节点文献

海上溢油遥感图像智能边缘检测算法的研究

The Study of Intelligent Edge Detection Algorithm for the Remote Sensing Image of Oil-Spills on the Sea

【作者】 焦阳

【导师】 安居白;

【作者基本信息】 大连海事大学 , 计算机应用技术, 2005, 硕士

【摘要】 海洋溢油污染是各种海洋污染中影响范围最广、危害时间最长、对生态环境破坏最大的一种。针对目前的溢油污染状况,国家海洋局已有两架海监飞机用于海上溢油污染监测,并且即将配备遥感监测设备-红外传感器和紫外传感器,但是在溢油信息智能处理的软件方面仍存在空缺。本论文对红外或紫外遥感扫描图像进行处理,根据处理后的信息,获取海上溢油情况。 图像的边缘检测是遥感图像处理的重要手段。溢油图像具有两个主要特点:一个是溢油由于在风,浪,流的作用下,具有动态的特性,所以溢油图像没有固定的形状;另一个是海水与溢油的对比度不大,溢油图像上油与水的边界不是很清楚,具有一定的模糊性。针对这两个主要的特点,本论文对Pal.King边缘检测算法进行一定的改进,提出了一种模糊增强理论与遗传算法相结合的新的智能溢油遥感图像边缘检测算法。Pal.King算法通过提高图像边缘两侧的对比度,对模糊边缘和细小边缘具有很好的检测能力,但是一方面它的模糊变换算法复杂,计算量大;另一方面增强算法中固定的阀值对形状不固定的溢油图像检测效果不理想。本论文提出的算子包含两部分:利用改进的模糊变换算子简化了Pal.King算子中复杂的G和G-1运算,提高了计算速度;另一方面采用二次寻优遗传算法根据不同的溢油图像获取相应的模糊增强阀值,提高了图像处理的自适应性。 本论文最后给出了该算子和其他算子的溢油遥感图像边缘检测结果,通过对比我们发现,该算子得到的溢油图像边缘比其他算子得到的边缘更加清晰和连贯,根据该边缘检测信息,海上溢油能够更加容易和快速地被识别。

【Abstract】 Edge detection is an important part of remote sensing image processing for the detection of oil-spills on the sea. To detect the oil-spills more effectively, in this paper the author makes some improvements to the Pal. King fuzzy edge detection algorithm and presents a new algorithm that combines fuzzy enhancement theory and Genetic Algorithms (GA) to get the edge of oil-spills image according to the two main characteristics of the oil-spills image-one is the shape of oil-spills is unfixed and the other is the edge between the oil and water is fuzzy.The Pal. King algorithm has a good capacity for the detection of fuzzy edge and thin edge through improving the contrast between the two sides of edge, but with complex and big calculation on the one hand and the fixed threshold value doesn’t fit for the different kind of images processing on the other hand. The algorithm presented here is composed of two parts: improved fuzzy enhancement algorithm that simplifies the complex G and G-1 calculation in the Pal. King algorithm, twice-optimization GA that can get the threshold value in the enhancement algorithm precisely and quickly to the different kind of image processing.At last, the author gives both its image edge processing result and other algorithms’. Through comparison, we can see that the oil’s edge is more legible and continuous than the other algorithms’ and oil-spills on the sea is easier and faster to be identified.

  • 【分类号】TP751
  • 【被引频次】5
  • 【下载频次】450
节点文献中: