KR20120041819A - Method for generating 3-d high resolution ndvi urban model - Google Patents

Abstract

PURPOSE: A three-dimensional high definition normalization city model generating method is provided to develop a three-dimensional city model related to vegetation distribution of a city through a mapping technique and high-definition NDVI image map manufacturing. CONSTITUTION: A three-dimensional building model(712) is generated. An NDVI building model(713) is generated by combination of the three-dimensional building model and an NDVI(Normalized Difference Vegetation Index) categorization image(723). Other facilities are manufactured in a three-dimensional model(703). An NDVI model is generated by combination of the three-dimensional model and a normalized vegetation map image. The NDVI categorization image is overlapped with a city model(721).

Description

{Method for Generating 3-D High Resolution NDVI Urban Model}

The present invention relates to a three-dimensional high-resolution normalized vegetation index urban model generation method, and more particularly to a three-dimensional normalized vegetation index urban model production method combining the three-dimensional urban model production method and the normalized vegetation index image production method.

Drawing is a categorization of the extraction of the ideas of interest shown in the three-dimensional environment based on the performance of aerial triangulation accuracy performed through aerial photography, reference point survey and aerial image internal expression, mutual expression and absolute expression calculation process. After that, it is a process of describing the required spatial form and extracting the attribute values. Digital drawing is the computerization completion phase of drawing business, and all work related to the generation of spatial information based on image information in digital environment and materials acquired from GPS / INS system and digital module camera equipment are digitized. outstanding.

, roll), 전후요동(

, pitch) 및 수평요동(

, yaw) 등과 같은 자세의 왜곡 상태를 기록하게 된다. 레이저 스캐너에 의한 관측 시간 간격은 GPS의 1초 수신간격에 비해 매우 짧기 때문에 관측된 자료는 항공기의 자세에 따라서 영향을 받게 된다. 이러한 영향을 최소화하기 위해서는 GPS 수신간격에 비해 짧은 간격으로 센서의 자세를 측정하기 위해 INS가 사용된다.GPS / INS (Global Positioning System / Inertial Navigation System) complements the weaknesses and strengths of both systems by combining the interpretation and correction of location information from GPS and INS's attitude information recording technology to obtain accurate position and attitude. The system is increased. GPS is a state-of-the-art navigation system that uses satellites to determine the position of sensors. It is a system that can measure position, velocity and time in a reference coordinate system regardless of climate anywhere on earth. The INS system is a device that measures postures in three axes using a three-axis gyro and an Inertial Measurement Unit (IMU) with an accelerometer. To determine the facial expressions of an aircraft, the INS simulates the fluctuations in the aircraft during flight.

, roll), swing back and forth (

, pitch) and horizontal fluctuations (

, yaw) and the like to record the distortion state of the posture. Since the observation time interval by the laser scanner is very short compared to the GPS one second reception interval, the observed data is affected by the attitude of the aircraft. To minimize this effect, INS is used to measure the attitude of the sensor at shorter intervals than the GPS interval.

Aerial digital surveying refers to a process of imaging and acquiring electromagnetic wave information of the earth's surface by attaching a digital image acquisition device represented by an aerial digital modular camera to an aircraft. Aviation digital cameras consist of more than eight multiple cameras. Electromagnetic wave information incident from the sun and reflected from the earth's surface is analyzed and informed by the magnitude of the signal in each of eight or more CCD sensors. The camera can generate high resolution color images through image fusion from a group of cameras (more than 4) for generating high resolution panchromatic images through mosaic and high resolution aerial photographs. It consists of a group of cameras (four or more).

The 3D city model is a model of the real city or the unreal city in a complex real world in a three-dimensional form so that it can be easily recognized. Methodologically, the real world can be constructed by using coordinate and projection-based survey data, and the unreal world can be divided into virtual coordinate and projection-based construction. In the past, 3D modeling was performed based on the observation of real-world surveying and three-dimensional description construction based on inefficiency. In recent years, the development of surveying technology and digitalization through digital photography have been performed. Although the actual city is created and reviewed through the development of technology, the city model is generated using only the aerial image extracted using the visible light band sensors represented by the red (R), green (G), and blue (B) wavelengths. Only the three-dimensional structure visible to the human eye can be identified, and the distribution of green resources in the city can be identified only with green photographs. The structure of the city is very complicated in form, but it is also very diverse in terms of the materials that make up the city. When you see the city through visible light, you can see the distinction between the green artificial cloth soccer field and the natural grass soccer field. It is very difficult to distinguish rooftops of buildings painted with paints and rooftops made of paved grass and plants. This is because all eyes are recognized as green information. Recently, green is also used for plastics and rebars. When surveying the current urban greening conditions for urban greening planning and design from natural color aerial photographs It is an important part that causes errors in judgment. In addition, in recent years, the structure of cities has become higher and three-dimensional, making it more difficult to investigate.

