All-day star sensor system and detection method based on dual-band image fusion

Abstract

The present application belongs to the field of optical detection technology, and particularly relates to an all-weather star sensor system and a detection method based on dual-band image fusion. The system is as follows: a target light beam is transmitted to a visible light detector through a visible light lens group, and at the same time, the target light beam is also transmitted to a short-wave infrared detector through a short-wave infrared lens group; a synchronous trigger controls the visible light detector and the short-wave infrared detector to synchronously expose at the current time, so that the visible light image and the short-wave infrared image are spatio-temporally aligned; and the visible light image and the short-wave infrared image are synchronously transmitted to a dual-band image data fusion processing module at the current time for processing, and high-precision attitude information of the aircraft at the current time is output. The present application adopts a dual-cylinder dual-band star sensor optical structure, and the whole system has no modulation structure and is high in stability, and is more suitable for strong vibration platforms such as ships and aircrafts.

Description

Technical Field

[0001] This invention belongs to the field of optical detection technology, and particularly relates to an all-weather star-sensor system and detection method based on dual-band image fusion. Background Technology

[0002] A star sensor is an attitude measurement and navigation device based on optical methods. Its working principle is to identify and track stellar targets in the sky, using the stars as fixed reference points to determine the azimuth and attitude angles of the star sensor itself and its carrier platform.

[0003] Compared to other navigation systems such as GPS and inertial navigation, star sensors offer advantages such as high measurement accuracy, no base station dependence, high long-term stability, and strong resistance to electromagnetic interference, leading to their widespread application in aerospace navigation and attitude determination for satellites and spacecraft. However, in daytime applications in near-Earth space, star sensor technology is highly susceptible to strong solar stray light. The solar radiation intensity near Earth's orbit can reach approximately 1361 watts per square meter (solar constant). Such high background radiation causes the intensity of background stray light received by the star sensor system to far exceed the signal from the target star, resulting in the star signal being submerged in background noise and difficult to detect and identify effectively. Therefore, at present, star sensor technology is only used as a supplementary navigation method for airborne, shipborne, and vehicle-mounted near-Earth platforms at night, and cannot meet the requirements for all-day applications.

[0004] Star sensors acquire digital images of stellar targets using their imaging optical systems and detection terminals. Through matching and calculation with existing star database data, they achieve positioning and attitude determination functions for themselves and their carrier platforms. With their advantages of high accuracy, strong anti-interference capabilities, and the ability to achieve autonomous attitude and orientation without relying on other equipment, they are widely used in satellites, spacecraft, and other platforms. However, due to the strong background stray light in near-Earth space during the day, star sensor devices on ships, aircraft, and other platforms can only operate at night, failing to meet the requirements for all-day applications. Therefore, it is essential to invent an all-day star sensor capable of detecting faint stellar targets against a strong background. Summary of the Invention

[0005] In view of this, the present invention aims to provide an all-weather star-sensor system and detection method based on dual-band image fusion, to address the shortcomings of existing technologies that typically employ background suppression methods such as small-field-of-view scanning mirrors, polarization modulation, and field-of-view gating to reduce strong solar stray light interference during the day, thereby improving the daytime detection capability of star sensors. However, the above solutions mostly rely on non-fixed modulation mechanisms, which not only increases the structural complexity of the star sensor system and reduces the overall operational stability, but also inevitably causes energy loss of the stellar target during the suppression of background stray light, resulting in limited improvement in daytime detection performance and failing to meet the application requirements of ships, airborne and other moving platforms for all-weather, high-reliability star sensor operation. The present invention adopts a dual-tube dual-band star sensor optical structure, with no modulation structure in the entire system, resulting in high stability and making it more suitable for ships, airborne and other high-vibration platforms.

