论文部分内容阅读
油气平台作为海上油气资源勘探开发的主要设备之一,其数量与空间分布反映了一个区域油气资源的开发状况。普通光学影像易受天气状况影响,而雷达数据成本较高,这给油气平台的检测带来了一定的困难。由于油气平台作业过程中,需要灯光照明,同时伴气的燃烧也产生很强的灯光,因此可以通过检测灯光来实现油气平台的提取。针对VIIRS数据具有强夜间光探测能力,本文提出了一种卷积运算临界值法对海上油气平台进行提取。首先对2期不同时相的VIIRS数据进行卷积运算,对亮像元进行了增强处理,对背景像元进行弱化处理,从而明确区分疑似目标与背景,然后以0值为分界点,对疑似目标进行提取,最后利用油气平台相对静止的特性,通过邻域分析,实现油气平台的提取。结果表明,本文提出的卷积运算临界值法可以有效地提取油气平台,提取准确率约为85.4%,同时可以有效地减少由于经验阈值对提取结果造成的误差。
As one of the major equipment for exploration and development of offshore oil and gas resources, the quantity and spatial distribution of oil and gas platforms reflect the development of oil and gas resources in a region. Ordinary optical images are easily affected by the weather conditions, while the radar data cost is high, which brings certain difficulties to the testing of oil and gas platforms. As the oil and gas platform operation process, the need for lighting, accompanied by the combustion of light also produces a strong light, so you can detect the light to achieve oil and gas platform extraction. In view of VIIRS data has strong nighttime light detection capability, this paper presents a convolution operation threshold method for offshore oil and gas platform extraction. Firstly, the VIIRS data of two different phases are convolutioned, the bright pixels are enhanced and the background pixels are weakened to distinguish the suspected target from the background clearly. Then the value of 0 is used as the cut-off point, Target extraction, and finally the relative static characteristics of oil and gas platforms, through the analysis of the neighborhood, to achieve the extraction of oil and gas platform. The results show that the convolution calculus proposed in this paper can effectively extract oil and gas platforms with an accuracy of about 85.4% and can effectively reduce the error caused by the empirical threshold on the extraction results.