In the domain of infrared small target detection (IRSTD), the challenges revolve around detecting small and faint targets from infrared images. These targets lack distinct textures and morphology exist in complex backgrounds with numerous distractions. Current deep-learning methods typically prioritize preserving target features while neglecting the crucial background context, ultimately resulting in false alarms and miss detection. To tackle this issue, we propose a novel approach involving separately focusing on candidate target responses and background context during the encoding stage and aligning them during the decoding stage. Specifically, we introduce the progressive background-aware transformer (PBT) which adopts an asymmetric encoder-decoder architecture. The encoder with task-specific frequency domain priors extracts candidate target responses and background context features separately from shallow and deep blocks, respectively. The following hierarchical decoder progressively refines the candidate target responses under the guidance of rich background context stage by stage, leading to more accurate results. Our experiments demonstrate that PBT surpasses state-of-the-art IRSTD methods across various datasets. The code and dataset are available athttps://github.com/Heron0625/PBT.

The lunar mineral spectrometer (LMS) onboard the Chang’e-5—China’s lunar exploration mission—collected in-situ multispectral data up to 3 μm over a large area of the lunar surface for the first time. As Chang’e-5 operated near the Moon’s noon local time, the spectral data are influenced by thermal radiation. In particular, the LMS multispectral data are in the form of characteristic spectral bands with discontinuities that make it difficult to directly apply existing empirical models for thermal removal. Here, we present a thermal removal method based on the one-dimensional heat conduction model and real terrain data. A centimeter-scale temperature distribution map of the multispectral measurement region is calculated, which ranged from 330 to 362 K. We calculated the temperatures of four hyperspectral regions using an empirical model, compared them with those of the same region using the proposed method, and obtained a relative error of less than 1%. Afterwards, we removed the effects of thermal radiation in the multispectral data using a thermal removal model with the temperature of the multispectral region. We compared multispectral data after thermal removal using our proposed method with hyperspectral data after thermal removal using an empirical model at similar locations. The results showed good consistency between the data corrected by both methods. This paper proposed method does not depend on the setting of the spectral band, it can provide a reference for thermal removal of Chang’e-5 multispectral data and that of subsequent in-situ non-continuous spectral detection data.

The filter performance will be degraded in the measurements with time-varying and unknown noise statistics. To combat the above challenges, a modified strong tracking slide window variational adaptive Kalman filter algorithm is proposed in this article. First, the multiple fading factors are integrated into the proposed algorithm to adjust the error covariance. Next, an improved adaptive slide window method is designed for variational Bayesian (VB) Kalman filtering by adaptively adjusting the slide window size and correcting the previous state according to the later state, which improves the estimation accuracy and computational efficiency. Finally, the inverse Wishart distribution is considered for modeling process and measurement noise, and the state vector, as well as noise statistics, are inferred via the VB technique without prior noise covariance information. Simulation results demonstrate that the proposed filter algorithm is more robust than existing filters in counteracting measurement and process noise uncertainties.

The lunar mineralogical spectrometer (LMS) is the primary scientific payload of the Chang’E-5 mission of the China Lunar Exploration Program. The LMS is responsible for thein situspectral detection and analysis of the sampling areas of interest on the Moon’s surface. The LMS needs to adapt to a wide range of temperature conditions, varying between the lunar morning to noon, based on the time, location, and work process of the LMS. Thus, ensuring consistency in the spectral data obtained at high temperatures and a wide range of operating environments is a major challenge. In this study, a thermal analysis model of the LMS is built, and the temperature variation of the LMS during its operation on the lunar surface is simulated using the instrument workflow that is based on the simulation results. Independent experiments were also carried out for the temperature-sensitive components in the LMS, and subsequently, the temperature correction model and model coefficients for each temperature-sensitive component were obtained. The final correction result was a quantitative assessment of the ground test data and thein situlunar detection data. For the ground tests at different temperatures, the average errors were found to be 0.57% and 1.2% for the short- and medium-wave data, respectively, after applying the temperature correction. When the LMS was operating on the lunar surface, the average errors for the short- and medium-ware data after the temperature correction were calculated to be 0.59% and 0.8%, respectively.

The Visible and Near-infrared Imaging Spectrometer (VNIS) is one of the scientific payloads onboard the Yutu-2 lunar rover of the Chang’E-4 mission. Based on Acousto-Optic Tunable Filter (AOTF) dispersion, a moving platform and a calibration unit, the VNIS was designed for lunar surface observations of the far side of the Moon for composition exploration and spectral experiments in the wavelength range of 0.45–2.4μm. Large differences in radiance caused by the discrepancy between the laboratory and lunar environment could affect the spectral shape and features. Therefore, in-flight calibration of the VNIS is critical for the interpretation of the acquired spectra. To precisely map the digital output to the spectral radiance, a calibration unit was installed on the VNIS to facilitate radiometric calibration measurements on the lunar surface. After the elimination of stray light and the shadow effect, the measured radiance of the calibration unit was derived by digital multiplication of the pre-flight coefficients. Considering the observation geometry, Moon-Sun distance, and the bi-directional reflectance factor (BRF) of the calibration panel, the theoretical radiance at the entrance pupil of VNIS was calculated. Using linear regression, in-flight calibration of the coefficients was performed and used to correct the spectral data. The performance of the spectrometer was evaluated using 12 lunar-day data. It was determined that the coefficients gradually converged and the variation was less than 1%, indicating that the in-flight calibration was robust. The proposed in-flight calibration method facilitated the acquisition of reliable data for lunar composition interpretation.