[0006] To achieve the above objectives, the technical solution created by this invention is implemented as follows: An all-weather star-sensing system based on dual-band image fusion, mounted on a spacecraft, includes: a visible light lens group, a short-wave infrared lens group, a visible light detector, a short-wave infrared detector, a synchronization trigger, and a dual-band image data fusion processing module, wherein: The target beam is transmitted to a visible light detector via a visible light lens group. The visible light detector is located at the defocus position of the visible light lens group and is used to acquire a visible light image containing a broadband skylight background signal. Simultaneously, the target beam is also transmitted to a short-wave infrared detector via a short-wave infrared lens group. The short-wave infrared detector is located at the focus position of the short-wave infrared lens group and is used to acquire a short-wave infrared image containing the target signal and skylight background noise. The visible light detector and the short-wave infrared detector have the same pixel array. A synchronization trigger controls the visible light detector and the short-wave infrared detector to expose synchronously at the current moment, so that the visible light image and the short-wave infrared image are spatiotemporally aligned. At the current moment, the visible light image and the short-wave infrared image are synchronously transmitted to the dual-band image data fusion processing module for processing, and outputs the high-precision attitude information of the aircraft at the current moment.

[0007] Furthermore, the dual-band image data fusion processing module includes a calibration data storage unit, an image registration unit, a background prediction unit, a phase-locked elimination unit, and an attitude calculation unit, wherein: The calibration data storage unit is used to construct a calibration image pair dataset based on visible light images and corresponding shortwave infrared images acquired under different light intensities, and to calibrate the spectral mapping function based on the calibration image pair dataset; The image registration unit is used to register the visible light image acquired at the current moment to the image coordinate system of the shortwave infrared image, so as to obtain the registered visible light image; The background prediction unit is used to input the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current time to obtain the residual image, and to extract the target signal from the residual image. The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

[0008] Furthermore, the specific operations of the calibration data storage unit include: A broadband halogen tungsten lamp light source is used to simultaneously and uniformly illuminate the target surface of the visible light detector and the target surface of the short-wave infrared detector. Under different light intensities, visible light images and corresponding short-wave infrared images of the same scene were collected to construct a calibration image pair dataset; A linear mapping model is constructed by sequentially inputting each calibrated image pair from the calibrated image pair dataset into the linear mapping model to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibrated image pair satisfies the following: S_swir(x1, y1) = α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ); Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of the shortwave infrared detector, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of the visible light detector, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

[0009] Furthermore, the incident end of the visible light lens group is equipped with a visible light shielding tube, and the incident end of the short-wave infrared lens group is equipped with a short-wave infrared shielding tube.

[0010] A sky-based, time-sensitive star-sensing method based on dual-band image fusion is implemented using a sky-based, time-sensitive star-sensing system based on dual-band image fusion, and specifically includes the following steps: S1: The all-day star-sensing system is mounted on the spacecraft. The target beam is transmitted to the visible light detector through the visible light lens group, and at the same time, the target beam is also transmitted to the short-wave infrared detector through the short-wave infrared lens group. S2: The synchronization trigger controls the visible light detector and the short-wave infrared detector to expose synchronously at the current moment, so that the visible light image and the short-wave infrared image are spatiotemporally aligned, and transmits the visible light image and the short-wave infrared image synchronously to the dual-band image data fusion processing module for processing at the current moment, and outputs the high-precision attitude information of the aircraft at the current moment. S3: Determine if the task has ended. If so, end the star-sensor detection. Otherwise, replace the current time with the next time and repeat step S2 until the star-sensor detection ends.

[0011] Furthermore, step S2 specifically includes: S21: The calibration data storage unit constructs a calibration image pair dataset based on visible light images and corresponding short-wave infrared images collected under different light intensities, and obtains the spectrum mapping function based on the calibration image pair dataset; S22: The image registration unit registers the visible light image acquired at the current moment to the image coordinate system where the shortwave infrared image is located, and obtains the registered visible light image; S23: The background prediction unit inputs the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; S24: The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current moment to obtain the residual image and extract the target signal in the residual image. S25: The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

[0012] Furthermore, step S21 specifically includes: S211: Use a broadband halogen tungsten lamp light source to simultaneously and uniformly illuminate the target surface of the visible light detector and the target surface of the short-wave infrared detector. S212: Under different light intensities, collect visible light images and corresponding short-wave infrared images of the same scene to construct a calibration image pair dataset; S213: Construct a linear mapping model. Input each calibration image pair contained in the calibration image pair dataset into the linear mapping model in sequence to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibration image pair satisfies: S_swir(x1, y1)=α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ); Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of the shortwave infrared detector, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of the visible light detector, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); S214: Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

