KR101586568B1 - Planning method of stereo CCTV for 3D map construcing and apparatus thereof - Google Patents

Abstract

The method includes the steps of displaying image data corresponding to the region to be analyzed by receiving data on the region to be analyzed, setting a monitoring candidate region in the image data, A step of setting a plurality of virtual cameras corresponding to the maximum number of layouts at arbitrary positions in the monitoring candidate area, Searching the virtual camera for each monitoring condition for maximizing a total sum of visible areas for buildings existing in the monitoring candidate area using the monitoring condition as a variable; And if the total sum of the visual areas is greater than or equal to the threshold value, And providing information for each of the plurality of virtual cameras to the three-dimensional map.
According to the method and apparatus for designing a stereo camera for constructing the three-dimensional map, more accurate and precise data can be obtained by providing a design environment in which a stereo camera suitable for an outdoor environment can be arranged at an optimal position, There is an advantage that it can be improved.

Description

[0001] The present invention relates to a stereoscopic camera,

The present invention relates to a stereo camera design method and apparatus for constructing a three-dimensional map, and more particularly, to a stereo camera design method and apparatus for constructing a three-dimensional map. More specifically, the present invention relates to a stereo camera suitable for an outdoor environment for building an HLM (Hyper Live Map) And more particularly, to a stereo camera design method and apparatus for three-dimensional map construction.

In order to extract real-time video data of a building in real time, CCTV cameras that can be monitored in a fixed state can be used. In addition, two pairs of cameras can be placed side by side, and a 3D cloud point of the building can be acquired when image processing is performed.

To extract the 3D cloud point for real-time digital map (RDM) construction, it is possible to utilize the existing CCTV and various multi-sources. Generally, in order to improve the quality of the data obtained from the camera, the surveillance performance of the camera with respect to the surveillance target area or the building should be high, and the surveillance performance of the camera is greatly influenced by the installation position, arrangement and angle.

Therefore, at the time of building the RDM, each of the cameras should be disposed at an optimal position to enhance the monitoring performance as a whole. However, there is a problem in that a method of searching for an optimum position while visually checking and collating installed images after installing a camera actually wastes time, manpower, and cost.

Therefore, it is necessary to construct an experimental environment that can test the optimal placement of the camera by the installation target zone in advance before installing the camera. If such an experimental environment is established, the deployment efficiency of CCTV will be improved.

The technology which is a background of the present invention is disclosed in Korean Patent Laid-Open Publication (KOKAI).

It is an object of the present invention to provide a stereo camera design method and apparatus for three-dimensional map construction that can provide a design environment in which a stereo camera suitable for an outdoor environment can be disposed at an optimum position.

The method includes the steps of receiving and displaying image data including geographical information on an area to be analyzed, setting a surveillance candidate area in the image data, displaying image data including geographical information on the area to be analyzed A step of setting a plurality of monitoring conditions for a virtual camera to be installed in the surveillance candidate region to be displayed as a variable; setting a maximum number of layouts of the virtual cameras; A step of searching the virtual camera for the surveillance condition for maximizing the total sum of visible areas for buildings existing in the surveillance candidate area with the surveillance condition as a variable, Comparing the sum of the visible areas with a preset threshold value, and comparing the visible area If the sum is more than the threshold value, and provides a stereo camera designed for building a three-dimensional map comprises the step of providing the information of the search condition monitored by the virtual camera.

The plurality of monitoring conditions may include an installation position, an installation altitude, an angle of view, and an azimuth angle of the virtual camera.

Also, if the total sum of the visible areas is less than the threshold value, the optimal arrangement number may be increased and the search may be re-executed.

The step of searching the virtual camera for the monitoring condition for maximizing the total sum of the visible areas of the buildings may include extracting images of the buildings existing in the monitoring candidate area from the input image data, Acquiring the monitoring condition for maximizing a ratio of the total area of the building areas monitored by the virtual cameras to the total area of the buildings, comparing the total of the visible areas with a predetermined threshold value The step may compare the ratio with the threshold value.

