CN104700367A - Geometric correction method of on-board hyperspectral push-broom imaging data - Google Patents

Abstract

The invention provides a geometric correction method of on-board hyperspectral push-broom imaging data. The method comprises the steps of calculating space coordinates of a middle point of a hyperspectral camera scanning point through airship POS data; calculating the space coordinates of all pixels of a distoration image through the airship POS data; building the nearest neighbor correspondence relationship of the pixel 1 from the distoration image to a correction image; correcting the image by the mode of combining the nearest neighbor interpolation and 8 neighboring-region interpolation values. According to the method, the mode of calculating the space coordinates of the pixel of the distoration image is carried out, the correspondence relationship of the pixels of the distoration image and the correction image is built, and the mode of combining the nearest neighbor interpolation and 8 neighboring-region interpolation values is also carried out, so as to quickly correct the image.

Description

A kind of ship carries the geometric correction method of EO-1 hyperion push-broom imaging data

Technical field

The present invention relates to technical field of image processing, relate to Hyperspectral imagery processing technical field special, for a kind of ship carries the geometric correction method of EO-1 hyperion push-broom imaging data.

Background technology

Positioning and orientation system (position and orientation system, POS) utilizes global positioning system inertial measuring unit of unifying directly to determine the integrated technology of sensor space position and attitude.

The eighties in 20th century, along with the develop rapidly of remote sensing technology, new high light spectrum image-forming technology of rising becomes the powerful of the apparent survey of space to ground.High light spectrum image-forming data while provide about the rich space information of topographical surface feature and spectral information for people, also satellite remote sensing date high Precision Processing and advanced treating demand have been risen to one higher level.High light spectrum image-forming technology can obtain the quasi-continuous spectroscopic data of a large amount of narrow wave band, and each pixel has a subcontinuous curve of spectrum.Compared with traditional multispectral romote sensing technology, high-spectrum remote-sensing can provide more abundant atural object observation information, can be used for generating more complicated observation model, improves comprehensive distinguishing and the analysis ability of topographical surface feature situation.

But space remote sensing platform, by the restriction of track, is fixed through the band observation area time, emergent observation cannot be realized; Aerial remote sensing examination and approval procedures are very complicated, affect comparatively large, and it are higher to obtain image data cost by weather condition.Dirigible earth observation platform can provide the multisource synchronization earth observation ability than spaceborne or airborne platform more horn of plenty; The dynamic earth observation ability that Space-borne can be provided to be difficult to realize and meticulous earth observation ability; Airborne platform can be provided to be difficult to the spot hover earth observation ability realized; There is safe, recyclable, easy care, can cruise, replaceable load etc. a little, one of following main remote-sensing flatform will be become.

But dirigible affects comparatively large by air-flow crosswind, self stability is poor.And EO-1 hyperion instrument generally adopts push-broom imaging pattern, in imaging process, the flight attitude of every scan line is all in change, makes high-spectral data produce very large geometric deformation.And the geometric deformation of this geometric deformation and normal image recovers there is larger difference, how to recover to be the key issue that dirigible remote-sensing flatform needs to solve to this image.

Summary of the invention

In order to overcome the shortcoming of above-mentioned prior art, the object of the present invention is to provide a kind of ship to carry the geometric correction method of EO-1 hyperion push-broom imaging data, solving the problem that existing algorithm process ship carries push-broom imaging data geometry correction weak effect.

To achieve these goals, the technical solution used in the present invention is:

Ship carries a geometric correction method for EO-1 hyperion push-broom imaging data, comprises the steps:

Step 1, for a pending view data, read its corresponding ship and carry POS data, within the scope of the longitude and latitude of airship flight route, select wherein as true origin, the latitude and longitude coordinates of ship being carried POS data is converted to ground absolute coordinates, ship is carried the sea level elevation coordinate conversion of POS data for distance floor level coordinate;

Step 2, utilizes ship to carry the volume coordinate of POS data calculating EO-1 hyperion camera scanning line mid point; Preferably, its implementation is as follows:

Calculate the EO-1 hyperion camera scanning line neutral point deviation caused by dirigible pitching, rolling reason respectively, and calculate the volume coordinate of EO-1 hyperion camera scanning line mid point according to dirigible course angle.