[0013] Compared with the prior art, the present invention can achieve the following beneficial effects: (1) The present invention creates an all-day star-sensor system and detection method based on dual-band image fusion. The system uses physically separated binocular optical channels to acquire sky images in the visible light band and short-wave infrared band respectively. The visible light detector is deliberately placed in a defocused position so that its image mainly reflects the spatial distribution of the broad spectrum sky light background as a "reference signal". The short-wave infrared detector is precisely focused to acquire the "test signal" containing the target signal and sky light background noise. The spatiotemporal alignment of the dual-band images is ensured by synchronous triggering. The background noise in the short-wave infrared image is modeled and precisely removed in real time using the visible light image, thereby extracting the star signal submerged in the noise and achieving a signal-to-noise ratio improvement effect similar to lock-in amplification.

[0014] (2) The present invention creates an all-day star-sensor system and detection method based on dual-band image fusion. The visible light detector is fixedly installed on the defocusing plane of the visible light lens group. It aims to actively blur point targets such as stars, so that the sky background component in the visible light image is absolutely dominant, thereby minimizing the interference of the all-day star-sensor target signal of stars on the subsequent background noise modeling and ensuring that the obtained "reference signal" is pure.

[0015] (3) The all-day star-sensor system and detection method based on dual-band image fusion described in this invention, before the all-day star-sensor system is put into operation, a series of dual-band calibration images under different intensities are collected by uniformly illuminating the dual detectors with a broadband light source (such as a halogen tungsten lamp); based on these image data, the spectrum mapping coefficient moment matrix and offset matrix from the visible light background gray value to the short-wave infrared background gray value are independently fitted for each pixel, thereby establishing a quantitative conversion relationship (spectral mapping function) between the two bands of background radiation.

[0016] (4) The all-day star-sensor system and detection method based on dual-band image fusion described in this invention, when working online, substitutes the real-time acquired visible light image into the spectrum mapping function to predict and generate the background noise image that should exist in the short-wave infrared image at the current moment; then, performs pixel-by-pixel difference operation on the actual acquired short-wave infrared image and the predicted background noise image to directly cancel the common-mode background noise and obtain a residual image dominated by star signals.

[0017] (5) The all-day star-sensor system and detection method based on dual-band image fusion described in this invention includes a synchronization trigger for controlling the visible light detector and the short-wave infrared detector to start and end exposure at the same moment, ensuring that the two images reflect the sky state at the same instant; at the same time, the system performs real-time spatial registration of the dual-band images through pre-calibrated geometric transformation parameters, ensuring that the same sky area is aligned at the pixel level, providing a prerequisite for accurate pixel-level background subtraction.

[0018] (6) The all-day star-sensor system and detection method based on dual-band image fusion described in this invention include a physically independent and parallel visible light lens group and a short-wave infrared lens group, each equipped with a dedicated light shield. This design avoids light energy loss and stray light interference caused by common optical path splitting, and can independently optimize optical design and stray light suppression for different bands.

[0019] (7) The all-day star-sensing system and detection method based on dual-band image fusion described in this invention mainly includes a calibration data storage unit (used to store the spectrum mapping coefficient matrix and offset matrix), an image registration unit, a background prediction unit (calculates and predicts the background based on the spectrum mapping function and the real-time visible light image), and a phase-locked elimination unit, which together complete the real-time processing flow from dual-band image input to target star point image output.

[0020] (8) The all-weather star-sensor system and detection method based on dual-band image fusion described in this invention adopts a dual-tube dual-band star sensor optical structure. The all-weather star-sensor system has no modulation structure, has high stability, and is more suitable for strong vibration platforms such as ships and airborne platforms. Furthermore, it utilizes the spectral mapping function of the visible light band and the short-wave infrared band to achieve real-time removal of background noise from strong daytime skylight, enabling the short-wave infrared detector terminal to improve the detection signal-to-noise ratio without reducing the target signal strength. Attached Figure Description

[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 A schematic diagram of the structure of the all-day star-sensor system based on dual-band image fusion as described in the embodiments of the present invention; Figure 2 A schematic diagram of the structure used to construct the spectrum mapping function as described in the embodiments of the present invention; Figure 3 This is a flowchart illustrating the all-day star-sensitive detection method based on dual-band image fusion, as described in an embodiment of the present invention.

