Turbulence degradation image semantic segmentation method based on boundary perception and adversarial learning

Abstract

The invention relates to a turbulence degradation image semantic segmentation method based on boundary perception and adversarial learning, and the method comprises the following steps: (1), combininga simulated turbulence-degraded image with a real turbulence-degraded image based on a turbulence imaging physical model for a turbulence-degraded image, and constructing a turbulence-degraded imagesemantic segmentation data set; (2) for the data set obtained in the step (1), constructing a DeepLabV3 + semantic segmentation model based on boundary perception in combination with fuzzy and distortion characteristics of a turbulence degradation image; (3) taking the DeepLabV3 + semantic segmentation model based on boundary perception in the step (2) as a generator, and constructing a boundary perception generative adversarial network GAN model based on adversarial learning in combination with a discriminator composed of five convolutional layers; and (4) for the GAN model obtained in the step (3), performing model training on the turbulence degradation image semantic segmentation data set obtained in the step (1) to obtain a trained semantic segmentation GAN model, and performing semantic segmentation on the turbulence degradation image by using the trained semantic segmentation adversarial network GAN model to obtain a prediction segmentation graph.

Description

technical field [0001] The present invention relates to a method for semantic segmentation of turbulent degraded images based on boundary perception and confrontation learning, which is a deep model that combines boundary perception algorithms and generative confrontation networks (GAN), and is suitable for semantic segmentation of images degraded by atmospheric turbulence Task. Background technique [0002] Semantic segmentation is widely used in various industries such as intelligent driving, security monitoring, and industrial inspection, and is a very challenging task in the field of computer vision. Semantic segmentation of turbulent degraded images refers to the task of pixel-level classification of degraded images affected by atmospheric turbulence. Atmospheric turbulence will have a serious impact on the imaging performance of the optical system, causing degradation phenomena such as distortion and blurring of the observed image, thereby reducing the accuracy of the...

AI Technical Summary

Problems solved by technology

However, there are few methods for semantic segmentation of turbulent degraded images. The current research difficulties mainly lie in the following aspects: (1) Turbulent degraded images have two characteristics: blur and distortion. Compared with conventional images, the image quality is lower. poor, blurred edges, object distortion, low contrast, difficult to express texture information, and often contains noise, it is difficult to perform accurate semantic segmentation; Double challenges, not only need to use global information to solve the problem of semantic discrimination, but also need to use local information to solve the problem of detail positioning; (3) At present, the data sets for semantic segmentation of turbulent degraded images are still very scarce, and turbulent flow needs to be manually obtained and established Degraded image datasets are more difficult; (4) Existing semantic segmentation methods based on deep learning generally only target higher-quality images of common scenes, and are not suitable for turbulent degraded images

Images