A video data providing unit for receiving and displaying video data including geographical information on a region to be analyzed; a surveillance region setting unit for setting a surveillance candidate region in the video data; A plurality of virtual cameras corresponding to the maximum number of layouts are set in the monitoring candidate area, and a plurality of virtual cameras corresponding to the maximum number of layouts are set in the monitoring candidate area An optimal placement unit for searching the virtual camera for each monitoring condition for maximizing a total sum of visible areas for buildings existing in the monitoring candidate area using the monitoring condition as a variable; A threshold value comparing unit for comparing the sum of the visible areas with a preset threshold value, Group is more than the threshold value, and provides a stereo camera design apparatus for building a three-dimensional map including a service allocation information to provide information of the search condition monitored by the virtual camera.

Here, the optimal arrangement unit may extract an image of the buildings existing in the surveillance candidate area from the input image data, and then extract an image of a building area monitored by the virtual cameras with respect to a total area of the buildings , And the visible area comparison unit may compare the ratio with the threshold value.

According to the method and apparatus for designing a stereo camera for constructing a three-dimensional map according to the present invention, more accurate and precise data can be obtained by providing a design environment in which a stereo camera suitable for an outdoor environment can be arranged at an optimum position There is an advantage that data quality can be improved.

FIG. 1 is a screen configuration diagram of a program according to an embodiment of the present invention.
2 is a block diagram of a stereo camera designing apparatus for three-dimensional map construction according to an embodiment of the present invention.
3 is a flowchart of a stereo camera design method using FIG.
FIG. 4 is a diagram illustrating a screen illustrating step S320 of FIG. 3. FIG.
Figure 5 shows an example of a camera in six different states with azimuth and angle of view as variables.
6 is a diagram illustrating an example of a screen showing step S340 of FIG.
FIG. 7 is a diagram illustrating an optimal layout result through step S350 of FIG. 3. FIG.
FIG. 8 is a diagram illustrating a screen in which text information corresponding to the optimal placement result of FIG. 7 is output.
FIG. 9 is an exemplary diagram of an optimal placement result using the embodiment of the present invention.
10 is a conceptual diagram for calculating position coordinates of two cameras from an optimal monitoring position of a stereo camera according to an embodiment of the present invention.
FIG. 11 shows image data obtained by arranging an actual CCTV using the result of FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention.

The present invention relates to a method and apparatus for designing a stereo camera for three-dimensional map construction, and provides a design environment for optimal placement of a camera in the construction of an HLM (Hyper Live Map) based on RDM (Realtime Digital Map).

RDM research aims to acquire real time images using CCTV, which is a state-of-the-art urban infrastructure, and to provide three-dimensional spatial information by supplementing closed area images of CCTV images using an online image (Add Image). Here, the add image may refer to a mobile phone image, a black box image, an aerial photograph, and the like. Stereo When overlapping 3D Cloud Point created by CCTV (stereo camera) through image processing is overlapped, it is possible to connect with adjacent CCTV 3D Cloud Point.

In this embodiment, it is aimed to optimally arrange a stereo camera for real-time map construction. To this end, embodiments of the present invention provide an analysis tool for optimal placement of a camera. Before describing the present invention in detail, the function of the analysis tool will be briefly described as follows.

2 shows a screen configuration of an analysis program according to an embodiment of the present invention. The tool box in Fig. 2 is a window showing the contents in which the CCTV is arranged. It shows detailed information such as name, location, azimuth angle, angle of view, installation height, visible area, overlapping area and DMZ area of the CCTV to be installed. The bottom bar provides the ability to identify the building area within the aerial image. And the surveillance area and environment setting part can set the CCTV surveillance area, and can control the weather condition, the ground height, the detection azimuth angle, and the detection angle of view. And the candidate area selection part is a part that analyzes the setting area and environment setting value in a single or a batch. Open candidate area analysis performs save and recall of candidate areas. Optimal placement and probabilistic analysis are necessary for optimal placement and analysis of the result data. Finally, aerial photographs represent aerial photographs. Candidate regions are directly selected, and result data can be visually visualized.

Hereinafter, embodiments of the present invention will be described in detail. FIG. 3 is a block diagram of a stereo camera designing apparatus for three-dimensional map construction according to an embodiment of the present invention, and FIG. 4 is a flowchart of a stereo camera designing method using FIG.