[0022] Explanation of reference numerals in the attached figures: 1. Visible light lens group; 2. Short-wave infrared lens group; 3. Visible light detector; 4. Short-wave infrared detector; 5. Synchronization trigger; 6. Dual-band image data fusion processing module; 7. Visible light shield; 8. Short-wave infrared light shield; 9. Broadband halogen tungsten lamp light source. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0024] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0025] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0026] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0027] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0028] like Figure 1 As shown, this invention proposes an all-weather star-sensor system based on dual-band image fusion. The all-weather star-sensor system is mounted on a spacecraft and includes: a visible light lens group 1, a short-wave infrared lens group 2, a visible light detector 3, a short-wave infrared detector 4, a synchronization trigger 5, and a dual-band image data fusion processing module 6, wherein: The target beam is transmitted to the visible light detector 3 via the visible light lens group 1. The visible light detector 3 is located at the defocus position of the visible light lens group 1 and is used to acquire a visible light image containing a broadband sky light background signal. At the same time, the target beam is also transmitted to the short-wave infrared detector 4 via the short-wave infrared lens group 2. The short-wave infrared detector 4 is located at the focus position of the short-wave infrared lens group 2 and is used to acquire a short-wave infrared image containing the target signal and sky light background noise. The visible light detector 3 and the short-wave infrared detector 4 have the same pixel array. The synchronization trigger 5 controls the visible light detector 3 and the short-wave infrared detector 4 to be exposed synchronously at the current moment, so that the visible light image and the short-wave infrared image are spatiotemporally aligned. At the current moment, the visible light image and the short-wave infrared image are synchronously transmitted to the dual-band image data fusion processing module 6 for processing, and outputs the high-precision attitude information of the aircraft at the current moment.

[0029] It should be noted that the all-weather star-sensing system proposed in this invention is mounted on an aircraft platform. The system employs a dual-path parallel detection architecture using visible light and short-wave infrared light. A synchronization trigger 5 is used to achieve synchronous exposure of the two detectors, ensuring precise spatiotemporal alignment of the dual-band images. By setting a differentiated layout—with the visible light detector 3 performing defocused acquisition and the short-wave infrared detector 4 performing focused imaging—effective information on both the sky background and star targets is acquired. The dual-band image fusion processing module performs joint data processing, effectively suppressing interference from complex sky backgrounds, improving star extraction accuracy, and ultimately outputting high-precision attitude data for the aircraft, enabling stable and reliable all-weather star attitude measurement.

[0030] In some embodiments, the dual-band image data fusion processing module 6 includes a calibration data storage unit, an image registration unit, a background prediction unit, a phase-locked elimination unit, and an attitude calculation unit, wherein: The calibration data storage unit is used to construct a calibration image pair dataset based on visible light images and corresponding shortwave infrared images acquired under different light intensities, and to calibrate the spectral mapping function based on the calibration image pair dataset; The image registration unit is used to register the visible light image acquired at the current moment to the image coordinate system of the shortwave infrared image, so as to obtain the registered visible light image; The SIFT (Scale-invariant feature transform) image registration algorithm is used to register the visible light image acquired at the current moment to the image coordinate system of the shortwave infrared image, thus obtaining the registered visible light image.

[0031] The background prediction unit is used to input the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current time to obtain the residual image, and to extract the target signal from the residual image. The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

[0032] In some embodiments, the specific operation of the calibration data storage unit includes: The target surfaces of the visible light detector 3 and the short-wave infrared detector 4 are simultaneously and uniformly illuminated by a broadband halogen tungsten lamp light source 9. Under different light intensities, visible light images and corresponding short-wave infrared images of the same scene were collected to construct a calibration image pair dataset; A linear mapping model is constructed by sequentially inputting each calibrated image pair from the calibrated image pair dataset into the linear mapping model to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibrated image pair satisfies the following: S_swir(x1, y1) = α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ); Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of shortwave infrared detector 4, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of visible light detector 3, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

[0033] In some embodiments, the incident end of the visible light lens group 1 is equipped with a visible light shielding tube 7, and the incident end of the short-wave infrared lens group 2 is equipped with a short-wave infrared shielding tube 8.

