A method for on-orbit field calibration of a high-orbit large-area optical remote sensor

Abstract

This invention proposes an on-orbit field calibration method for a high-orbit, large-area optical remote sensor. Utilizing a uniform ground-based radiometric calibration field, the method calculates relative radiometric calibration coefficients from stitched images using a flat-field approach. The method involves selecting a suitable usable range as the calibration field based on the resolution of the payload to be calibrated; estimating the traversal time based on the usable range, the swath width of the payload, and the satellite's maneuverability; adjusting the satellite's attitude control parameters according to the satellite camera's optical axis pointing and image plane detector distribution characteristics to obtain a set of related images where the calibration field region is distributed across different areas of the camera image; performing geometric correction on the obtained set of related images; extracting the ground-based calibration field region from each image; and stitching them together according to their corresponding positions on the image plane to form a complete calibration image; and finally, using the flat-field approach, the relative radiometric calibration coefficients are obtained through calculations within sub-regions, calculations between sub-regions, and multiplication of sub-regions and sub-regions.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of remote sensing data application, and relates to an on-orbit field calibration method for a high-orbit large-array optical remote sensor. BACKGROUND

[0002] Image data acquired by a spaceborne imaging sensor needs to be radiometrically calibrated to determine the response characteristics of the sensor to a radiation source. In the prior art, the on-orbit calibration method for a high-orbit large-array optical remote sensor adopts a semi-optical-path on-satellite calibration method combined with a full-optical-path local-aperture calibration method, but has the following obvious defects:

[0003] 1. The semi-optical-path on-satellite calibration method does not cover the whole optical path, and does not take into account the influence of atmospheric transmission and environment on the space-ground transmission path, so that the reliability of the calibration result is affected.

[0004] 2. The full-optical-path local-aperture calibration method does not cover the whole field of view, and the reliability of the data extrapolation of the uncovered field of view is affected.

[0005] If the ground calibration data is extrapolated, the accuracy of the ground simulation light source itself will affect the accuracy of the calibration coefficient, and when the satellite enters on-orbit operation, the usability of the ground calibration coefficient will further decrease due to the influence of the space environment and the decay of the device itself. If the semi-optical-path on-satellite calibration data is extrapolated, the confidence error described in defect 1 is introduced, which comprehensively affects the usability of the system calibration coefficient after transmission.

[0006] It can be seen that there is no calibration method for a full-link, full-aperture, full-field-of-view, and full-dynamic-range high-orbit large-array optical remote sensor. The biggest difficulty lies in that the available geometric range of the current ground calibration field is too small to directly meet the calibration requirements of the large-aperture of the high-orbit large-array optical remote system. In addition, the calibration of high-resolution imaging systems requires higher uniformity of the calibration field, and the effective range of the calibration field with good uniformity is further reduced.

[0007] Therefore, it is necessary to improve the technology and method to realize the on-orbit radiometric calibration of a high-resolution large-array optical remote satellite. SUMMARY

[0008] The application provides an on-orbit field calibration method for a high-orbit large-array optical remote sensor, which uses a ground uniform radiometric calibration field to calculate the relative radiometric calibration coefficient for the spliced image by using the flat-field method.

[0009] The application is implemented by the following technical scheme.

[0010] An on-orbit field calibration method for a high-orbit large-array optical remote sensor, comprising the following steps:

[0011] Step one, according to the to-be-calibrated load resolution, select the available range suitable as the calibration field;

[0012] Step two, according to the available range of the calibration field selected in step one and the to-be-calibrated load width, the satellite maneuvering ability, estimate the time required for traversal;

[0013] Step three, within the time required for traversal estimated in step two, according to the satellite camera optical axis pointing and the image plane probe element distribution characteristics, adjust the satellite attitude control parameters, so that the field of view area in the satellite width range traverses the ground uniform calibration field in turn, and a set of related images with the calibration field area distributed in different regions of the camera image are obtained;

[0014] Step four, geometrically correct the set of related images obtained in step three, cut out the ground calibration field area image in each image, and splice the corresponding positions in the image plane to form a whole calibration image;

[0015] Step five, according to the calibration image obtained in step four, use the flat field method to obtain the relative radiation calibration coefficient according to the sub-region calculation, sub-region calculation, and multiplication calculation of sub-regions.