3 and 4, a stereo camera designing apparatus 100 for three-dimensional map construction according to an exemplary embodiment of the present invention includes an image data providing unit 110, a surveillance area setting unit 120, An optimal placement unit 150, an optimum placement unit 150, and a placement information providing unit 170. The placement information providing unit 130 may be a database,

First, the image data providing unit 110 receives and displays image data including geographical information on an area to be analyzed (S310). The step S310 is a process of receiving a data file for a region of interest in which a camera is to be disposed.

The input data file can be an aerial image, a satellite image, and the like, and includes not only simple image information but also geographical information. The geographical information may include a DEM (digital elevation data) image file, a numerical elevation data texture file, an image texture file, a building area area classification file, and the like. The input information takes into account the actual coordinates, topographical and environmental factors of the area where the camera is to be arranged.

Table 1 is an example of a data file input at the step S310.

File name Contents KonkukDEM.img DEM image file KonkukDEMTexture.dds Texture file ImageTexture.dds Video texture file DEM.raw Konkuk University DEM 1m Digital Elevation Model Kku_dmz.raw Konkuk University building area area file Side.raw Basic classification of area for superposition of Konkuk University building area Compass.dds Compass texture image BuildingArea.raw Area of building area

In addition, the input information may include a resource file. The resource file may correspond to an arrow model file for a DirectX-specific GUI, a basic effect file used by the DirectX Ut, or a texture used for GUI controls (buttons, list boxes, etc.) used in the DirectX Ut.

Next, the surveillance area setting unit 120 sets a surveillance candidate area within the video data (S320). The setting of the surveillance candidate area can be randomly selected based on the road line with reference to the geographic information in the image.

Since the road line is included in the geographical information, the surveillance candidate area can be selected at random by referring to the road line from the information input in step S310. Surveillance may be a building around the road, etc.

Of course, the surveillance candidate area can be selected by using functions such as weather condition, installation altitude, installation azimuth, detection angle, detection distance, overlap area, road line extraction, etc. of the area to be analyzed.

FIG. 4 is a diagram illustrating a screen illustrating step S320 of FIG. 3. FIG. In the case of FIG. 4, the surveillance candidate region is indicated as green at Konkuk University site. In Fig. 4, 1000 observation candidates are selected based on information such as road lines. Since the individual monitoring candidate areas are very small, several surveillance candidate areas are overlapped on the screen. Of course, in step S320, the monitoring candidate area may mean a whole area including the above 1000 areas.

The surveillance candidate region may be arbitrarily set as a surveillance candidate region from the user in the aerial image, and the surveillance candidate region may be selected as a random number based on the road line included in the set site. In addition, the selectable area (area where the road line exists) corresponding to the mouse movement position may be provided overlapping with a specific surface, line, image, or the like in advance. Embodiments related to the setting of the surveillance candidate region may be more various.

On the other hand, in the case of this embodiment, as a tool to test (simulate) the optimal arrangement of the stereo camera before actually installing the stereo camera, the camera used for the test corresponds to a virtual camera other than a real camera.

Then, the monitoring condition setting unit 130 sets a plurality of monitoring conditions for the virtual camera to be installed in the set monitoring candidate area as variables (S330). Here, the plurality of monitoring conditions may include the installation position of the virtual camera, the installation altitude, the view angle, and the azimuth angle.

In general, CCTV can not measure all 360 ° of the deployed area. In addition, even in the same camera, the area to be measured may be different depending on the azimuth angle, angle of view, installation position, height, etc. of the camera.

When a plurality of monitoring conditions are combined, various numbers (combination of variables) may be derived. The optimal combinations of variables that can provide optimal monitoring performance among various combinations of variables may be searched by applying the above combinations of variable numbers to arbitrary virtual cameras installed in arbitrary locations.

If the surveillance condition of the surveillance camera is set in various variables in advance as described above, the test can be performed for each surveillance condition. Actually, in FIG. 4, 1000 candidate areas in which the angle of view is fixed at 50 degrees and various azimuth angles are arbitrarily designated in each virtual camera are displayed. Of course, in this embodiment, the cameras are not arranged in all of these areas, and the optimal camera placement results having the optimum number of layouts and optimal monitoring conditions (position, altitude, azimuth angle, etc.) are provided through simulation.

