawings.
[0008]
FIG. 1 shows a configuration example of an urban area extraction processing apparatus according to this embodiment. In the figure, 1 is an urban area extraction processing device comprising a CPU and a memory, 10 is an input image storage unit for storing an image of the ground surface such as a satellite photograph image or an aerial photograph image, and 11 is a predetermined pixel centered on the pixel of interest. A block acquisition processing unit that acquires a block of a size, 12 an edge tangent parameter calculation processing unit that obtains an edge tangent parameter at each point in the block, and 13 a two-dimensional vote by performing a Hough transform on the obtained edge tangent parameter A Hough conversion processing unit for voting in a space, 14 is a voting space memory for a created voting space, 15 is an urban area determination processing unit that determines whether it is an urban area based on the voting space, and 16 is an urban area of the determination result Information.
[0009]
FIG. 2 shows a processing flow of the urban area extraction processing apparatus 1 shown in FIG. The block acquisition process 21 is a process performed by the block acquisition processing unit 11. Here, the pixel of interest for the input image 20 obtained by photographing the ground surface such as a satellite photograph image or an aerial photograph image stored in the input image storage unit 10. Get a block of a certain size centered on. The shape of the block is arbitrary, such as a square or a circle, but in this embodiment, a circular block is acquired. This is because it is convenient for the Hough conversion of the edge tangent parameter performed thereafter.
[0010]
An example of a circular block is shown in FIG. The size of the circular block 30 is preferably large enough for a building to enter, and varies depending on the resolution of the image, but is a size such as several tens to hundreds of pixels.
[0011]
The edge tangent parameter calculation process 22 is a process for obtaining an edge tangent at each point in the block. Here, as a process performed by the edge tangent parameter calculation processing unit 12 in FIG. 1, first, an edge vector is obtained by performing a difference calculation in the horizontal direction and the vertical direction on an image, and an edge tangent parameter is obtained based on the obtained edge vector. Will be explained. Before performing the difference calculation, the influence of noise can be reduced by performing blurring processing with a Gaussian filter or the like as preprocessing.
[0012]
FIG. 4 is a diagram for explaining an edge vector calculation method. Difference calculation in the horizontal direction is performed by subtracting the left and right pixel values of the target pixel, and difference calculation in the vertical direction is performed by subtracting the upper and lower pixel values. Hereinafter, the horizontal direction is expressed as the X direction, and the vertical direction is expressed as the Y direction.
[0013]
In the example shown in FIG. 4A, the edge vector C (C _X , C _Y ) is adjacent to the pixel C (point C) with the horizontal difference value C _X and the vertical difference value C _Y. From the pixel values of pixels A, B, D, and E,
C _X = (D−B) / 2
C _Y = (EA) / 2
Is calculated.
[0014]
In addition, for pixels at the edge of the image, there may be no vertical, horizontal, and adjacent pixels. In this case, the difference between the adjacent pixel and the pixel of interest itself is calculated and used as an edge vector component. Such an example is shown in FIG. The edge vector G (G _X , G _Y ) at the pixel G (point G) is obtained from the pixel values of the adjacent pixels F, H, I.
G _X = HG
G _Y = (IF) / 2
Is calculated.
[0015]
FIG. 5 shows an example of edge vector information. 5A shows the pixel value of the original image, and FIG. 5B shows the value of the edge vector component.
[0016]
Next, an edge tangent parameter is calculated from the obtained edge vector. Here, assuming that the center of the block is the origin, the distance r from the origin to the tangent and the angle θ that the normal vector of the tangent makes with the X direction are the edge tangent parameters to be obtained.
[0017]
FIG. 6 is a diagram for explaining an example of the edge tangent parameter. As is apparent from FIG. 6, the edge tangent parameters θ _P and r _{P at} the point _P are easily calculated from the edge vector components P _X and P _Y at the point P.
[0018]
The Hough conversion process 23 is a process of performing Hough conversion on the calculated edge tangent parameter and voting to the two-dimensional voting space 24, and is performed by the Hough conversion processing unit 13 in FIG. FIG. 7 is a diagram for explaining an example of the voting process. Edge tangent parameter theta _P and r _P at point P shown in FIG. 7, be voted to the coordinates (theta _P, r _P) in voting space.