Step 3, utilizes ship to carry the volume coordinate of the whole pixel of POS data calculated distortion image; Preferably, its implementation is as follows:

First the volume coordinate of pixel on every bar sweep trace is calculated by sweep trace;

Then calculating pixel on every bar sweep trace relative to the distance of sweep trace mid point obtains the volume coordinate of the whole pixel of distorted image.

Wherein when calculating distance relative to sweep trace mid point of pixel on every bar sweep trace, consider because dirigible is not face the sweep trace affined transformation that imaging causes.

Step 4, sets up the 1 pixel arest neighbors corresponding relation of distorted image to correcting image; Preferably, its implementation is as follows:

The each pixel of distorted image is found to 4 nearest neighbor pixels in correcting image, and set up 4 pairs of corresponding relations;

For each pixel of correcting image, find apart from minimum a pair in the corresponding relation comprising this pixel, and give up other and comprise this pixel corresponding relation.

Step 5, the mode adopting arest neighbors interpolation to combine with 8 neighbor interpolation carrys out correcting image, and preferably, its implementation is as follows:

Arest neighbors interpolation is carried out to the pixel that there is 1 pixel arest neighbors corresponding relation in correcting image;

8 neighbor interpolation are carried out to other pixel iteratives in correcting image, namely in capture element 8 neighborhoods, there is the mean value of the point of pixel value.

Compared with prior art, the present invention, by the volume coordinate utilizing ship to carry POS data calculating EO-1 hyperion camera scanning central point volume coordinate and the whole pixel of calculated distortion image, establishes the corresponding relation from distorted image pixel to correcting image pixel.The mode utilizing arest neighbors interpolation to combine with 8 neighbor interpolation carrys out correcting image, fast and do not recover the geometric deformation of image with not losing precision.

Accompanying drawing explanation

Fig. 1 ship disclosed in the embodiment of the present invention 1 carries the geometric correction method process flow diagram of EO-1 hyperion push-broom imaging data.

Fig. 2 is a line example that ship carries POS data.

Fig. 3 ship disclosed in the embodiment of the present invention 2 carries the geometric correction method process flow diagram of EO-1 hyperion push-broom imaging data.

Fig. 4 is three attitude angle of dirigible POS data record, the schematic diagram of roll angle, the angle of pitch, crab angle.

Fig. 5 is that camera scanning line offsets the affined transformation schematic diagram produced by rolling.

Fig. 6 sets up the 1 pixel arest neighbors corresponding relation schematic diagram of distorted image to correcting image.

Fig. 7 is the schematic diagram of distorted image scan line distribution situation.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.

The invention discloses the geometric correction method that a kind of ship carries EO-1 hyperion push-broom imaging data, carry the problem of push-broom imaging data geometry correction weak effect with the algorithm process ship solving prior art.Its embodiment is as described below:

Embodiment one

Ship disclosed in the present embodiment carries the flow process of the geometric correction method of EO-1 hyperion push-broom imaging data as shown in Figure 1,

Step S11, utilize ship to carry volume coordinate that POS data calculates EO-1 hyperion camera scanning line mid point;

If (x _l, y _l) for being carried the dirigible volume coordinate of the longitude and latitude conversion of POS data record by ship, height is calculated apart from floor level by the height above sea level degree in dirigible POS data, (roll, pitch, yaw) be three attitude angle of dirigible POS data record, roll angle, the angle of pitch, crab angle, as shown in Figure 4.Utilize formula 1.1, calculate the volume coordinate (x of EO-1 hyperion camera scanning line mid point _m, y _m),

offset _roll＝height*tan(-roll)

offset _pitch＝height/cos(roll)*tan(pitch) (0.1)

x _M＝x _L+offset _roll*cos(yaw)+offset _pitch*sin(yaw)

y _M＝y _L-offset _roll*sin(yaw)+offset _pitch*cos(yaw)

