This paper presentsDiffFuSR, a modular pipeline for super-resolving all 12 spectral bands of Sentinel-2 Level-2A imagery to a unified ground sampling distance (GSD) of 2.5 meters. The pipeline comprises two stages: (i) a diffusion-based super-resolution (SR) model trained on high-resolution RGB imagery from the NAIP and WorldStrat datasets, harmonized to simulate Sentinel-2 characteristics; and (ii) a learned fusion network that upscales the remaining multispectral bands using the super-resolved RGB image as a spatial prior. We introduce a robust degradation model and contrastive degradation encoder to support blind SR. Extensive evaluations of the proposed SR pipeline on the OpenSR benchmark demonstrate that the proposed method outperforms current SOTA baselines in terms of reflectance fidelity, spectral consistency, spatial alignment, and hallucination suppression. Furthermore, the fusion network significantly outperforms classical and learned pansharpening approaches, enabling accurate enhancement of Sentinel-2’s 20 m and 60 m bands. This work proposes a novel modular framework Sentinel-2 SR that utilizes harmonized learning with diffusion models and fusion strategies. Our code and models can be found at https://github.com/NorskRegnesentral/DiffFuSR.

Diffusion models have obtained photo-realistic results on various super-resolution tasks. However, existing approaches typically require the availability of high-resolution and paired training data, which often is not readily available in many remote sensing scenarios. To enhance multi-spectral Sentinel 2 (S2) satellite images – at a ground sampling distance (GSD) ranging from 10m to 60m – without requiring high-resolution or paired training data, we therefore propose and evaluate a novel set of approaches to leverage traditional pansharpening within a diffusion model context to simulate the required training data. We extensively compare the proposed methods and demonstrate that by utilizing unpaired Spot-6/7 data, we are able to produce photo-realistic S2 images at a resolution of 2.5m.

Welding is a fabrication process used to join or fuse two mechanical parts. Modern welding machines have automated lasers that follow a pre-defined weld seam path between the two parts to create a bond. Previous efforts have used simple computer vision edge detectors to automatically detect the weld seam edge on an image at the junction of two metals to be welded. However, these systems lack reliability and accuracy resulting in manual human verification of the detected edges. This paper presents a neural network architecture that automatically detects the weld seam edge between two metals with high accuracy. We augment this system with a pre-classifier that filters out anomalous workpieces (e.g., incorrect placement). Finally, we justify our design choices by evaluating against several existing deep network pipelines as well as proof through real-world use. We also describe in detail the process of deploying this system in a real-world shop floor including evaluation and monitoring. We make public a large, well-labeled laser seam dataset to perform deep learning-based edge detection in industrial settings.

We present SIT-SR 3D, a novel self-supervised method for 3D single image super-resolution (SISR). Scaling 2D SISR networks to 3D SISR requires code redesign, high computing resources, and 3D ground-truth. However, we circumvent this by (1) using a pre-trained 2D SISR for indirect supervision and (2) using a novel consistency loss to learn frame interpolation. Any pre-trained state of the art 2D SISR method can replace the 2D SISR used in SIT-SR 3D, thus transferring the merits of 2D to 3D and ensuring modularity. We trained two end-to-end 3D baselines in a supervised setting; a 3D RRDBNet trained only with L1 loss and a 3D ESRGAN trained with adversarial and perceptual loss. We show that the proposed pipeline's self-supervised version is qualitatively better than the baselines. When trained in a supervised setting, SIT-SR 3D achieves better PSNR than its counterparts. Furthermore, our pipeline uses fewer parameters compared to the baselines. We demonstrate our results on an open-source digital rock CT dataset. Our code and pre-trained models will be made publicly available.

Our research aims to reduce the cost of pressure sensor calibration through machine learning. Pressure sensor calibration is a standard process whereby freshly manufactured pressure sensors are subjected to various controlled temperature and pressure setpoints to compute a mapping between the sensor's output and true pressure. Traditionally this mapping is calculated by fitting a polynomial with calibration data. Obtaining this data is costly since a large spectrum of temperature and pressure setpoints are required to model the sensor's behavior. We present a machine learning approach to predict a pre-defined calibration polynomial's parameters while requiring only one-third of the calibration data. Our method learns a pattern from past calibration sessions to predict the calibration polynomial's parameters from partial calibration setpoints for any newly manufactured sensor. We design a novel polynomial hypernetwork coupled with Fourier features and a weighted loss to solve this problem. We perform extensive evaluations and show that the current industry-standard method fails under similar conditions. In contrast, our approach saves two-thirds of the calibration time and cost. Furthermore, we conduct comprehensive ablations to study the effect of Fourier mapping and weighted loss. Code and a novel calibration dataset validated by calibration engineers are also made public.

We present RL-GAN-Net, where a reinforcement learning (RL) agent provides fast and robust control of a generative adversarial network (GAN). Our framework is applied to point cloud shape completion that converts noisy, partial point cloud data into a high-fidelity completed shape by controlling the GAN. While a GAN is unstable and hard to train, we circumvent the problem by (1) training the GAN on the latent space representation whose dimension is reduced compared to the raw point cloud input and (2) using an RL agent to find the correct input to the GAN to generate the latent space representation of the shape that best fits the current input of incomplete point cloud. The suggested pipeline robustly completes point cloud with large missing regions. To the best of our knowledge, this is the first attempt to train an RL agent to control the GAN, which effectively learns the highly nonlinear mapping from the input noise of the GAN to the latent space of point cloud. The RL agent replaces the need for complex optimization and consequently makes our technique real time. Additionally, we demonstrate that our pipelines can be used to enhance the classification accuracy of point cloud with missing data.