[0034] In terms of hardware, the present invention mainly comprises three parts: a dual-barrel dual-band star sensor optical system, a back-end dual-band image fusion detector, and a dual-band image data fusion processing module 6.

[0035] (1) Binocular dual-band star sensor optical system The dual-barrel dual-band star sensor optical system mainly consists of a visible light lens group and a short-wave infrared lens group. Stray light is suppressed by corresponding visible light shielding tubes 7 and short-wave infrared shielding tubes 8, thereby reducing the impact of stray light on the downstream dual-band image fusion detector. The dual-barrel dual-band star sensor optical system is primarily responsible for the acquisition and transmission of target signals and the suppression of stray light outside the field of view.

[0036] (2) Back-end dual-band image fusion detector The back-end dual-band image fusion detector mainly consists of a visible light detector 3, a short-wave infrared detector 4, and a synchronization trigger 5. After the light beam passes through the dual-bullet dual-band star sensor optical system, the visible light detector 3 acquires the sky background signal in the 500nm~850nm band, while the short-wave infrared detector 4 is responsible for acquiring the target signal and sky background signal in the 1000~1700nm band. The two detectors are controlled by the synchronization trigger 5 to ensure synchronous exposure and simultaneously transmit the data to the dual-band image data fusion processing module 6. The visible light detector 3 is located at the defocus position of the visible light lens group 1, thereby reducing the influence of the target signal on the sky background detection.

[0037] (3) Dual-band image data fusion processing module 6 The dual-band image data fusion processing module 6 is the core of the system's computation and control. It receives synchronized and registered visible light image Ivis(x, y) and short-wave infrared image Iswir(x, y) from the back-end dual-band image fusion detector. Its data processing flow is mainly divided into two parts: spectrum mapping calibration and real-time background removal. The core idea is based on the spectrum mapping relationship of skylight background noise in the visible light and short-wave infrared bands, using the defocused visible light image as a "reference signal" to model and remove background noise in the short-wave infrared image. This process is similar to a spatial domain image lock-in amplification. The calibration system used is as follows... Figure 2 As shown.

[0038] The specific process is as follows: 1) Spectrum mapping calibration: Modeling spectrum mapping relationships This operation is performed in the laboratory before the all-day star-sensor system is put into use or periodically. Its purpose is to establish a mapping model of the sky background radiation intensity from the visible light band to the shortwave infrared band, i.e., the spectral mapping function F. The main process is as follows: a) Calibration light source: A broadband halogen tungsten lamp light source 9 is used to uniformly illuminate the target surfaces of the two detectors. The spectrum of this light source covers 400-2500nm and can simultaneously excite the responses of the visible light detector 3 and the short-wave infrared detector 4 to simulate the background radiation of the sky.

[0039] b) Dual-band image acquisition: Under various lighting conditions of different intensities (achieved by adjusting the light source power or neutral density filter), a series of defocused visible light images C_vis_i and corresponding short-wave infrared images C_swir_i (i=1,2,...N) for calibration are simultaneously acquired. Since the visible light images are in a defocused state, they do not contain point targets, but only uniform or known-distributed gray levels.

[0040] c) Mapping function fitting: For corresponding pixels in the visible light image and the corresponding short-wave infrared image, which are pixels with the same spatial position on the target surfaces of visible light detector 3 and short-wave infrared detector 4, curve fitting is performed on all collected calibration image data pairs (C_vis_i(x, y), C_swir_i(x, y)). Typically, since background radiation has a high linear correlation between the two bands, a linear mapping model can be used for fitting. S_swir(x1, y1) = α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ); Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of shortwave infrared detector 4, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of visible light detector 3, x1=x2, y1=y2, α(x 2→1 , y 2→1 β(x) represents the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1), reflecting the conversion ratio of the background radiation of the pixel at that spatial location from visible light to short-wave infrared. 2→1 , y 2→1 The offset is the distance between pixel (x2, y2) and pixel (x1, y1) when linearly mapping them, mainly reflecting the differences in dark current and bias between the two detectors. Finally, two scaling matrices are obtained: the spectrum mapping coefficient matrix A and the offset matrix B, which together constitute the spectrum mapping function F.

