基于EBPNN模型的遥感图像变化检测研究

对不同时段获取的特定图像进行自动变化检测是遥感图像研究的主要问题;通过自适应中值滤波(AMF)去除遥感图像中的噪声,结合Tamura和Law掩模方法提取图像中的次级特征,并将研究区域划分为植被、水域和城区三类,利用增强型反向传播神经网络(EBPNN)对特征提取结果进行分类并实现不同时期遥感图像的变化检测;与现有的FFNN和CNN分类技术相比,利用EBPNN进行分类可以有效地检测出图像中的变化且具有更好的检测性能。

Remote Sensing Image Change Detection based on EBPNN Model

Automatic change detection of specific images acquired in different periods is the main problem of remote sensing image research. Adaptive median filter (AMF) is used to remove the noise in remote sensing image, and Tamura and law mask methods are used to extract the secondary features of the image. The study area is divided into vegetation, water area and urban area. The enhanced back propagation neural network (ebpnn) is used to classify the feature extraction results and realize the change detection of remote sensing images in different periods. Compared with the existing FFNN and CNN classification techniques, ebpnn can effectively detect the changes in the image and has better detection performance.