基于骨架的垂测电离图F层自动判读算法

Method on F layer of ionogram autoscaling based on skeleton

为了实现F层参数的自动度量,为电离层短期监测提供实时可靠的数据,以MATLAB7.6.0为仿真平台,首先对F层描迹图像进行预处理去除离散噪声点;针对F层描迹常见的寻常波和非寻常波重叠的现象,根据描迹形态特征,采用基于形态学算子的骨架提取方法以及骨架分解算法提取出寻常波;利用最小二乘分离F1层和F2层;结合形态学重建、Radon变换以及图像投影方法读取参数。该方法可以实时自动读取F层主要参数并获得较好的识别率。在不考虑Es层的多次反射遮蔽情况下,实验验证了方法的可行性同时具有一定的自抗干扰能力。

In order to realize the autoscaling of F parameters and offer the real-time and reliable data for the short term monitoring of the ionosphere, this paper firstly pretreated the trace of F layer to remove discrete noise using MATLAB7.6.0 simulation platform. In F layer , the trace of ordinary wave and extraordinary wave overlap commonly and had morphological characteristics. Aiming at this phenomenon, an algorithm of extracting skeleton based on morphological operator and a skeleton decomposition approach were Used to extract ordinary wave. Then applied the least-square method to fit the trace of ordinary wave and separated F1 and F2. Finally, combined morphological reconstruction, Radon transform and image projection method to get F parameters. This method could provide main parameters in real time automatically and had high recognition accuracy. The ex- periment confirms the feasibility of this method without considering the case that Es layer blankets F1 layer seriously and it has certain degree of self-anti-interference ability in the absence of noise.