[0041] 2) Background noise removal based on lock-in amplification This step describes the specific process for real-time processing of each frame of synchronously acquired dual-band image when the system is in orbit or operating during the day.

[0042] a) Image Preprocessing and Registration: First, inherent non-uniformity correction is performed on the input visible light image I_vis and shortwave infrared image I_swir. Next, geometric registration is performed. Due to the slight field-of-view and pointing differences in the binocular dual-band star sensor optical system, a pre-calibrated transformation matrix (obtained using the SIFT (Scale-invariant feature transform) image registration algorithm, a common existing image transformation method) is used to accurately register the visible light image to the shortwave infrared image coordinate system, ensuring spatial consistency. The registered visible light image is denoted as I_vis_reg(x, y).

[0043] b) Modeling of skylight background noise: ① The real-time acquired, registered, defocused visible light image I_vis_reg is used as the "direct measurement value" of the sky background in the visible light band at the current moment. Due to defocus, point targets such as stars have been blurred into a nearly uniform background. Therefore, I_vis_reg mainly contains sky background information B_vis and a small amount of blurred target residue T_vis_residual.

[0044] ② Using the spectral mapping function F obtained from offline calibration, predict the image B_swir_pred corresponding to the skylight background noise in the shortwave infrared band: B_swir_pred(x, y) = A(x, y) × I_vis_reg(x, y) + B(x,y). This step is the core of "phase-locked loop". I_vis_reg serves as a reference signal, and using the known spectral relationship F, predicts the in-phase background noise in the shortwave infrared channel.

[0045] c) Background Removal and Signal Extraction (Lock-in Amplification): From the actual acquired shortwave infrared image I_swir(x,y), the predicted background noise image B_swir_pred(x,y) is subtracted to obtain the residual image ΔI(x,y): ΔI(x,y) = I_swir(x,y) - B_swir_pred(x,y). Ideally, most of the relevant background noise in ΔI(x,y) is canceled out, while the target signal (mainly present in the shortwave infrared image and contributing very little in the defocused visible light image) is preserved and significantly enhanced as the "signal to be measured". This "background reduction" process is equivalent in the image spatial domain to "phase-sensitive detection" in lock-in amplification technology, demodulating the target signal submerged in strong background noise.

[0046] d) Stellar Target Detection and Attitude Calculation: Thresholding segmentation, star centroid extraction, and magnitude calculation are performed on the residual image ΔI(x, y). Due to significant background suppression, the signal-to-noise ratio is greatly improved, enabling reliable star detection even in strong daylight. The detected stars are matched and identified with the navigation star catalog to ultimately calculate the spacecraft's high-precision attitude information at the current moment.

[0047] like Figure 3 As shown, an all-day star-sensing detection method based on dual-band image fusion is implemented using an all-day star-sensing system based on dual-band image fusion, and specifically includes the following steps: S1: The all-day star-sensing system is mounted on the aircraft. The target beam is transmitted to the visible light detector 3 through the visible light lens group 1, and at the same time, the target beam is also transmitted to the short-wave infrared detector 4 through the short-wave infrared lens group 2. S2: Synchronization trigger 5 controls the visible light detector 3 and shortwave infrared detector 4 to expose synchronously at the current moment, so that the visible light image and the shortwave infrared image are spatiotemporally aligned, and transmits the visible light image and the shortwave infrared image synchronously to the dual-band image data fusion processing module 6 for processing at the current moment, and outputs the high-precision attitude information of the aircraft at the current moment. S3: Determine if the task has ended. If so, end the star-sensor detection. Otherwise, replace the current time with the next time and repeat step S2 until the star-sensor detection ends.