Figure 5 shows the camera arrangement by azimuth and angle of view in an embodiment of the present invention. Figure 5 shows an example of a camera in six different states with azimuth and angle of view as variables. Of course, in this embodiment, the angle of view of the virtual camera is fixed at 50 °, and the azimuth angle is set in the range of 0 to 359 °. Of course, the present invention is not limited thereto.

Then, the arrangement number setting unit 140 sets the maximum number of arrangement of the virtual cameras from the user (S340). In other words, it is inputted how many virtual cameras are to be arranged with respect to the set monitoring candidate area. At this time, the number of cameras to be arranged is selected in consideration of cost and various variables.

For example, the user selects the number of stereo CCTVs that can be installed in maximum considering the regional features, limitations, and economy, and inputs the number of iterations of the algorithm to increase accuracy.

In the case of the algorithm repetition times, this means the maximum repetition times in this embodiment. The maximum number of repetitions is determined by arranging the maximum number of layouts that are set up in the virtual candidate region, and if the monitoring performance is below the threshold, the desired performance is not satisfied. Therefore, do.

6 is a diagram illustrating an example of a screen showing step S340 of FIG. FIG. 6 shows an example in which the maximum number of layouts is set to 30. If the maximum number of arrangement is set to 30 for 1000 surveillance candidate areas, there are many areas where the virtual camera is not arranged. However, a small number of cameras may or may not provide the most images of the university site. In other words, if there is a blind spot with only 30 batches of test results, or if it is difficult to obtain a clear image, the test may be repeated by increasing the number. The number of repetitions is set to 100. In this case, if you are dissatisfied after verifying your 30s, you can retest by increasing the number of cameras from 31 to 130.

Next, the optimal placement unit 150 places a plurality of virtual cameras corresponding to the maximum number of layouts at arbitrary positions in the monitoring candidate area, and sets the monitoring condition (ex, installation position, installation height, angle of view, The optimum monitoring condition for maximizing the total sum of the visible areas of the buildings existing in the surveillance candidate area is searched for by the virtual camera at step S350.

In the case of the present embodiment, in the monitoring condition item, the angle of view was fixed to 50 degrees as described above, and the azimuth angle gave various variables. Of course, the placement position and height of the camera are also applied as variables of the monitoring condition.

In step S350, the optimal monitoring condition is searched while changing the monitoring condition for the 30 virtual cameras by simulation. Here, the optimal monitoring condition of each virtual camera means a condition for maximizing the sum of the visible areas of buildings by each virtual camera.

Here, the optimal arrangement unit 150 compares the total sum of the visible areas with a preset threshold value (S360). That is, it is determined whether the total sum of the visible areas is greater than or less than the threshold (S370).

As a result, if it is less than the threshold value, the optimum arrangement number is increased and then the search is performed again. That is, the placement number setting unit 140 adjusts the optimal placement number, for example, it can increase to 31 from 30 existing ones (S380). The maximum number of batches that can be increased (the number of retests) is determined by the maximum number of iterations.

On the other hand, if the sum of the visible areas is greater than or equal to the threshold value, the arrangement information providing unit 170 provides the information of the searched monitoring condition for each virtual camera. That is, the optimal placement result can be provided on the screen in the form of video, text, and the like.

In the present embodiment, iterative experiments are performed until the target threshold value is reached, and a genetic algorithm can be used for this purpose. This is a representative field of evolutionary computation, one of the research methods used in the simulation. It is an algorithm that iterates the four processes of selection, intersection, variation, and substitution until the target or frequency is reached. Batch analysis can be performed.

FIG. 7 is a diagram illustrating an optimal layout result through step S350 of FIG. 3. FIG. FIG. 7 corresponds to the result of an image in which thirty cameras corresponding to the maximum number of arrangement in the monitored area shown in FIG. 6 are arranged under the optimum monitoring condition. Here, the optimum angle of view and azimuth angle of each virtual camera can be visually confirmed from the shape of the fan-shaped shape.

In this embodiment, when the condition values (the maximum number of layouts and the maximum number of repetitions) are specified and the optimum arrangement program is executed, a desired number of virtual cameras are arranged at optimal positions corresponding to the condition values, Lt; / RTI > Of course, the resulting value can also be provided in the form of text. FIG. 8 is a diagram illustrating a screen in which text information corresponding to the optimal placement result of FIG. 7 is output. This allows you to see the location coordinates, visible area, and left and right coordinates for each virtual camera.

