Zero-Shot Remote Sensing Image Dehazing Based on a Re-Degradation Haze Imaging Model

Abstract

Image dehazing is crucial for improving the advanced applications on remote sensing (RS) images. However, collecting paired RS images to train the deep neural networks (DNNs) is scarcely available, and the synthetic datasets may suffer from domain-shift issues. In this paper, we propose a zero-shot RS image dehazing method based on a re-degradation haze imaging model, which directly restores the haze-free image from a single hazy image. Based on layer disentanglement, we design a dehazing framework consisting of three joint sub-modules to disentangle the hazy input image into three components: the atmospheric light, the transmission map, and the recovered haze-free image. We then generate a re-degraded hazy image by mixing up the hazy input image and the recovered haze-free image. By the proposed re-degradation haze imaging model, we theoretically demonstrate that the hazy input and the re-degraded hazy image follow a similar haze imaging model. This finding helps us to train the dehazing network in a zero-shot manner. The dehazing network is optimized to generate outputs that satisfy the relationship between the hazy input image and the re-degraded hazy image in the re-degradation haze imaging model. Therefore, given a hazy RS image, the dehazing network directly infers the haze-free image by minimizing a specific loss function. Using uniform hazy datasets, non-uniform hazy datasets, and real-world hazy images, we conducted comprehensive experiments to show that our method outperforms many state-of-the-art (SOTA) methods in processing uniform or slight/moderate non-uniform RS hazy images. In addition, evaluation on a high-level vision task (RS image road extraction) further demonstrates the effectiveness and promising performance of the proposed zero-shot dehazing method.

Self-Supervised Remote Sensing Image Dehazing Network Based on Zero-Shot Learning

“…Kar et al's [16] method requires no training data and can restore images with various types of degradation, such as haze, rain, and snow; however, the enormous training time is a major issue. Wei et al [15] propose a three-branch RS image-dehazing network regulated by a re-degradation haze imaging model to achieve zero-shot learning; however, this model is less effective in non-uniform hazy conditions.…”

“…(min(J (x), 0.1) + min(t (x), 0.1)) (15) where x represents the image pixel, and N denotes the total number of image pixels. The L min ensures that the pixel intensity of J and t is not lower than 0.1, compensating for the over-clipping of pixel intensity caused by L DCP .…”

“…Additionally, some dehazing models trained on synthetic hazy datasets fail to produce satisfactory results on real-world hazy images, resulting in domain-shift issues. To address these challenges, zero-shot dehazing methods [13][14][15][16] that use a single hazy image for model training have been proposed to improve model generalizability. For instance, Wei et al [15] proposed a re-degradation haze imaging model for remote sensing (RS) image dehazing without any pre-training or fine-tuning on specific datasets.…”

Hierarchical Semantic-Guided Contextual Structure-Aware Network for Satellite Image Dehazing

“…For instance, DehazeFormer [8] exemplifies this trend with its innovative normalization layer and spatial information aggregation scheme. In the past two years, some dehazing methods tailored for satellite images [18][19][20][21][22] are explored, promoting better applications of satellite images and sharpening our view from space.…”