Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new s...Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new systems require large amounts of labeled data for model retraining,but collecting sufficient data in a short time is often infeasible.To address this contradiction,this paper proposes a data generation and transfer framework,integrating a stable diffusion model with attention distillation,that leverages historical SAR data to synthesize training data tailored to the unique characteristics of new SAR systems.Specifically,we fine-tune the low-rank adaptation(LoRA)modules within the multimodal diffusion transformer(MM-DiT)architecture to enable class-controllable SAR image generation guided by textual prompts.To ensure that the generated images reflect the statistical properties and imaging characteristics of the target SAR system,we further introduce an attention distillation mechanism that transfers sensor-specific features,such as spatial texture,speckle distribution,and structural patterns,from real target-domain data to the generative model.Extensive experiments on multi-class aircraft target datasets from two real spaceborne SAR systems demonstrate the effectiveness of the proposed approach in alleviating data scarcity and supporting cross-sensor remote sensing applications.展开更多
Recent exploration results indicate that a significant exploration potential remains in the Dongying Depression of the Bohai Bay Basin and the undiscovered oil and gas are largely reservoired in subtle traps including...Recent exploration results indicate that a significant exploration potential remains in the Dongying Depression of the Bohai Bay Basin and the undiscovered oil and gas are largely reservoired in subtle traps including turbidite litholigcal traps of the Sha-3 Member. In order to effectively guide the exploration program targeting turbidites, this study will focus on the depositional models of the Sha-3 Member turbidites and oil/gas accumulation characteristics in these turbidites. Two corresponding relationships were found. One is that the East African Rift Valley provides a modern analog for the depositional systems in the Dongying Depression. The other is that the depositional models of line-sourced slope aprons, single point-source submarine fan and multiple source ramp turbidite, established for deep-sea turbidites, can be applied to interpret the depositional features of the turbidite fans of three different origins: slope turbidite aprons, lake floor turbidite fans and delta-fed turbidite fans in the Sha-3 Member. Updip sealing integrity is the key factor determining whether oil/gas accumulates or not in the slope aprons and lake floor fans. The factors controlling oil/gas migration and accumulation in the delta-fed turbidite fans are not very clear. Multiple factors rather than a single factor probably played significant roles in these processes.展开更多
基金supported in part by the National Natural Science Foundations of China(Nos.62201027,62271034)。
文摘Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new systems require large amounts of labeled data for model retraining,but collecting sufficient data in a short time is often infeasible.To address this contradiction,this paper proposes a data generation and transfer framework,integrating a stable diffusion model with attention distillation,that leverages historical SAR data to synthesize training data tailored to the unique characteristics of new SAR systems.Specifically,we fine-tune the low-rank adaptation(LoRA)modules within the multimodal diffusion transformer(MM-DiT)architecture to enable class-controllable SAR image generation guided by textual prompts.To ensure that the generated images reflect the statistical properties and imaging characteristics of the target SAR system,we further introduce an attention distillation mechanism that transfers sensor-specific features,such as spatial texture,speckle distribution,and structural patterns,from real target-domain data to the generative model.Extensive experiments on multi-class aircraft target datasets from two real spaceborne SAR systems demonstrate the effectiveness of the proposed approach in alleviating data scarcity and supporting cross-sensor remote sensing applications.
文摘Recent exploration results indicate that a significant exploration potential remains in the Dongying Depression of the Bohai Bay Basin and the undiscovered oil and gas are largely reservoired in subtle traps including turbidite litholigcal traps of the Sha-3 Member. In order to effectively guide the exploration program targeting turbidites, this study will focus on the depositional models of the Sha-3 Member turbidites and oil/gas accumulation characteristics in these turbidites. Two corresponding relationships were found. One is that the East African Rift Valley provides a modern analog for the depositional systems in the Dongying Depression. The other is that the depositional models of line-sourced slope aprons, single point-source submarine fan and multiple source ramp turbidite, established for deep-sea turbidites, can be applied to interpret the depositional features of the turbidite fans of three different origins: slope turbidite aprons, lake floor turbidite fans and delta-fed turbidite fans in the Sha-3 Member. Updip sealing integrity is the key factor determining whether oil/gas accumulates or not in the slope aprons and lake floor fans. The factors controlling oil/gas migration and accumulation in the delta-fed turbidite fans are not very clear. Multiple factors rather than a single factor probably played significant roles in these processes.
