基于遗传算法改进 BP 神经网络的遥感影像分类研究

遥感影像分类是遥感信息提取的重要手段，是目前遥感技术中的热点研究内容，有效地选择合适的分类方法是提高遥感影像分类精度的关键。BP神经网络具有收敛快和自学习、自适应性强的特点。在遥感图像分类中，BP神经网络能充分利用样本集的信息，自动建立分类模型，但由于BP神经网络的权值和阀值能直接影响BP神经网络模型的分类精度，因此该研究通过遗传算法来确定BP神经网络的最优权值和阀值，从而提高BP神经网络的分类精度。以LandsatTM遥感图像作为数据源，以长江中游一武汉市为研究地区，建立了基于BP神经网络模型的遥感分类模型和基于遗传算法改进BP神经网络模型的分类模型，对分类结果进行了定量分析。结果表明：在样本相同的情况下，基于遗传算法改进BP神经网络的遥感影像分类精度要高于BP神经网络的遥感影像分类精度。

Remote Sensing Image Classification of Improved BP Neural Network Based on the Genetic Algorithm

Remote sensing image classification is an important means of remote sensing information extraction, content of research is the hotspot in the remote sensing technology, effectively selecting the appropriate classification method is the key to improve the precision of remote sensing image classification. The BP neural network has quick convergence and self learning, the characteristics of strong adaptability. In remote sensing image classification, the BP neural network can make full use of the information of sample set, automatic classification model is set up, but as a result of the BP neural network weights and thresholds can directly influence the classification precision of BP neural network model, so the opti- mal BP neural network weights and thresholds were determined by the genetic algorithm, so as to improve the classification precision of BP neural network. With Landsat TIM remote sensing image as data sources, the middle reach of Yangtze River in Wuhan City as research area, the classifi- cation model of the remote sensing classification model based on BP neural network model and improved BP neural network model based on genet- ic algorithm was established, the quantitative analysis of the classification result was conducted. The results showed that under the condition of the same sample, the remote sensing image classification accuracy of improved BP neural network based on genetic algorithm is higher than that of BP neural network remote sensing image classification precision.