In operation S 350, the monitoring condition for maximizing the total sum of the visible areas of the buildings is searched for each virtual camera. The monitoring images are extracted from the input image data, And obtains the monitoring condition for maximizing the ratio of the total area of the visible areas of the building area monitored by the virtual cameras to the total area of the buildings.

That is, depending on the monitoring condition (installation position, altitude, azimuth angle, and angle of view) of the camera, the entire building may come in view of the camera, or only a part of the building may come in. In other words, the camera may capture only 10% of the building area or more according to monitoring conditions. In order to analyze the visible region of the camera, there are a line drawing method, a shadowing method, etc. In this embodiment, the visible region is calculated using the Breshenem method of the line drawing algorithm.

In this embodiment, the ratio of the sum of the visible areas monitored by each camera can be calculated in comparison with the total area of all the buildings to be monitored by the cameras, and when the monitoring condition is changed By computing these, we can find the surveillance condition of cameras that can maximize the ratio. It is clear that the higher the ratio, the higher the monitoring efficiency and performance. If the ratio is compared with the threshold value and then the threshold value is satisfied, the results shown in FIGS. 7 and 8 can be provided.

That is, in the present embodiment, only the area of the optimal placement part in the optimal placement progression at the entire candidate points is shown on the map, and the method of FIG. 7 in which a new window is opened, It is possible to provide all the methods of Fig. 8, which show numerical values such as the ratio of the visible area to the area, in text or the like. As described above, in this embodiment, it is possible to select conditions such as the position of the camera using the ratio of the building area visible by the camera to the entire building area.

FIG. 9 is an exemplary diagram of an optimal placement result using the embodiment of the present invention. Of the 1000 candidates, 80 candidate dormitories were found in the university dormitory area. The total area of the dormitory buildings was found to be 2461 by checking the optimal layout of each dormitory building. Optimized placement of 80 candidate points resulted in an optimization curve. As a result, we could see that the increase in CCTV layout was greatly reduced by 8 points. In addition, the left / right position coordinates of the actual stereo camera can be obtained based on the extracted optimum position, which can be used as a reference for placing the stereo CCTV.

Hereinafter, a method of acquiring the positional coordinates of the two cameras using the extracted optimal position will be described. 10 is a conceptual diagram for calculating position coordinates of two cameras from an optimal monitoring position of a stereo camera according to an embodiment of the present invention. The stereo camera is implemented with two left and right cameras. The optimum monitoring position obtained through the test according to the present embodiment corresponds to a point (A) where the two cameras cross each other. In addition, the visible area by the stereo camera may mean the entire overlapping area of the two cameras. Using the intersection points of the two cameras, the coordinates (B, C) of the two cameras can be calculated. The coordinates (B, C) of the two computed cameras are used as placement positions of the two cameras in the installation work of the actual camera.

Here, the distance between the two cameras is considered to be 1.1 m. Of course, this is changeable. In Fig. 10, 0.55 corresponds to half of the interval. In the formulas of B and C, 25 corresponds to half of the angle of view 50 used.

Using the formula shown in FIG. 10, the coordinates of the points B and C varying as the azimuth angle increases from the azimuth angle of 270 degrees in the coordinates of the point A are as follows. Here, the values at which the azimuth angle value exceeds 360 degrees can be considered as 360 degrees = 0 degrees, 450 degrees = 90 degrees, and 540 degrees = 180 degrees.

azimuth The formula [A = (a, b)] 270˚

360 ° (= 0 °)

450 ° (= 90 °)

540 ° (= 180 °)

The coordinates of B and C are B = ((α + T cos (25-n), (β + Tsin + Tsin (335-n)) This calculation formula is based on the azimuth angle of 270 degrees, and as the azimuth angle increases, the n value increases to 0, 1, 2, 3, ... The positions B and C of the two cameras can be automatically calculated by the simple mathematical formula described above.

FIG. 11 shows image data obtained by arranging an actual CCTV using the result of FIG. 11 is based on CCTV left and right coordinates (TM), azimuth angle, azimuth (50 degrees), and installation altitude according to the optimal arrangement. The result of the CCTV shooting experiment of Konkuk University dormitory building was able to shorten the shooting time and obtain more accurate data. As a result of CCTV shooting like this, it is possible to obtain the image data of the optimal area.

