A method for extracting the height of vertical objects in satellite CCD remote sensing images

Abstract

The invention discloses a method for extracting the height of an upright object on a satellite CCD remote sensing image. The method applies a strict imaging geometric relationship model of a satellite CCD remote sensing image, calculates a coordinate rotation coefficient and a projection scale coefficient according to information such as the satellite's height and quaternion data, and then uses the coordinate rotation coefficient and the projection scale coefficient to calculate the vertex image coordinates and the bottom image coordinates of the upright object to obtain the height of the upright object. It is unnecessary to extract other features of the upright object and it is also unnecessary to use two image data. Therefore, the invention can simplify the process of extracting the height information of various upright objects on a satellite CCD remote sensing image.

Description

Technical Field

[0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a method for extracting the height of vertical objects on a satellite CCD remote sensing image. Background Art

[0002] Satellite CCD (Charge Coupled Device) images acquired based on the linear array push-scanning method have the advantages of high resolution and wide ground coverage, which makes satellite linear array CCD the mainstream of optical satellite remote sensing and has been widely used. When applying satellite CCD remote sensing images, it is often necessary to analyze and extract certain thematic information of interest, and the vertical objects on satellite CCD remote sensing images are a special type of thematic objects, such as various types of industrial chimneys, various tower-type structures, transmission poles, high-rise buildings, etc. These objects usually stand vertically above the ground, and their height is much greater than their length or width. When using satellite CCD remote sensing images to analyze such vertical objects, it is often necessary to extract their height values.

[0003] At present, the methods for extracting the height of ground objects from satellite CCD remote sensing images mainly include extraction methods based on single satellite CCD remote sensing images and extraction methods based on satellite stereo CCD images.

[0004] The method of extracting the height of the object based on a single satellite CCD remote sensing image mainly utilizes the information of the shadow characteristics formed by the upright objects on the satellite CCD image. The root cause of the shadow of the upright objects on the image is the different azimuths and altitudes of the sun and the satellite. When the azimuth and altitude of the sun and the satellite are the same or the solar altitude angle is greater than or equal to the satellite altitude angle, there is no shadow on the image; when the solar altitude angle is less than the satellite altitude angle, a shadow is formed on the image, which is manifested as a low brightness value on the image, which is conducive to identification. According to the different positions of the satellite when the satellite CCD image is imaged, the shadow information on the image has a clear geometric relationship with the sun and satellite altitude angles. Therefore, the height of the object can be solved by using the clear geometric relationship between the shadow of the upright object and its altitude, the solar altitude angle and azimuth, the satellite altitude angle and azimuth and other parameters, and finally the height information of the upright object can be inferred according to the altitude angle and azimuth of the satellite and the sun when the satellite CCD remote sensing image is imaged, and the length of the shadow of the upright object on the image.

[0005] The method based on satellite stereo CCD images is a good method to extract the height of the object by stereo mapping by constructing a stereo image through two satellite remote sensing images. Moreover, with the development of satellite remote sensors, the height accuracy obtained is getting higher and higher. These satellites mostly use linear array CCD detectors to scan and image, and the imaging method is flexible. They can obtain stereo images in different orbits, or swing back and forth along the orbit direction to obtain stereo images in the same orbit. Satellites with stereo mapping capabilities mainly include IKONOS, QuickBird, GeoEye, WorldView-I / II, etc., while Japan's ALOS, France's SPOT, India's IRS-P5, Israel's EROS-B and other satellites also have stereo mapping capabilities. High-resolution satellite stereo imaging is becoming more and more mature, constantly changing the traditional way of obtaining three-dimensional information. The method of extracting the height of upright objects based on satellite stereo CCD images requires obtaining two satellite CCD images of the same object to form a stereo image pair. According to the two reverse projection rays of the same-name image points of the upright object in the two images, the height of the object point can be determined based on the principle of forward intersection. Its essence is to geometrically reverse the photography process to obtain the height of the ground object. The prerequisite for extracting the height of upright ground objects based on satellite stereo CCD images is that there must be two images (two images obtained at different positions) and there is a certain degree of overlap between them. The overlapping parts constitute a "stereo image pair". Therefore, to obtain the three-dimensional position of the ground using a stereo method, it is necessary to: (1) obtain at least two images at two positions; (2) extract the points with the same name on the two images, which is generally achieved by image matching; (3) restore the spatial orientation of the two corresponding straight lines, and finally obtain the height value of the ground object point.

