基于视觉加权处理的SAR干扰效果评估

摘要

传统SAR图像质量评价指标没有考虑像素点间的相关性和人眼视觉系统感知特性,导致主客观不一致的现象时有发生.鉴于此,提出了一种新的评估方法.该方法将对比度敏感度作为主客观评价的链接纽带,利用小波变换与人类视觉系统多通道特性相匹配的特点,对目标图像进行多级小波分解,由分解后的各小波分量提取各频段相应的传统评估指标,然后将其与各频段加权系数进行非线性合并,并以此作为综合评价的度量指标.仿真实验结果表明,新方法较之传统评估手段明显更具优势,不仅能够反映出干扰对图像细节的影响,而且能够与人类视觉系统保持高度一致,提升了整套评估系统的效率和准确率.%Without considering correlation of pixels and the perceptual characteristics of human vision system(HVS), traditional SAR image quality evaluation indexes are limited, so that the subjective and the objective means can not come to consistent. On the basis, a new evaluation method is proposed. It makes good use of CSF as juncture between the subjective and the objectivc cvaluation methods, as well as the characteristic of wavelet transformation which matches the human vision system's multichannel model. Multi-level wavelet decomposition is then applied to the image for obtaining traditional evaluation indexes in sub-bands.Then a synthctic indcx is acquired by nonlinear combination, which is used as final comprehensive evaluation index. The results from experiments indicate that the new method has an obvious advantage over traditional evaluation methods. Besides reflecting the influences of jamming on detail information in image, it can keep identical to human vision system, and furthermore, it has improved the efficiency and accuracy of entire eval uation system.