强噪声背景下在遥感云图中提取地球圆盘的快速算法

A FAST ALGORITHM OF EARTH-EDGE DETECTION FOR SPACE REMOTE SENSING IMAGES UNDER STRONG NOISES

提出一种新的图像边缘提取的快速算法———通过建立马尔可夫预测模型对目标边缘位置进行前向预测 ,并根据边缘特性 ,采用最小均方误差分类器进行同步校正 .与传统的边缘提取算法相比 ,该算法具有很强的噪声抑制和实时处理能力 .实践表明 ,该算法对于任何非突变闭合区域均有良好的边缘跟踪特性 。

A new fast algorithm of the edge detection for the images-the Markov prediction model is presented, which predicts the positions of the edge-points. Based on the merits of the target's edge, the LMS classifier is used to correct the predicted outputs at the same time. Comparing with the traditional method, the algorithm has a stronger ability of restraining noise and real-time processing. The experimental results show that the algorithm has a better detection to any zone of non-mutation and closed. It has been applied to the real-time acquisition and processing system for the space remote sensing successfully.