海岸工程对海岸带经济发展和生态环境影响很大。随着海岸工程建设迅猛的发展,采用遥感的方法对海岸工程变化进行遥感监测显得尤为重要。本文以北京一号小卫星(BJ-1)资料为数据源,利用多种变化监测的方法对天津港和曹妃甸港区2006年和2010年的海岸工程变化进行监测。结果显示,波段替换法与SVM分类相结合的方法在2个重点研究区域精度最高,其总体精度和Kappa系数分别为92.35%和0.7902;面向对象的方法精度和稳定性其次,其总体精度和Kappa系数分别为91.77%和0.7732。

The coastal engineering exerts a great impact on the economic development and ecological environment of the seacoast. Thus, coastal engineering monitoring is a focus in coast zone remote sensing and monitoring. Since the 1980s, satellite remote sensing has become an indispensable technique in detecting the dynamic changes of coastal engineering. The accuracy of changes in coastal engineering is determined by the applicability of data obtained from remote sensing system and the feasibility of the methods in detecting the changes. As a satellite developed by China, the BJ-1 small satellite has already obtained numerous achievements in environment and disaster monitoring, urban management and construction and national land resource surveying. However, little has been investigated concerning the utilization of BJ-1 small satellite in monitoring the coast engineering. We compared various typical detection methods, and summarized a highly accurate and stable method in monitering the costal engineering with BJ-1 small satellite remote sensing data. Different detection methods were applied to investigate the changes in 2 key areas-Tianjin Port and Caofeidian Port costal engineering from 2006 to 2010 based on the characteristics of BJ-1 small satellite data, and evaluated the detected results. Our findings showed that among the detecting methods with BJ-1 remote sensing data, the maximum precision was obtained when waveband substitution technique was combined with SVM classification to detect the changes in costal engineering. The second precisest and stablest method was object-oriented analysis. These results indicate that, BJ-1 remote sensing data meet the requirements for accuracy in different costal engineering monitoring. Meanwhile, combination of waveband substitution and SVM classification technique, as well as object-oriented analysis, has the highest accuracy and stability in different costal engineering monitoring.