[0006] It can be seen that the method of extracting the height of upright objects based on a single satellite CCD remote sensing image is essentially based on the shadow characteristics formed by the upright objects blocking the sun because they are higher than the ground. The height of the object is inverted and calculated by the characteristic length of the shadow of the upright objects. Therefore, to obtain the height of the object, there must be a shadow on the satellite CCD image, and the formation of the shadow depends on the relationship between the sun and the sun's altitude angle and the satellite altitude angle. Once there is no shadow on the satellite CCD remote sensing image, this method cannot extract the height of the object. In some cases, the actual satellite CCD remote sensing image cannot obtain the shadow of the object on the image due to various reasons. In this case, the height of the upright object cannot be extracted. The method of extracting the height of upright objects based on satellite stereo CCD images needs to extract the height of the ground in a stereo manner, that is, it is necessary to obtain two images at two locations, extract the same-name object points on the two images, and calculate the height of the object points. When there is only one satellite CCD remote sensing image, the height information of the object cannot be extracted by the stereo method.

[0007] In summary, the existing method of extracting the height of vertical objects based on a single satellite CCD remote sensing image has the problem of relying on the shadow of the object, while the method of extracting the height of vertical objects based on satellite stereo CCD images has the disadvantage of requiring at least two images for analysis. Summary of the invention

[0008] In view of this, the present invention provides a method for extracting the height of a vertical object on a satellite CCD remote sensing image, which can calculate the height of a vertical object based on a satellite CCD remote sensing image.

[0009] The present invention provides a method for extracting the height of vertical objects on a satellite CCD remote sensing image, comprising the following steps:

[0010] Extract the bottom image coordinates and vertex image coordinates of the vertical feature to be analyzed in the satellite CCD remote sensing image, and obtain the satellite altitude and quaternion data corresponding to the vertical feature to be analyzed; calculate the coordinate rotation coefficient according to the quaternion data; calculate the projection scale coefficient according to the coordinate rotation coefficient, the bottom image coordinates of the vertical feature to be analyzed and the focal length of the aerial camera; calculate the height value of the vertical feature to be analyzed according to the image coordinates of the vertical feature to be analyzed, the coordinate rotation coefficient and the projection scale coefficient.

[0011] Furthermore, the method of calculating the coordinate rotation coefficient according to the quaternion data is to use formula (1) to calculate:

[0012]

[0013] Among them, (q0, q1, q2, q3) is the quaternion value of the satellite's quaternion data, (a1, a2, a3, b1, b2, b3, c1, c2, c3) is the coordinate rotation coefficient;

[0014] According to the calculated coordinate rotation coefficients (a1, a2, a3, b1, b2, b3, c1, c2, c3), the bottom image coordinates (x1, y1) of the vertical feature to be analyzed and the focal length f of the satellite CCD camera, the coordinate transformation coefficients (l1, l2, m1, m2, n1, n2) are calculated using formula (2):

[0015]

[0016] Wherein, (x1, y1) is the bottom point image coordinate of the vertical feature to be analyzed.

[0017] The height h of the vertical object to be analyzed is calculated using formula (3) according to the image coordinates of the vertical object to be analyzed, the coordinate rotation coefficient and the projection scale coefficient:

[0018]

[0019] Among them, H s The height of the ground.

[0020] Furthermore, the method of extracting the bottom image coordinates and the top image coordinates of the vertical object to be analyzed on the satellite CCD remote sensing image is to use image analysis software to analyze the satellite CCD remote sensing image to obtain the bottom image coordinates and the top image coordinates.

[0021] Furthermore, the three-dimensional position and attitude of the satellite are obtained by reading an auxiliary data file of the satellite CCD.

[0022] Furthermore, the push-sweep line T corresponding to the vertical object to be analyzed can be calculated using formula (4):

[0023]

[0024] Among them, x1 is the image coordinate of the bottom point of the vertical feature to be analyzed, and x2 is the image coordinate of the vertex of the vertical feature to be analyzed.

[0025] Beneficial effects:

[0026] The present invention applies a strict imaging geometric relationship model of satellite CCD remote sensing images, calculates coordinate rotation coefficients and projection scale coefficients based on information such as the satellite's altitude and quaternion data, and then uses the coordinate rotation coefficients and projection scale coefficients to calculate the vertex image coordinates and the bottom image coordinates of the upright object to obtain the height of the upright object. There is no need to extract other features of the upright object, and there is no need to use two image data. Therefore, the present invention can simplify the process of extracting height information of various upright objects on satellite CCD remote sensing images. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention provides a flow chart of a method for extracting the height of vertical objects on a satellite CCD remote sensing image. DETAILED DESCRIPTION

[0028] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0029] The present invention provides a method for extracting the height of an upright object on a satellite CCD remote sensing image. The core idea of ​​the method is: using the position information of the upright object to be analyzed extracted from the satellite CCD remote sensing image to obtain the height and quaternion data of the satellite, then calculating the coordinate rotation coefficient according to the quaternion data, and then calculating the projection scale coefficient according to the coordinate rotation coefficient, the bottom point image coordinates and the camera focal length, and finally calculating the height of the upright object to be analyzed according to the image coordinates, the coordinate rotation coefficient and the projection scale coefficient.

