Wild animal image denoising method based on sparse error constraint representation

Abstract

A wild animal image denoising method based on sparse error constraint representation comprises the following steps: firstly, performing high and low frequency decomposition on a wild animal image by adopting 2D-DCT transformation, and performing Gaussian multi-scale transformation on a high-frequency image; then, respectively constructing a structured dictionary of a high-frequency image and a high-frequency multi-scale image based on regularization sparse error constraint, and solving a target function containing a non-convex but smooth p norm by using a Riemann conjugate gradient minimization method; on the basis, a PALM algorithm is introduced to solve the minimization problem of a non-convex and non-smooth function in a high-frequency image noise reduction model, and the PALM algorithm is used for updating a dictionary and a sparse coefficient; and finally, fusing the high-frequency image and the low-frequency image to generate a denoised image. The method is suitable for denoising of animal images shot in the field, and has important research significance and application value.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a wild animal image denoising method based on sparse error constraint representation. Background technique [0002] In recent years, many species of wild animals are on the verge of extinction, and people are paying more and more attention to how to protect wild animals. The use of cameras to monitor the behavior of wild animals has become an important auxiliary means of wildlife protection research. Through the research on wild animal images, such as image segmentation, feature extraction, target recognition, etc., it is helpful to discover new species, identify animal types and behaviors, etc. However, animal images taken in the wild will inevitably be subject to external interference during the shooting and transmission process, resulting in noise in the image, and most existing noise reduction methods are to artificially add noise to the clean image, and then add noi...

AI Technical Summary

Problems solved by technology

However, animal images taken in the wild will inevitably be subject to external interference during the shooting and transmission process, resulting in noise in the image, and most existing noise reduction methods are to artificially add noise to the clean image, and then add noise to the image. image noise reduction

The noise of animal images taken in the wild is random mixed noise, and will be affected by bad weather such as strong wind and fog. The existing noise reduction algorithms are not suitable for denoising wild animal images. Noise method has important research significance and application value

Images