So far, there is no standard related to Stereo CCTV deployment for 3D Cloud Point acquisition of buildings, so there is a need for optimal placement technology for this. In order to solve this problem, input data input, candidate area selection, monitoring condition setting, optimum layout number designation, optimal layout, and visualization of result data extraction are sequentially performed in this embodiment. Finally, the location of the camera optimally arranged in the entire image and the building monitoring ratio of the photographed image are shown. By efficiently installing the camera, the embodiment of the present invention can be utilized as an underlying technology of real-time 3D MAP implementation.

According to the method and apparatus for designing a stereo camera for constructing a three-dimensional map according to the present invention, a more accurate and precise data can be obtained by providing a design environment in which a stereo camera suitable for an outdoor environment can be arranged at an optimum position And can improve data quality.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

100: Stereo camera design device for 3D map construction
110: video data providing unit 120: surveillance area setting unit
130: Monitoring condition setting unit 140:
150: optimal placement unit 160: visible area comparison unit
170:

Claims (8)

Receiving and displaying image data including geographical information on an analysis target area;
Setting a monitoring candidate region in the video data;
Setting a plurality of monitoring conditions for a virtual camera to be installed in the monitoring candidate region as variables;
Setting a maximum number of layouts of the virtual cameras;
A plurality of virtual cameras corresponding to the maximum number of layouts are arranged at arbitrary positions in the monitoring candidate region, Searching the monitoring condition by the virtual camera;
Comparing the total sum of the visible areas with a preset threshold value; And
And providing information of the searched monitoring condition for each virtual camera when the total sum of the visible areas is greater than or equal to the threshold value. The method according to claim 1,
Wherein the plurality of monitoring conditions include:
A method of designing a stereo camera for constructing a three-dimensional map including an installation position, an installation altitude, an angle of view, and an azimuth angle of the virtual camera. The method according to claim 1,
And increasing the maximum number of layouts until the sum of the visible areas satisfies the threshold value or more, if the total sum of the visible areas is less than the threshold, And repeating the re-performing operation through the step of performing the three-dimensional map construction. The method according to claim 1,
Wherein the step of searching the virtual camera for the monitoring condition for maximizing the sum of the visible areas for the buildings comprises:
Extracting an image of the buildings existing in the monitoring candidate region from the input image data,
Acquiring the monitoring condition for maximizing a ratio of a total area of visible areas of a building area monitored by the virtual cameras to a total area of the buildings,
Wherein comparing the sum of the visible areas with a predetermined threshold comprises:
And comparing the ratio with the threshold value. An image data providing unit for receiving and displaying image data including geographical information on an analysis target area;
A surveillance region setting unit for setting a surveillance candidate region within the video data;
A monitoring condition setting unit configured to set a plurality of monitoring conditions for a virtual camera to be installed in the monitoring candidate area as variables;
A layout number setting unit configured to set a maximum number of layouts of the virtual cameras;
A plurality of virtual cameras corresponding to the maximum number of layouts are arranged at arbitrary positions in the monitoring candidate region, An optimum arrangement unit for searching the monitoring condition by the virtual camera;
A visible area comparing unit for comparing the total sum of the visible areas with a predetermined threshold value; And
And a placement information providing unit for providing the information of the searched monitoring condition for each of the virtual cameras when the total sum of the visible areas is equal to or greater than the threshold value. The method of claim 5,
Wherein the plurality of monitoring conditions include:
A stereo camera designing device for constructing a three-dimensional map including an installation position, an installation altitude, an angle of view, and an azimuth angle of the virtual camera. The method of claim 5,
And increasing the maximum number of layouts until the sum of the visible areas satisfies the threshold value or more, if the total sum of the visible areas is less than the threshold, And repeating the re-performing operation through the second step. The method of claim 5,
Wherein the optimum arrangement unit comprises:
Extracting an image of the buildings existing in the monitoring candidate region from the input image data,
Acquiring the monitoring condition for maximizing a ratio of a total area of visible areas of a building area monitored by the virtual cameras to a total area of the buildings,
Wherein the visible-
And comparing the ratio with the threshold value.

Images