[0019]
Although the value to be voted may be +1, in this embodiment, the edge power value itself, that is, the square root of (P _X ² + P _Y ² ) is used as the vote value. By this method, the parameters of the edge having a strong power existing in the urban area are more emphasized.
[0020]
For the purpose of suppressing the adverse effects of noise and the like, it is also effective to use a voting process when the edge power is below a certain value, or to give a voting value by adding some weight to the edge power.
[0021]
The voting space 24 is a voting space obtained as a result of performing the Hough conversion processing 23 for all the pixels in the block. An example is shown in FIG. FIG. 8A shows an original image, and FIG. 8B shows an example of a voting space. In the example shown in FIG. 8A, voting is performed for all pixels in a circular block having a diameter of 50 pixels. In the voting space of FIG. 8B, the magnitude of the vote value is shown by changing the density. The horizontal axis of the voting space is the axis of the angle parameter θ (range 0 to 360 °), and the vertical axis is the axis of the distance parameter r (range 0 to 25). It can be seen that the straight line in the image is converted as a point having a high vote value in the voting space.
[0022]
The intra-block process end determination 25 is a process for determining whether or not the edge tangent parameter calculation process 22 and the Hough conversion process 23 have been completed for all pixels in the block. If there is an unprocessed pixel, the process returns to the edge tangent parameter calculation process 22 again to process the unprocessed pixel.
[0023]
The urban area determination process 26 is a process of determining whether or not the representative point of the block is an urban area based on the information of the voting space 24 in the voting space memory 14, and is performed by the urban area determination processing unit 15 of FIG. As a representative point of the block, there are a central pixel of the block, a pixel near the center of the block, all pixels constituting the block, a pixel having an edge power exceeding a certain value in the block, and the like. In this embodiment, a method of using a histogram (hereinafter referred to as an angle histogram) that is projected in the X-axis direction (angle parameter axis direction) and reduced in one dimension without using the two-dimensional voting space as it is. explain.
[0024]
FIG. 9 shows an angle histogram created from the voting space shown in FIG. There is also a method of obtaining an urban area from statistics such as the maximum value, minimum value, average value, and variance value of the angle histogram. Here, (1) maximum / minimum value and (2) determination processing based on edge tangent orthogonality explain.
[0025]
(1) Maximum value / minimum value The maximum value and minimum value of the angle histogram are obtained, and the ratio is calculated by dividing the maximum value by the minimum value. If this value is greater than or equal to a certain value, the representative point of the block is determined to be an urban area. Since the origin at the time of tangent parameter calculation is set at the center of the block, even if the apparent inclination is the same straight line, the angle parameter θ has a difference of 180 ° when it is located on the opposite side of the block center. . Therefore, by adding the 180-360 ° portion of the angle histogram to the 0-180 ° portion and setting the angle histogram range to 0-180 °, straight lines having the same inclination are corrected so as to have the same angle parameters. A method is also conceivable.
[0026]
For images with random edge directions, such as random noise, the maximum / minimum values are small, and this value is large when there is a straight line with strong edge power in the block. It is.
[0027]
(2) When an orthogonal straight line exists in the edge tangent orthogonal block, two peaks having an interval of 90 ° are generated in the angle histogram. Whether there are two peaks at 90 ° intervals can be determined by the following method.
[0028]
FIG. 10 is an explanatory diagram of the determination method. The histogram shown in FIG. 10 is obtained by overlapping the 180-360 ° portion with the 0-180 ° portion so that the angle range of the angle histogram is 0-180 ° as described in (1) above. _First , the angle θ ₁ taking the maximum value in this histogram is obtained. Next, a second peak search area S _A having an arbitrary width around an angle θ ₂ having a difference of θ ₁ to 90 ° is set, and the maximum value (second peak value) in that area is obtained. In the example of FIG. 10, a width of ± 10 ° with θ ₂ as the center is set as the range of the second peak search region S _A.
[0029]
After obtaining the second peak value, this value is divided by the minimum value of the histogram. If the ratio is equal to or greater than a certain value, it is determined that the representative point of the block is an urban area. In addition to the ratio between the second peak value and the minimum value, the value of the second peak value itself, the difference between the second peak value and the minimum value, etc., can be used as a determination scale, or the maximum value itself can be added to the determination scale. Is also possible.