[0016] The beneficial effects of the present application are:

[0017] 1. The present application is aimed at the system of earth stationary orbit satellite staring imaging and the flexible maneuvering attitude control ability, so that the satellite can image multiple times in a short time, traverse the whole field of view area of the ground uniform calibration field in turn, and then obtain a complete satellite calibration image with a large width through image splicing; since it is multiple imaging in a short time, the environment and radiation parameters of the calibration field remain basically constant, so the relative radiation calibration coefficient can be calculated for the spliced image using the flat field method;

[0018] 2. The present application solves the calibration problem of earth stationary orbit high-resolution large-array optical remote sensor due to the field of view range being much larger than the ground uniform field, and fills the technical blank of ground calibration of high-orbit optical remote sensor in the whole chain, full aperture, full field of view, and full dynamic range;

[0019] 3. The present application sets the fixed imaging parameters according to the calibration requirements, which can ensure the stability of the data obtained by the to-be-calibrated instrument during the acquisition of a set of calibration data images, avoid introducing response errors, and ensure the effectiveness of the calibration data;

[0020] 4. The present application considers different width, different resolution, different maneuvering ability, and different size of the ground calibration field available range, etc., and realizes the applicability analysis of the splicing calibration under the premise that the illumination is as constant as possible, allows a certain change in illumination between the start and end of the splicing calibration group image acquisition, and still maintains good calibration accuracy, thereby improving the on-orbit usability of the method;

[0021] 5、The field calibration method provided by the application is the basis for subsequent cross calibration and absolute radiation calibration combined with other high-precision radiation measurement satellite data, and has important value for quantitative monitoring and data application of high-orbit optical remote sensing satellite data. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A flow chart of the on-orbit field calibration method of the high-orbit large-array optical remote sensor in the application;

[0023] Figure 2 A field-of-view splicing schematic diagram for acquiring a full-frame calibration image by splicing in the specific embodiment;

[0024] Figure 3 An image splicing schematic diagram for acquiring a full-frame calibration image by splicing in the specific embodiment. DETAILED DESCRIPTION

[0025] The exemplary embodiments of the application will be described in detail below with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary, and are intended to illustrate the principles and spirits of the application, rather than limit the scope of the application.

[0026] As shown in the drawings, Figure 1 The on-orbit field calibration method of the high-orbit large-array optical remote sensor in the application specifically includes the following steps:

[0027] Step one, according to the resolution of the to-be-calibrated load, select a suitable available range as a calibration field; the specific steps are as follows:

[0028] S101, the to-be-calibrated sensor predicts and acquires a uniform calibration field image;

[0029] S102, select different ranges of the calibration field according to the calculation method of the coefficient of variation CV, divide the to-be-selected area into multiple small windows, calculate each small window separately, then combine the small windows, gradually expand the available range, and until the coefficient of variation does not meet the uniformity requirement, then determine the selected available calibration area size as the available range of the calibration field.

[0030] In this embodiment, the CV is the standard deviation of the sample divided by the average value, expressed in percentage, and the calculation formula is as follows:

[0031]

[0032]

[0033] Wherein, σ is the standard deviation of the sample, E is the average value of the sample, CV i is the coefficient of variation in a window, CV aveThe mean value of all CVs in the region of interest.

[0034] Taking the observation data of Dunhuang calibration site in China and Landsat8-Operational Land Imager (L8-OLI), Sentinel2-MSI (S2-MSI) and GF2-PMS sensors (spatial resolutions are 30m, 10m, 3.2m respectively) as examples, the statistical analysis of the variation of the site radiation spatial difference coefficient with the pixel resolution is carried out.

