摘要
提出将高斯拉普拉斯算子应用在光电联合相关变换器中进行谱面图像的增强处理。光电混合联合变换器可实现对目标的实时探测、识别及自动定位,但由于实际中采集到的图像的对比度较低,且存在大量背景噪音,影响了目标的识别率。根据高斯拉普拉斯变换对高斯噪声不敏感的特性,结合了自适应阈值、边界跟踪和细化技术,对图像噪声进行滤波的同时,对图像进行了增强处理,这样最大限度地保留了光谱图像的细节信息,提高了光电联合相关系统的目标识别率。
Gauss-Laplace operator was used in the hybrid photoelectric joint transform correlator for image enhancement of joint power spectrum. Using hybrid photoelectric joint transform correlation can realize real-time detection, recognition and automatic localization of objects, but the images capture from practice have lower contrast and mass background noises, which affect the recognition rate of objects. According to the characteristic of Gauss-Laplace transform is not sensitive to Gaussian noise, and using adaptive threshold, boundary tracking, thin algorithm, we can filter the image noises, at same time, enhance the images. It can reserve the detail information of joint power spectrum, and improve the object recognition rate of optoelectronical hybrid joint transform correlation system.
出处
《半导体光电》
EI
CAS
CSCD
北大核心
2007年第6期881-883,共3页
Semiconductor Optoelectronics
基金
部级预先研究项目(41317030106)
关键词
联合相关变换器
高斯拉普拉斯
算子
图像增强
joint transform correlation
Gauss-Laplace
operator
image enhancement