Remote sensing image dehazing method in mountaineering equipment.

This article focuses on developing an improved Retinex algorithm-based haze removal model to enhance the quality of remote sensing images used by intelligent mountaineering equipment. The model integrates a multiscale Retinex with color restoration (MSRCR) algorithm, Gaussian kernel filtering, and guided filtering to address haze interference and improve image texture and detail. Experimental evaluation using a large dataset of Landsat 8 satellite images demonstrates that the improved Retinex model (I-Retinex) outperforms traditional Retinex, generative adversarial networks (GANs), and Alex neural network models in metrics such as normalized information entropy, mean squared error, mean absolute error, peak signal-to-noise ratio, and mean gradient, though it requires moderate computational time and memory. The study concludes that the proposed model effectively removes haze and preserves image details, thereby supporting enhanced positioning and recognition functions in intelligent mountaineering devices, with potential for further research involving larger datasets.