Therefore, there is a growing need to develop a technology capable of accurately indicating and quantifying the distribution state of three-dimensional and green resources by integrating three-dimensional city green resource distribution information with 3D city model construction technology.

Therefore, as a result of our efforts to develop 3D urban modeling technology into a technique for producing a normalized vegetation index capable of quantitatively grasping the green resources of a city, the present inventors have made numerical diagrams using images acquired through aerial digital photography. After the implementation, by combining normalized vegetation index extracted through the calculation between bands from multiple images, it was confirmed that it is possible to accurately classify and quantify a material prone to error of determination from color images, thereby completing the present invention.

An object of the present invention is to provide a method for generating a three-dimensional high-resolution normalized vegetation index (NDVI) urban model that can grasp the green resources of a city in a three-dimensional environment.

In order to achieve the above object, the present invention comprises the steps of (a) acquiring a multi-band image using an aerial digital module camera; (b) acquiring the shooting time at the time of acquiring the image using the aviation digital module camera, the position and attitude information of the aircraft and the camera through a GPS / INS system, and calculating an external expression element; (c) performing aerial triangulation using the external expression elements and ground control point survey results; (d) generating a three-dimensional urban model by performing numerical mapping based on the results of steps (a) to (c); (e) generating a high resolution NDVI image from the multi-band image; And (f) combining the generated high resolution NDVI image with a 3D city model to generate a 3D high resolution NDVI city model.

According to the present invention, by developing and distributing 3D city model related to vegetation distribution of a city by using a technology of creating a 3D city model and a high resolution NDVI image mapping and mapping technology using a digital camera, visible light that can be read by the eyes The distribution structure for vegetation outside the band can be identified along with the three-dimensional structure of the city. This program helps to grasp the utilization of the rooftop part, which is the best available place for the three-dimensional urban structure, and to identify the three-dimensional green resource information of the city that cannot be obtained from the planar information. It will contribute to providing 3D urban green information for engineering.

1 is a flowchart of a method for manufacturing a 3D normalized vegetation index urban model according to an embodiment of the present invention.
2 is a flowchart illustrating a process of acquiring a multi-band high resolution aerial photograph with a aviation digital module camera.
3 is a flowchart illustrating a process of calculating an accurate exterior orientation (EO) element through position correction using aviation GPS / INS data.
4 is a flowchart illustrating a process of generating a 3D city model through aerial triangulation and numerical mapping.
5 is a flowchart illustrating a process of generating a normalized vegetation index image using high resolution aerial photograph multi-image data.
6 is an explanatory diagram comparing the difference between the normal aerial image and the normalized vegetation index image.
FIG. 7 is a flowchart illustrating a method of manufacturing a 3D high resolution normalized vegetation index urban model in which a high resolution NDVI image is combined with a 3D urban model constructed through numerical mapping.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart of a method for manufacturing a 3D normalized vegetation index urban model according to an embodiment of the present invention.

As shown in Figure 1, the method for producing a three-dimensional high-resolution normalized vegetation index urban model according to the invention step of obtaining a multi-band image data using an aviation digital camera 101, GPS / INS signal and individual image capture Computing the position and posture correction through the calculation by combining the position (102), through the process of analyzing the plurality of image data position three-dimensional alignment (103), the positional interconnection with adjacent images Numerical determination phase 104, ground control points (GCPs) and absolute representation phase 105 to determine the positional interconnection of each image, numerical representation phase to describe the shape of a three-dimensional building in a three-dimensional environment (106), a three-dimensional building model database construction step incorporating numerical description results (107), a three-dimensional three-dimensional city model generation step using a terrain (DEM) and a three-dimensional building model (108) and combines a high-resolution NDVI image Three-dimensional high-resolution normalized vegetation index city model generation step 109 can be configured.