In formula, offset _rollbe the sweep trace neutral point deviation caused by roll angle, be referred to as rolling skew; Offset _pitchbe the sweep trace neutral point deviation caused by the angle of pitch, be referred to as pitching skew; Write as plural form namely,

x _M+i*y _M＝x _L+i*y _L+offset _roll*e ^-yaw*i+i*offset _pitch*e ^-yaw*i(0.2)

Can find out, the volume coordinate (x of sweep trace mid point _m, y _m) namely by dirigible coordinate (x _l, y _l), add that rolling skew and pitching skew obtain.

Step S12, ship is utilized to carry the volume coordinate of the whole pixel of POS data calculated distortion image;

The present invention by the volume coordinate of the whole pixel of calculated distortion image of sweep trace, by EO-1 hyperion camera scanning line mid point volume coordinate (x _m, y _m) substitute into following formula, calculate the volume coordinate (x of pixel on distorted image sweep trace _line, y _line),

h＝height/cos(roll)/cos(pitch)

$\mathrm{\α}=a\tan \left(\tan \frac{\mathrm{FOV}}{2}\frac{1-\mathrm{samples}:2:\mathrm{samples}-1}{\mathrm{samples}}\right)$

$\mathrm{line}=\frac{h*\sin \left(\mathrm{\α}\right)}{\sin (a\cos ({\mathrm{offest}}_{\mathrm{roll}}/h)-\mathrm{\α})}---\left(0.3\right)$

x _line＝x _M+line*cos(yaw)

y _line＝y _M-line*sin(yaw)

In formula, FOV is the visual angle of EO-1 hyperion camera, samples is the sampling rate of EO-1 hyperion camera, the physical significance of h is the distance from EO-1 hyperion camera to sweep trace mid point, a is the angle of the point on sweep trace relative to image center line, line is for after have passed through affined transformation, and the point on a sweep trace is relative to the distance of sweep trace central point, and affined transformation as shown in Figure 5.Write as plural form as follows,

x _line+i*y _line＝x _M+i*y _M+line*e ^-yaw*i(0.4)

Further, when imaging faced by dirigible, when namely there is not rolling skew, offset _rollthe computing formula of=0, line is reduced to:

line＝h*tan(a) (0.5)

Step S13, set up the 1 pixel arest neighbors corresponding relation of distorted image to correcting image;

In order to carry out geometry correction to image, the spatial transform relation between distorted image to correcting image need be set up.But for push-broom imaging image, there is not analytical expression in this spatial alternation.The present invention utilizes absolute coordinate space, 1 pixel arest neighbors corresponding relation setting up from distorted image pixel to correcting image pixel.As a distorted image pixel (x _i, y _i) and a correcting image pixel (x' _j, y' _j) between distance be less than 1 pixel and (x _i, y _i) be (x' _j, y' _j) arest neighbors in distorted image, then claim pixel (x _i, y _i) and pixel (x' _j, y' _j) be that a pair 1 pixel arest neighbors are corresponding.

Step S14, demosaicing is carried out to correcting image;

In order to take into account efficiency and degree of accuracy, the mode that the present invention adopts 1 pixel arest neighbors interpolation to combine with 8 neighbor interpolation.For the pixel that there is 1 pixel arest neighbors corresponding relation in correcting image, adopt arest neighbors interpolation; To rest of pixels, adopt 8 neighbor interpolation, namely get in its 8 neighborhood and given the average of the pixel of gray-scale value.