[0030] The present invention provides a method for extracting the height of vertical objects on a satellite CCD remote sensing image, the process is as follows: Figure 1 As shown, the specific steps include:

[0031] Step 1: Use image analysis software to extract the bottom image coordinates (x1, y1) and top image coordinates (x2, y2) of the vertical feature to be analyzed contained in the satellite CCD remote sensing image, and read the satellite auxiliary data file to obtain the satellite altitude data and star sensor data, and use the interpolation method to calculate the satellite flight altitude value H corresponding to the push-sweep line T of the vertical feature to be analyzed. s And the quaternion values ​​of the star sensor (q0,q1,q2,q3).

[0032] The push-sweep line T corresponding to the vertical object can be calculated using formula (1):

[0033]

[0034] Step 2: According to the quaternion values ​​(q0, q1, q2, q3) obtained in step 1, use formula (2) to calculate the coordinate rotation coefficients (a1, a2, a3, b1, b2, b3, c1, c2, c3):

[0035]

[0036] Step 3: According to the coordinate rotation coefficients (a1, a2, a3, b1, b2, b3, c1, c2, c3) calculated in step 2, the bottom image coordinates (x1, y1) of the vertical object to be analyzed, and the focal length f of the satellite CCD camera, the projection scale coefficients K1 and K2 are calculated using formula (3):

[0037]

[0038] Among them, the focal length of the satellite CCD camera can be provided by the camera manufacturer.

[0039] Step 4: According to the image coordinates of the vertical object to be analyzed, the coordinate rotation coefficients (a1, a2, a3, b1, b2, b3, c1, c2, c3) and the projection scale coefficients K1 and K2 obtained in the above steps, the height h of the vertical object to be analyzed is calculated using formula (4):

[0040]

[0041] It can be seen that the method for extracting the height of upright objects on satellite CCD remote sensing images provided by the present invention applies a strict imaging geometric relationship model of satellite CCD remote sensing images. The geometric relationship is strict, no other features of the objects are extracted, and no two image data are used. The height of the upright objects can be directly calculated without initial value setting and iterative process.

[0042] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for extracting the height of vertical objects on a satellite CCD remote sensing image, characterized in that: The following steps are involved: Extracting the bottom image coordinates and the top image coordinates of the vertical object to be analyzed in the satellite CCD remote sensing image, obtaining the satellite height and quaternion data corresponding to the vertical object to be analyzed; calculating the coordinate rotation coefficient according to the quaternion data; Calculate the projection scale coefficient according to the coordinate rotation coefficient, the bottom image coordinates of the vertical feature to be analyzed and the focal length of the aerial camera; calculate the height value of the vertical feature to be analyzed according to the image coordinates of the vertical feature to be analyzed, the coordinate rotation coefficient and the projection scale coefficient; The method of calculating the coordinate rotation coefficient according to the quaternion data is to use formula (1) to calculate: Among them, (q0, q1, q2, q3) is the quaternion value of the satellite's quaternion data, (a1, a2, a3, b1, b2, b3, c1, c2, c3) is the coordinate rotation coefficient; According to the calculated coordinate rotation coefficients (a1, a2, a3, b1, b2, b3, c1, c2, c3), the bottom image coordinates (x1, y1) of the vertical feature to be analyzed and the focal length f of the satellite CCD camera, the coordinate transformation coefficients (l1, l2, m1, m2, n1, n2) are calculated using formula (2): Wherein, (x1, y1) is the image coordinate of the bottom point of the vertical feature to be analyzed; The height h of the vertical object to be analyzed is calculated using formula (3) according to the image coordinates of the vertical object to be analyzed, the coordinate rotation coefficient and the projection scale coefficient: Among them, H s The height of the ground.

2. The extraction method according to claim 1, characterized in that The method of extracting the bottom image coordinates and the top image coordinates of the vertical object to be analyzed on the satellite CCD remote sensing image is to use image analysis software to analyze the satellite CCD remote sensing image to obtain the bottom image coordinates and the top image coordinates.

3. The extraction method according to claim 1, characterized in that The satellite obtains the three-dimensional position and attitude by reading the auxiliary data file of the satellite CCD.

4. The extraction method according to claim 1, characterized in that The push-sweep line T corresponding to the vertical object to be analyzed can be calculated using formula (4): Among them, x1 is the image coordinate of the bottom point of the vertical feature to be analyzed, and x2 is the image coordinate of the vertex of the vertical feature to be analyzed.

Images