That is, the present invention comprises the steps of: (a) acquiring a multi-band image using an aerial digital module camera; (b) acquiring the shooting time at the time of acquiring the image using the aviation digital module camera, the position and attitude information of the aircraft and the camera through a GPS / INS system, and calculating an external expression element; (c) performing aerial triangulation using the external expression elements and ground control point survey results; (d) generating a three-dimensional urban model by performing numerical mapping based on the results of steps (a) to (c); (e) generating a high resolution NDVI image from the multi-band image; And (f) combining the generated high resolution NDVI image with the 3D city model to generate a 3D high resolution NDVI city model.

,

,

,

,

,

)를 산출하는 것이고, 상기 외부표정 요소와 지상기준점 측량 성과(

,

,

)를 절대기준으로 참조하는 항공삼각측량을 수행하면 외부표정 요소의 자세부분을 교정하는 상호표정 (

,

,

,

,

,

)및 외부표정 요소의 위치부분을 교정하는 절대표정(

,

,

,

,

,

)을 산출할 수 있다.In more detail, as is generally known, the calculation of the external expression element of step (b) may include mechanical external expression element (

,

,

,

,

,

), And the external expression factors and ground control point survey performance (

,

,

When performing an aerial triangulation referencing the absolute reference, the cross-reference (

,

,

,

,

,

) And absolute expressions to correct the positional part of the external expression element.

,

,

,

,

,

) Can be calculated.

2 is a flowchart illustrating a process of acquiring a multi-band high resolution aerial photograph with an aerial digital module camera.

In the present invention, the aviation digital module camera is characterized in that it comprises a camera control device 201, the camera body 202, and the lens cone, the lens cone 203 surrounding the lens has four full color (panchromatic) camera head 204 and four multispectral camera heads 205. The wavelength band and the number of sensors of each band of the digital image obtained using the aerial digital module camera are shown in Table 1 below.

Camera band images 400-580 nm Number of sensors Blue 400-580 nm One Green 500-650 nm One Red  590-675 nm One Near IR 740-850 nm One Panchromatic 400-580 nm 4

The four panchromatic camera heads 204 are configured in a 2 × 2 array, with each camera looking backward left (B / L) looking, backward right (B / R) looking, forward left (F / L). ) Looking, forward right (F / R) It is inclined little by little so that the target area can be photographed and each image 211 can be acquired at the same time as the photographing. Since photographing four sub-images is to compensate for the disadvantage that the area of one CCD is small, it is preferable to have overlapping areas 212 of about 1% of each other. Tie points 213 are generated in the overlapped region, and an image 214 is generated through a systematic mosaic.

The high-resolution black and white images 215 generated in this way are used for a 3D city model manufacturing work through aerial triangulation and numerical mapping.

In addition, the four multi-band camera heads 205 are configured in a 2 × 2 array, and can acquire red, green, blue, and near infrared band information, respectively. The images 221 of the green, red, blue, and near infrared wavelength bands acquired through each camera head are generated as multi-band images 222 combined into one image in which four bands are combined, and then normalized. A vegetation index (NDVI) image is generated (223) and used to combine with the 3D urban model.

In the present invention, the aviation digital module camera is set to acquire images of red, green, blue, and near infrared bands in addition to the black and white images during aerial photography. Since the three-dimensional change is severe as the city develops, it is preferable that about 75% overlap in the flight direction and about 50% overlap in the adjacent direction to generate the 3D city model. In unusual areas such as high-rise buildings, separate redundancy should be applied. For example, in high-rise buildings such as the Terran Road area in Seoul, it is difficult to obtain elevation information of buildings due to the vertically developed spatial structure. The image information of the elevation may be acquired through separate shooting such as shooting.

The aviation GPS / INS system uses the video ID, time information, camera position information (X, Y, Z), and posture information (ω) to be recorded at the moment the digital module camera shutter, which is an aviation sensor system, is opened. , φ, κ), and the geocoordinate system used are recorded on the storage medium along with the images.

