航天遥感相机星上自适应图像增强研究与实验

Study and Experiment of Adaptive Image Enhancement on Satellite for Space Remote Sensing Camera

摘要:

为了解决航天遥感图像在地面增强处理带来的滞后性和失真性,因此提出了星上自适应图像增强的方法和一种新的基于自适应线性拉伸的拉普拉斯滤波算法,来实现星上图像增强处理的实时性和自适应性;在拉普拉斯滤波的基础上,用自适应调整参数A和B对图像的对比度和边缘进行增强,该算法结构简单,运算量小；最后建立以FPGA为核心的硬件系统平台,采用流水线的处理方式,对该算法进行实验验证；实验结果表明,与其他增强算法相比,文章算法增强的图像目视效果更好,信息熵提高了10.21%,处理一幅图像所需的时间是79.6 ms,满足航天遥感相机星上自适应图像增强实时性的要求,达到了预期效果。

Abstract:

A method of Adaptive image enhancement on satellite is proposed to solve the problem of Hysteresis and complexity, which is caused by remote sensing image enhancement on the ground, and in order to achieve real-time and self-adaptive of image enhancement, this paper proposes a new Laplace filtering algorithm based on adaptive linear stretch. On the basis of the Laplace filter, enhance the contrast and edge of the image with adaptive parameters A and B. The algorithm is simple and has small amount of computation. A platform is established with FPGA as the core of the hardware and then verifies the feasibility of the algorithm by pipelining processing of two images. The results show that the information entropy improved 10.21% by using this algorithm, and the visual effects is better than the other algorithms. The operation speed of an image is 79.6ms. It meet the real-time requirements of adaptive image enhancement on satellite, we achieved the expected results.