Diffusion models (DM) have become state-of-the-art generative models because of their capability of generating high-quality images from noises without adversarial training. However, they are vulnerable to backdoor attacks as reported by recent studies. When a data input (e.g., some Gaussian noise) is stamped with a trigger (e.g., a white patch), the backdoored model always generates the target image (e.g., an improper photo). However, effective defense strategies to mitigate backdoors from DMs are underexplored. To bridge this gap, we propose the first backdoor detection and removal framework for DMs. We evaluate our framework Elijah on over hundreds of DMs of 3 types including DDPM, NCSN and LDM, with 13 samplers against 3 existing backdoor attacks. Extensive experiments show that our approach can have close to 100% detection accuracy and reduce the backdoor effects to close to zero without significantly sacrificing the model utility.

In the Arctic and Antarctic scientific expeditions, ice-penetrating radar is an effective method for studying the bedrock under the ice sheet and ice information within the ice sheet. Because of the low conductivity of ice and the relatively uniform composition of ice sheets in the polar regions, ice-penetrating radar is able to obtain deeper and more abundant data than other geophysical methods. However, it is still necessary to suppress the noise in radar data to obtain more accurate and plentiful effective information. In this paper, the entirely non-recursive Variational Mode Decomposition (VMD) is applied to the data noise reduction of ice-penetrating radar. VMD is a decomposition method of adaptive and quasi-orthogonal signals, which decomposes airborne radar data into multiple frequency-limited quasi-orthogonal Intrinsic Mode Functions (IMFs). The IMFs containing noise are then removed according to the information distribution in the IMF’s components and the remaining IMFs are reconstructed. This paper employs this method to process the real ice-penetrating radar data, which effectively eliminates the interference noise in the data, improves the signal-to-noise ratio and obtains the clearer inner layer structure of ice. It is verified that the method can be applied to the noise reduction processing of polar ice-penetrating radar data very well, which provides a better basis for data interpretation. At last, we present fine ice structure within the ice sheet based on VMD denoised radar profile.

In seismic refraction exploration, travel time tomography is the most widely used method in engineering and environmental geophysical exploration. In this paper, we mainly optimize the travel time tomography of refraction. First, with respect to the forward algorithm, we introduce a new travel time calculation method to improve the accuracy and efficiency of forward calculation. Based on the fast marching method (FMM), we introduce an improved forward calculation method called the multi-stencil fast marching method (MSFM). In the process of inversion, we propose a dynamic prior model composite constraint (DPMCC) method based on the T0 difference method from the idea of multi-scale inversion. Meanwhile, we use the prior information to improve the accuracy of inversion. Furthermore, we use the dynamic regularization factor selection method to make the inversion solution more stable and reliable. Finally, we test and analyze the synthetic data and the measured data to verify the effectiveness of the optimized travel time tomography algorithm.

On December 14, 2013, China’s Chang ’e-3 missions landed on the lunar surface, providing valuable information for better study of the moon. LPR (Lunar penetrating radar) is a very important and effective detection method. The lunar surface shallow radar data we obtained through the lunar exploration radar can be used to understand the geological structure of the lunar surface and the thickness of the lunar regolith. This paper analyzes the processed LPR data of “Chang’e-3” by using instantaneous attributes, and obtains the instantaneous amplitude attributes, instantaneous phase attributes, and instantaneous frequency attributes of the radar profile. According to the results of the instantaneous attribute analysis, the lunar regolith structure of the radar profile is divided, and it is found that it can be well matched with the previous research results.

Modeling plays an important role in engineering exploration. However, the traditional rectangular mesh cannot meet the requirement because the actual shape of the model may be very complex. Therefore, this paper chooses the Delaunay grid for modeling and applies this method to magnetic numerical simulation. At the same time, we make use of the advantages of the Delaunay grid and choose to use grid encryption technology in the boundary and complex area of the model. Then, the formula for calculating the magnetic anomaly of tetrahedron is deduced in detail, and the magnetic anomaly with surface fluctuation is calculated. For a synthetic data model, Delaunay grids with different densities are used to model and calculate surface anomalies. The two results are compared with the analytical solution of the model. The results show that the method has high accuracy. Then, the method is applied to the actual geological body modeling in the Jinchuan mining area. According to local needs, a three-dimensional model with inhomogeneous grid density is established and the surface magnetic field of the model is calculated. Finally, the simulation data are compared with the measured data. The results show that the Delaunay grid modeling method has strong applicability.