,

,

,

,

,

) 산출과정을 설명하면 다음과 같다. 먼저 항공 GPS/INS 시스템에서 항공 GPS/INS 원시 데이터 (301)가 생성된다. 다중 밴드 영상은 고속으로 움직이는 비행기에서 촬영되므로 정밀한 수치도화를 위해서는 GPS 위치보정이 수행되어져야 하며 이를 위해 항공 GPS/INS 원시 데이터는 새로운 GPS 데이터(302)와 INS 데이터(303)로 물리적으로 분할된다. 이 중 항공 GPS 데이터는 동일 시간에 지상에 미리 고정되어 세워져 있던 지상 기준국 GPS로부터 수신된 데이터(304)와 결합하여 항공기의 정확한 자세를 보정하여 계산한 Differential GPS 데이터 (305)를 생성하며 이는 항공 GPS 데이터 보정에 대한 최종 해(解)가 된다. 상기 Differential GPS 데이터 (305)는 항공 INS 데이터(303)와 다시 결합하여 보정이 완료된 최종 GPS/INS 데이터(306)를 생성하게 되며, 이로부터 항공삼각측량에 필요한 GPS/INS 외부표정요소(307)를 산출하게 된다. 3 is a flowchart illustrating a process of calculating a mechanical exterior orientation (EO) element through position correction using aeronautical GPS / INS data. Referring to FIG. 3, the external expression element (

,

,

,

,

,

The calculation process is as follows. First, aviation GPS / INS raw data 301 is generated in the aviation GPS / INS system. Since multi-band images are taken from a high-speed plane, GPS positioning must be performed for precise numerical mapping. For this purpose, aerial GPS / INS raw data is physically divided into new GPS data 302 and INS data 303. . Among these, the aviation GPS data is combined with the data 304 received from the ground reference station GPS, which is fixed on the ground at the same time, to generate differential GPS data 305 calculated by correcting the aircraft's correct attitude. This is the final solution to GPS data correction. The Differential GPS data 305 is combined with the aviation INS data 303 again to generate the final corrected GPS / INS data 306, from which the GPS / INS external representation element 307 required for aviation triangulation. Will yield.

, ,

,

,

,

) 산출 및 수치도화를 통한 3차원 도시 모델 생성과정 보여주는 흐름도이다.Figure 4 shows the corrected external marking elements (by air triangulation)

, ,

,

,

,

) This is a flow chart showing the process of creating a 3D city model through calculation and numerical mapping.

By using the aerial photograph black-and-white image 401 generated through the above process, the external expression element data 402 obtained from the GPS / INS correction result, and the ground control point performance 403 obtained through GPS / Level surveying for the target area. Perform an erial triangulation (AT) process that includes the inner orientation, relative orientation, and absolute orientation processes (404), and evaluate the accuracy of the results (405). After the step 1), the digitalization and the 3D city model generation are performed (406). In the accuracy evaluation step, if the position accuracy does not fall within the allowable range, the process returns to the aviation triangulation step, and the process of eliminating and adjusting the cause of the excessive error is stored. When the result of the aviation triangulation is within the allowable range, the relevant working environment and the final adjustment result are stored. And the relevant process is completed.

5 is a flowchart illustrating a process of generating an NDVI image using high resolution aerial photographic image data.

 In the present invention, a red band (RED) and a near infrared band (NIR) are extracted from the high resolution aerial photograph multi-band image 501 generated by the multi-band camera head 205 (502), and the equation (503) of the two bands is calculated. Calculate the NDVI for each cell, categorize it based on [Table 2], and then write a new raster image to create a high-resolution NDVI image and an NDVI categorized image 504. do.

NDVI Category Class NDVI value range Index code color Very high vegetation vitality 200 to 255 4 Blue High vegetation vitality 131 to 200 3 Green Low vegetation vitality 61-130 2 Yellow Very low vegetation vitality 0 to 60 One Red

In the present invention, the NDVI and NDVI images are obtained by obtaining a difference from two images of visible light and near infrared ray, and emphasize the reflection characteristics of vegetation, and normalize them by dividing them by the sum of the two images. Refers to the exponential and indexed image data that can be accurately identified.

In order to obtain NDVI from a multi-band image, a near infrared (NIR) wavelength band image (reflectance of 0.7 to 1.1 μm) and a red (RED) wavelength band image (reflection of 0.6 to 0.7 μm) are required. Real NDVI calculations can be obtained according to Equation 1 below. It is stored as a float image.