[0048] It should be noted that this invention utilizes detectors in both the visible and short-wave infrared bands to achieve dual-band image data fusion for low signal-to-noise ratio stellar targets against a strong sky background. A spectral mapping function is established based on the high- and low-frequency differences in sky background noise and target signal between the two band images, effectively suppressing strong background stray light interference and significantly improving the signal-to-noise ratio of stellar targets. A small-aperture star sensor device enables daytime detection of faint stellar targets, thus meeting the all-weather application requirements of star sensing. Compared to techniques such as field-of-view gating and polarization, this invention utilizes image data from both the visible and short-wave infrared bands, achieving spectral modeling and removal of sky background noise using dual-band fused images without losing the target signal. This is expected to significantly improve the daytime detection capability of all-weather star sensing systems.

[0049] In some embodiments, step S2 specifically includes: S21: The calibration data storage unit constructs a calibration image pair dataset based on visible light images and corresponding short-wave infrared images collected under different light intensities, and obtains the spectrum mapping function based on the calibration image pair dataset; S22: The image registration unit registers the visible light image acquired at the current moment to the image coordinate system where the shortwave infrared image is located, and obtains the registered visible light image; S23: The background prediction unit inputs the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; S24: The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current moment to obtain the residual image and extract the target signal in the residual image. S25: The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

[0050] In some embodiments, step S21 specifically includes: S211: Use a broadband halogen tungsten lamp light source 9 to simultaneously and uniformly illuminate the target surface of the visible light detector 3 and the target surface of the short-wave infrared detector 4. S212: Under different light intensities, collect visible light images and corresponding short-wave infrared images of the same scene to construct a calibration image pair dataset; S213: Construct a linear mapping model. Input each calibration image pair contained in the calibration image pair dataset into the linear mapping model in sequence to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibration image pair satisfies: S_swir(x1, y1)=α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ); Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of shortwave infrared detector 4, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of visible light detector 3, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); S214: Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

[0051] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0052] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A sky-based, time-and-space star-sensing system based on dual-band image fusion, wherein the sky-based, time-and-space star-sensing system is mounted on an aircraft, characterized in that: include: The system comprises a visible light lens group, a short-wave infrared lens group, a visible light detector, a short-wave infrared detector, a synchronization trigger, and a dual-band image data fusion processing module, wherein: The target beam is transmitted to a visible light detector via a visible light lens group. The visible light detector is located at the defocus position of the visible light lens group and is used to acquire a visible light image containing a broadband sky background signal. At the same time, the target beam is also transmitted to a short-wave infrared detector via a short-wave infrared lens group. The short-wave infrared detector is located at the focus position of the short-wave infrared lens group and is used to acquire a short-wave infrared image containing the target signal and sky background noise. The visible light detector and the short-wave infrared detector have the same pixel array. A synchronization trigger controls the visible light detector and the short-wave infrared detector to expose synchronously at the current moment, so that the visible light image and the short-wave infrared image are spatiotemporally aligned. At the current moment, the visible light image and the short-wave infrared image are synchronously transmitted to the dual-band image data fusion processing module for processing, and outputs the high-precision attitude information of the aircraft at the current moment. The dual-band image data fusion processing module includes a calibration data storage unit, an image registration unit, a background prediction unit, a phase-locked elimination unit, and an attitude calculation unit, wherein: The calibration data storage unit is used to construct a calibration image pair dataset based on visible light images and corresponding shortwave infrared images acquired under different light intensities, and to calibrate the spectral mapping function based on the calibration image pair dataset; The image registration unit is used to register the visible light image acquired at the current moment to the image coordinate system of the shortwave infrared image, so as to obtain the registered visible light image; The background prediction unit is used to input the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current time to obtain the residual image, and to extract the target signal from the residual image. The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

2. The all-day star-sensor system based on dual-band image fusion according to claim 1, characterized in that: The specific operations of the calibration data storage unit include: A broadband halogen tungsten lamp light source is used to simultaneously and uniformly illuminate the target surface of the visible light detector and the target surface of the short-wave infrared detector. Under different light intensities, visible light images and corresponding short-wave infrared images of the same scene were collected to construct a calibration image pair dataset; A linear mapping model is constructed by sequentially inputting each calibrated image pair from the calibrated image pair dataset into the linear mapping model to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibrated image pair satisfies the following: S_level(x1, y1) = α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ): Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of the shortwave infrared detector, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of the visible light detector, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

3. The all-day star-sensor system based on dual-band image fusion according to claim 1, characterized in that: The incident end of the visible light lens group is equipped with a visible light shielding tube, and the incident end of the short-wave infrared lens group is equipped with a short-wave infrared shielding tube.