Embodiment two

The flow process of the present embodiment to low-light (level) image processing method is described in detail, its flow process as shown in Figure 3,

Step S31, for a pending view data, read its corresponding ship and carry POS data, ship carries a line example of POS data as shown in Figure 2;

Step S32, within the scope of the longitude and latitude of airship flight route, select wherein (x ₀, y ₀) as true origin;

Step S33, the latitude and longitude coordinates of ship being carried POS data are converted to ground absolute coordinates, adopt following formula,

$\begin{array}{c}{x}_L=(x_{\mathrm{Long}}-x_0)*R*\cos \left(y_0\right)*\frac{\mathrm{\π}}{180}\\ y_L=(y_{\mathrm{Lat}}-y_0)*R*\frac{\mathrm{\π}}{180}\end{array}---\left(0.6\right)$

In formula, (x _long, y _lat) carry the latitude and longitude coordinates of POS data record, (x for ship _l, y _l) ground coordinate for converting to, R represents earth radius.

Step S34, to calculate by dirigible pitching, the EO-1 hyperion camera scanning line neutral point deviation that causes of rolling reason

Step S35, calculate the volume coordinate of EO-1 hyperion camera scanning line mid point according to dirigible course angle;

Step S34-step S35 can utilize formula 1.1 to calculate.

Step S36, each sweep trace for distorted image, perform step S37-step S38, the volume coordinate of the whole pixel of calculated distortion image, Fig. 7 gives a schematic diagram of distorted image scan line distribution;

On step S37, calculating sweep trace, pixel is relative to the angle of camera center line;

On step S38, calculating sweep trace, pixel is relative to the distance of sweep trace mid point, should be taken into account here because dirigible is not the affined transformation facing the sweep trace that imaging causes.

Step S36-step S38 can utilize formula 1.3 to calculate.

Step S39, execution step S310 for each pixel of distorted image, then perform step S311 to each pixel of correcting image, set up the 1 pixel arest neighbors corresponding relation of distorted image to correcting image;

Step S310, a pixel for distorted image, find its 4 nearest neighbor pixels in correcting image, and set up this 4 pairs of corresponding relations, as shown in Figure 6;

Step S311, a pixel for correcting image, find apart from minimum a pair, and give up other corresponding relations comprising this pixel in the corresponding relation comprising this pixel.

Step S312, arest neighbors interpolation is carried out to the pixel that there is 1 pixel arest neighbors corresponding relation in correcting image

Step S313,8 neighbor interpolation are carried out to other pixel iteratives in correcting image, namely in capture element 8 neighborhoods, there is the mean value of the point of pixel value.

In this instructions, each embodiment adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar portion mutually see.For device disclosed in embodiment, because it corresponds to the method disclosed in Example, so description is fairly simple, relevant part illustrates see method part.

Professional can also recognize further, in conjunction with unit and the algorithm steps of each example of embodiment disclosed herein description, can realize with electronic hardware, computer software or the combination of the two, in order to the interchangeability of hardware and software is clearly described, generally describe composition and the step of each example in the above description according to function.These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can use distinct methods to realize described function to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.

The software module that the method described in conjunction with embodiment disclosed herein or the step of algorithm can directly use hardware, processor to perform, or the combination of the two is implemented.Software module can be placed in the storage medium of other form any known in random access memory (RAM), internal memory, ROM (read-only memory) (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field.

To the above-mentioned explanation of the disclosed embodiments, professional and technical personnel in the field are realized or uses the present invention.To be apparent for those skilled in the art to the multiple amendment of these embodiments, General Principle as defined herein can without departing from the spirit or scope of the present invention, realize in other embodiments.Therefore, the present invention can not be restricted to these embodiments shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.

Claims (7)

1. ship carries a geometric correction method for EO-1 hyperion push-broom imaging data, it is characterized in that, comprises the steps:

Step 1, for a pending view data, read its corresponding ship and carry POS data, within the scope of the longitude and latitude of airship flight route, select wherein as true origin, the latitude and longitude coordinates of ship being carried POS data is converted to ground absolute coordinates, ship is carried the sea level elevation coordinate conversion of POS data for distance floor level coordinate;

Step 2, utilizes ship to carry the volume coordinate of POS data calculating EO-1 hyperion camera scanning line mid point;

Step 3, utilizes ship to carry the volume coordinate of the whole pixel of POS data calculated distortion image;

Step 4, sets up the 1 pixel arest neighbors corresponding relation of distorted image to correcting image;

Step 5, the mode adopting arest neighbors interpolation to combine with 8 neighbor interpolation carrys out correcting image.