[Equation 1]

Since real NDVI image consumes a lot of capacity for the same information display, NDVI should be stored as integer image by calculating integer data with a range of 0 to 255, and integer NDVI calculation is calculated according to [Equation 2]. Can be.

[Equation 2]

In the present invention, since the integer NDVI image generation is performed through a ratio calculation of a red band and a near infrared band obtained at the same point of photographing, the integer NDVI image may have a normalized value with minimum deviation in all images, and vegetation vitality in the NDVI categorized image. Can be easily read through the graded colors.

6 is an explanatory diagram comparing the difference between the normal aerial image and the NDVI image.

In the color image, the bright roof 601 is identified, but the roof of the apartment 612 and the neighboring forest 611 with the green paint waterproof packaging are not clearly distinguished by the same green color. On the other hand, in the NDVI image 602, both the bright roof 601 and the rooftop 612 with green paint waterproofing are shown in strong red and yellow 622, and the forest is blue 621, so that the vegetation vitality is visual. Is read. In other words, in the color image made through the red, green, and blue bands from the aerial photo, the color of the lawn and the paint painted for waterproofing on the roof of the apartment appear in the same green color. The color value of the green roof area of the apartment is small (about 192 of ~ 255) and the color of the green roof area of the apartment is small (about 53 of 0 ~ 255).

In the present invention, the three-dimensional high-resolution NDVI urban model capable of visually reading vegetation vitality may be generated by combining the generated high-resolution NDVI image and the three-dimensional urban model, wherein the generated high-resolution NDVI image and the three-dimensional The combination of urban models can be performed by overlaying NDVI categorized images on urban models having only geometric shapes.

7 is a flowchart illustrating a method of manufacturing a 3D high resolution NDVI urban model in which a high resolution NDVI image is combined with a 3D urban model constructed through numerical mapping.

From the result of the completion of aerial triangulation, two adjacent multi-band images with sufficient redundancy are opened, and the shape of the building required to generate the three-dimensional city is processed through the numerical mapping process in the stereoscopic environment 701. ) 711 is described one by one to generate a three-dimensional building model 712 and then combined with the previously produced NDVI categorization image to generate an NDVI building model 713. Other facilities such as bridges and overpasses (702) having a three-dimensional shape also form a three-dimensional model (703) by connecting the three-dimensional shape to a divided facet in a digitalized environment, Combine to generate NDVI model 704. Finally, by placing the NDVI categorized image 723 on a city model 721 having only geometric shapes, a 3D high resolution city model capable of visually reading vegetation vitality is generated.

201: camera control unit 202: camera body
203: lens cone 204: panchromatic camera head
205: multispectral camera head

Claims (6)

3D high resolution NDVI city model including the following steps:
(a) acquiring a multi-band image using an aerial digital module camera;
(b) acquiring the shooting time at the time of acquiring the image using the aviation digital module camera, the position and attitude information of the aircraft and the camera through a GPS / INS system, and calculating an external expression element;
(c) performing aerial triangulation using the external expression elements and ground control point survey results;
(d) generating a three-dimensional urban model by performing numerical mapping based on the results of steps (a) to (c);
(e) generating a high resolution NDVI image from the multi-band image; And
(f) generating a 3D high resolution NDVI urban model by combining the generated high resolution NDVI image and a 3D urban model.
The method of claim 1, wherein the aerial digital module camera comprises a camera controller, a camera body and a lens cone.
3. The method of claim 2, wherein the lens cone comprises four panchromatic camera heads and four multispectral camera heads.
The method of claim 1, wherein the generating of the high resolution NDVI image comprises extracting a red band (RED) and a near infrared band (NIR) from the multi-band image, calculating an integer NDVI using Equation 2, and then categorizing the high resolution NDVI. Method for producing a categorized image:
[Equation 2]

5. The method of claim 4, wherein the categorization is based on Table 3 below. NDVI Categorization Rank NDVI value range Index code color Very high vegetation vitality 200 to 255 4 Blue High vegetation vitality 131 to 200 3 Green Low vegetation vitality 61-130 2 Yellow Very low vegetation vitality 0 to 60 One Red
The method of claim 1, wherein combining the generated high-resolution NDVI image and the three-dimensional urban model comprises stacking the NDVI categorized image on an urban model having only geometric shapes.

Images