[0035]

[0036]

[0037] This step is due to the CV value presents a larger trend with the improvement of the resolution, but the CV values of different bands with the same resolution are different, so the range of the site suitable for calibration needs to be selected according to the resolution and spectral band of the camera to be tested; the smaller the CV value, the better the calibration accuracy requirement, but the smaller the CV value of the same band and the same resolution, the more limited the corresponding available calibration site range, and then the time of step two is lost, so steps one and two need to be weighed.

[0038] Step two, according to the available range of the calibration site selected in step one and the width of the load to be calibrated, the satellite maneuvering ability, the time required for traversal is estimated; in specific implementation, whether to use the mosaic calibration method is judged according to the influence degree of the change of illumination conditions in the traversal time range on the calibration accuracy;

[0039] This step is because the relative radiation calibration coefficient error is proportional to the non-uniformity of the light source, which is a 1:1 relationship. In this embodiment, considering the size limit of the ground uniform field, it is assumed that the time-sharing mosaic imaging mode is used to realize the observation of the calibration site by the satellite with 100km full field of view.

[0040] Assuming that the orbit height of the geostationary satellite is 36000km, the satellite observation pointing angle needs to be changed Δθ≈sin(100 / 36000)≈0.16°. Taking Dunhuang 20km×20km and relatively more uniform 10km×10km calibration site as examples, 5*5 times and 10*10 times of observation are needed for mosaic, and assuming that the corresponding mosaic time is about 5min and 20min respectively, the calibration error caused by 5min time difference and 20min time difference in each band (empirical value) is about:

[0041]

[0042] It can be seen that the longer the time required for traversal, the greater the influence of the change of sunlight conditions on the start and end of the acquisition of the image of the data taken, the smaller the precision error is less than the calibration precision requirement, and the faster the completion time of the traversal path is required.

[0043] Step three, according to the calibration precision requirement and the time required for traversal obtained in step two, the imaging parameters of the remote sensing camera carried by the high-orbit satellite are set, a certain spectral band, integration time and gain circuit parameters are selected, so that the camera keeps the setting parameters unchanged in the subsequent multi-imaging data acquisition process of the calibration of the stitching;

[0044] Step four, within the time required for traversal estimated in step two, according to the satellite camera optical axis pointing and the distribution characteristics of the image element, the satellite attitude control parameters are adjusted, so that the field of view area in the satellite width range traverses the ground uniform calibration field in turn, and a group of related images of the calibration field area distributed in different areas of the camera image are obtained.

[0045] As shown in Figure 2 , first, the satellite field of view is divided into N sub-zones matched with the size of the uniform field; second, the satellite pointing angle attitude is controlled, and at least N times of shooting are performed in succession, so that the uniform field is imaged in different sub-zones of the satellite field of view. In specific implementation, the whole satellite stepping momentum needs to ensure that the data of the available calibration field area are not missed.

[0046] Step five, the group of related images obtained in step four are geometrically corrected, the ground calibration field area image in each image is cut out, and a whole calibration image is formed by splicing according to the corresponding positions on the image plane.

[0047] As shown in Figure 3 , in this embodiment, the obtained image data is geometrically corrected, the ground calibration field area image in each image is cut out, and a whole calibration image is formed by splicing the cut-out sub-zone calibration field image according to its corresponding position on the image plane. Figure 3 The multiple imaging images of the uniform field are sorted in order according to their positions in the satellite field of view, and the full-field-of-view uniform field observation image is reconstructed by splicing.