[0030]
In addition, an angle histogram with a range of 0 to 180 °, and a 90 to 180 ° portion overlapped with a 0 to 90 ° portion to create a histogram with an angle range of 0 to 90 °, the maximum and minimum values of the histogram. It is also conceivable to use the ratio of values as a judgment scale.
[0031]
Here, a method for determining an urban area based on an angle histogram has been described, but a method for calculating an appropriate feature amount using a two-dimensional voting space as it is is also conceivable.
[0032]
The end determination process 27 is a process for determining whether or not the urban area determination process 26 has been performed for all pixels. If there is a pixel that has not yet been processed, the process returns to the block acquisition process 21 and the same processing is performed for the unprocessed pixel.
[0033]
By performing the above processing for all the pixels, the urban area information 16 to which a bitmap or label indicating whether or not the area is an urban area is obtained for each pixel.
[0034]
The processing performed by the above-described urban area extraction processing apparatus 1 can be realized by a computer and a software program, and the program is stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, or a hard disk that can be read by the computer. It can be executed by a computer by storing and reading from it. The program can be downloaded from another computer via a communication line, and can be installed and executed.
[0035]
【The invention's effect】
As described above, the present invention performs urban area extraction focusing on texture characteristics in urban areas, that is, characteristics in which the edge power is strong and the edge tangent direction tends to be orthogonal in the urban area. For this reason, it is difficult to extract a closed area stably, and even in the case where there is a problem in the conventional method, it is possible to stably extract an urban area.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of an urban area extraction processing device according to the present embodiment.
FIG. 2 is a diagram illustrating a processing flow of an urban area extraction processing device.
FIG. 3 is a diagram illustrating an example of a circular block.
FIG. 4 is a diagram illustrating an edge vector calculation method.
FIG. 5 is a diagram illustrating an example of edge vector information.
FIG. 6 is a diagram for explaining an example of edge tangent parameters;
FIG. 7 is a diagram illustrating an example of voting processing.
FIG. 8 is a diagram illustrating an example of a voting space.
FIG. 9 is a diagram illustrating an example of an angle histogram.
FIG. 10 is a diagram illustrating a second peak search process.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Urban area extraction processing apparatus 10 Input image memory | storage part 11 Block acquisition process part 12 Edge tangent parameter calculation process part 13 Hough conversion process part 14 Voting space memory 15 Urban area determination process part 16 Urban area area information

Claims (6)

A method for automatically extracting a city area from a photograph of the ground surface by a computer,
The computer is
A process for setting a block of local area around the image points,
A process of calculating the edge tangents respectively at each point in the block,
Performs Hough transform on the calculated edge tangent angle parameter and the distance parameter, the process of creating a two-dimensional voting space for the two parameters and 2 axes voting space memory,
Urban area extracting processing method, wherein a representative point of the block based on the voting space of the voting space in memory and a determining process whether the city. In the urban area extraction processing method according to claim 1,
An urban area extraction processing method, wherein an edge power value is used as a vote value when performing a Hough transform on the edge tangent parameter. In the urban area extraction processing method according to claim 1 or 2,
In determining process whether the representative point of the block is urban, the two-dimensional voting space angular parameters axis direction is degenerated into a one-dimensional histogram takes the projection, representative of the blocks based on the histogram An urban area extraction processing method characterized by determining whether or not a point is an urban area. A device that automatically extracts a city area from a photograph of the ground surface by a computer,
Means for setting a block in a local area centered on each point of the image;
Means for calculating an edge tangent at each point in the block;
Means for performing a Hough transform on the calculated angle parameter and distance parameter of the edge tangent, and creating a two-dimensional voting space having the two parameters as two axes;
Means for determining whether or not a representative point of the block is an urban area based on the obtained voting space.   An urban area extraction processing program for causing a computer to execute the urban area extraction processing method according to any one of claims 1 to 3. A computer-readable recording medium for an urban area extraction processing program, wherein a program for causing a computer to execute the urban area extraction processing method according to any one of claims 1 to 3 is recorded. .

Images