2. ship carries the geometric correction method of EO-1 hyperion push-broom imaging data according to claim 1, it is characterized in that, in described step 2, calculate the EO-1 hyperion camera scanning line neutral point deviation caused by dirigible pitching, rolling reason respectively, and calculate the volume coordinate of EO-1 hyperion camera scanning line mid point according to dirigible course angle.

3. ship carries the geometric correction method of EO-1 hyperion push-broom imaging data according to claim 1, it is characterized in that, in described step 2, and the volume coordinate (x of EO-1 hyperion camera scanning line mid point _m, y _m) computing formula as follows:

offset _roll＝height*tan(-roll)

offset _pitch＝height/cos(roll)*tan(pitch)

x _M＝x _L+offset _roll*cos(yaw)+offset _pitch*sin(yaw)

y _M＝y _L-offset _roll*sin(yaw)+offset _pitch*cos(yaw)

Wherein, offset _rollbe the sweep trace neutral point deviation caused by roll angle, be referred to as rolling skew; Offset _pitchbe the sweep trace neutral point deviation caused by the angle of pitch, be referred to as pitching skew, (x _l, y _l) for being carried the dirigible volume coordinate of the longitude and latitude conversion of POS data record by ship, height is calculated apart from floor level by the height above sea level degree in dirigible POS data, (roll, pitch, yaw) be three attitude angle of dirigible POS data record, i.e. roll angle, the angle of pitch, crab angle.

4. geometric correction method according to claim 1, is characterized in that, in described step 3:

First the volume coordinate of pixel on every bar sweep trace is calculated by sweep trace;

Then calculating pixel on every bar sweep trace relative to the distance of sweep trace mid point obtains the volume coordinate of the whole pixel of distorted image, when calculating distance relative to sweep trace mid point of pixel on every bar sweep trace, consider because dirigible is not face the sweep trace affined transformation that imaging causes.

5. geometric correction method according to claim 3, is characterized in that, in described step 3: the volume coordinate (x of pixel on distorted image sweep trace _line, y _line) computing formula as follows:

h＝height/cos(roll)/cos(pitch)

$\mathrm{\α}=a\tan \left(\tan \frac{\mathrm{FOV}}{2}\frac{1-\mathrm{samples}:2:\mathrm{samples}-1}{\mathrm{samples}}\right)$

$\mathrm{line}=\frac{h*\sin \left(\mathrm{\α}\right)}{\sin (a\cos ({\mathrm{offset}}_{\mathrm{roll}}/h)-\mathrm{\α})}$

x _line＝x _M+line*cos(yaw)

y _line＝y _M-line*sin(yaw)

Wherein, FOV is the visual angle of EO-1 hyperion camera, samples is the sampling rate of EO-1 hyperion camera, the physical significance of h is the distance from EO-1 hyperion camera to sweep trace mid point, a is the angle of the point on sweep trace relative to image center line, line is for after have passed through affined transformation, and the point on a sweep trace is relative to the distance of sweep trace central point.

6. geometric correction method according to claim 1, is characterized in that, in described step 4:

The each pixel of distorted image is found to 4 nearest neighbor pixels in correcting image, and set up 4 pairs of corresponding relations;

For each pixel of correcting image, find apart from minimum a pair in the corresponding relation comprising this pixel, and give up other and comprise this pixel corresponding relation.

7. geometric correction method according to claim 1, is characterized in that, in described step 5:

Arest neighbors interpolation is carried out to the pixel that there is 1 pixel arest neighbors corresponding relation in correcting image;

8 neighbor interpolation are carried out to other pixel iteratives in correcting image, namely in capture element 8 neighborhoods, there is the mean value of the point of pixel value.