4. A sky-based, time-sensitive star-sensing method based on dual-band image fusion, implemented using the sky-based, time-sensitive star-sensing system based on dual-band image fusion as described in claim 1, characterized in that: Specifically, the steps include the following: S1: The all-day star-sensing system is mounted on the spacecraft. The target beam is transmitted to the visible light detector through the visible light lens group, and at the same time, the target beam is also transmitted to the short-wave infrared detector through the short-wave infrared lens group. S2: The synchronization trigger controls the visible light detector and the short-wave infrared detector to expose synchronously at the current moment, so that the visible light image and the short-wave infrared image are spatiotemporally aligned, and transmits the visible light image and the short-wave infrared image synchronously to the dual-band image data fusion processing module for processing at the current moment, and outputs the high-precision attitude information of the aircraft at the current moment. S3: Determine if the task has ended. If so, end the star-sensor detection. Otherwise, replace the current time with the next time and repeat step S2 until the star-sensor detection ends.

5. The all-day star-sensitive detection method based on dual-band image fusion according to claim 4, characterized in that: Step S2 specifically includes: S21: The calibration data storage unit constructs a calibration image pair dataset based on visible light images and corresponding short-wave infrared images collected under different light intensities, and obtains the spectrum mapping function based on the calibration image pair dataset; S22: The image registration unit registers the visible light image acquired at the current moment to the image coordinate system where the shortwave infrared image is located, and obtains the registered visible light image; S23: The background prediction unit inputs the registered visible light image into the spectrum mapping function to obtain the predicted background noise image of the shortwave infrared image; S24: The phase-locked elimination unit is used to subtract the predicted background noise image from the shortwave infrared image acquired at the current moment to obtain the residual image and extract the target signal in the residual image. S25: The attitude calculation unit is used to compare the target signal with the navigation satellite catalog, calculate the attitude of the aircraft based on the comparison results, and output the high-precision attitude information of the aircraft at the current moment.

6. The all-day star-sensitive detection method based on dual-band image fusion according to claim 5, characterized in that: Step S21 specifically includes: S211: Use a broadband halogen tungsten lamp light source to simultaneously and uniformly illuminate the target surface of the visible light detector and the target surface of the short-wave infrared detector. S212: Under different light intensities, collect visible light images and corresponding short-wave infrared images of the same scene to construct a calibration image pair dataset; S213: Construct a linear mapping model. Input each calibration image pair contained in the calibration image pair dataset into the linear mapping model in sequence to obtain the spectral mapping coefficient matrix and the offset matrix. The pixel mapping relationship of each calibration image pair satisfies: S_level(x1, y1) = α(x 2→1 , y 2→1 )×S_vis(x2, y2)+β(x 2→1 , y 2→1 ): Where S_swir(x1, y1) is the grayscale value of the shortwave infrared image at pixel (x1, y1) on the target surface of the shortwave infrared detector, and S_vis(x2, y2) is the grayscale value of the visible light image at pixel (x2, y2) on the target surface of the visible light detector, x1=x2, y1=y2, α(x 2→1 , y 2→1 Let β(x) be the spectral mapping coefficients corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1). 2→1 , y 2→1 ) represents the offset corresponding to the linear mapping from pixel (x2, y2) to pixel (x1, y1); S214: Construct a spectrum mapping function based on the spectrum mapping coefficient matrix and the offset matrix: F = A·M + B; Where F is the grayscale matrix of the predicted background noise image of the shortwave infrared image, A is the spectrum mapping coefficient matrix, M is the grayscale matrix of the visible light image, and B is the offset matrix.

Images