[0048] Step six, according to the calibration image obtained in step five, the relative radiation calibration coefficient is obtained by using the flat field method according to the sub-zone calculation, sub-zone calculation, and multiplication calculation of the sub-zone and the sub-zone. The specific steps are as follows:

[0049] S601, calculate the average value DN of the signal value of the Kth sub-zone of the calibration image K ; the specific formula is as follows:

[0050]

[0051] DN r，K，j DN K，j DN

[0052] S602、According to the average value DN K , the relative radiation calibration coefficient NG K，j of the jth pixel in the Kth sub-region is calculated; the specific formula is as follows:

[0053]

[0054] S603、According to the average value DN K , the average value of the signal value of each sub-region in the full field of view is calculated ; the specific formula is as follows:

[0055]

[0056] S604、According to the average value of the signal value of each sub-region in the full field of view , the sub-region relative radiation calibration coefficient NG K of the Kth sub-region is calculated; the specific formula is as follows:

[0057]

[0058] S605、According to the sub-region relative radiation calibration coefficient NG K , the total relative radiation calibration coefficient of the ith pixel in the full field of view is calculated; the specific formula is as follows:

[0059] NG i = NG K，j · NG K .

[0060] Those skilled in the art will realize that the embodiments described herein are for the purpose of helping the reader understand the principles of the present application, and should be understood as not limiting the scope of protection of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspiration disclosed in the present application without departing from the essence of the present application, and these modifications and combinations are still within the scope of protection of the present application.

Claims

1. An on-orbit ground calibration method for a high-orbit large-array optical remote sensor, characterized in that, The method comprises the following steps: Step one: according to the resolution of the to-be-calibrated load, select a suitable available range as the calibration field; the specific steps are as follows: S101: the to-be-calibrated sensor predicts and acquires a uniform calibration field image; S102: according to the calculation method of the coefficient of variation CV of different ranges of the calibration field, the to-be-selected region is divided into a plurality of small windows, the signal value of each small window is calculated separately, and then the small windows are combined to gradually expand the available range, until the coefficient of variation does not meet the uniformity requirement, then the size of the selected available calibration region is determined as the available range of the calibration field; Step two: according to the available range of the calibration field selected in step one and the width of the to-be-calibrated load and the satellite maneuvering capability, estimate the time required for traversal; Step three: according to the calibration accuracy requirement and the time required for traversal obtained in step two, set the imaging parameters of the remote sensing camera carried by the high-orbit satellite, select a certain spectral band, integration time and gain circuit parameters, so that the camera keeps the setting parameters unchanged in the subsequent multi-imaging data acquisition process of the calibration of the stitching; Step four: within the time required for traversal estimated in step two, according to the optical axis pointing direction of the satellite camera and the distribution characteristics of the image elements on the image plane, adjust the satellite attitude control parameters, so that the field of view region in the satellite width range traverses the ground uniform calibration field in turn, and a group of related images with the calibration field region distributed in different regions of the camera image are obtained; Step five: geometrically correct the group of related images obtained in step four, cut out the ground calibration field region image in each image, and splice the images in the corresponding positions on the image plane to form a whole calibration image; Step six: according to the calibration image obtained in step five, use the flat field method to obtain the relative radiation calibration coefficient according to the calculation in the sub-region, the calculation between the sub-regions, and the multiplication calculation of the sub-region and the sub-region; the specific steps are as follows: S601: calculate the average value of the signal value of the Kth sub-region of the calibration image; S602: calculate the relative radiation calibration coefficient of the jth image element in the Kth sub-region according to the average value; S603: calculate the average value of the signal value of each sub-region of the full field of view according to the average value; S604: calculate the sub-region relative radiation calibration coefficient of the Kth sub-region according to the average value of the signal value of each sub-region of the full field of view; S605: calculate the total relative radiation calibration coefficient of the ith image element of the full field of view according to the sub-region relative radiation calibration coefficient.

2. The on-orbit field calibration method for high-orbit large-array optical remote sensor according to claim 1, characterized in that, In step five, the satellite field of view is first divided into N sub-regions matched with the size of the uniform field; secondly, the satellite pointing angle attitude is controlled to quickly and continuously take at least N pictures, so that the uniform field is imaged in different sub-regions of the satellite field of view.

Images