无人机可见光遥感影像地物目标提取技术研究

Research on ground object extraction technology from visible light remote sensing images of unmanned aerial vehicles

摘要:

无人机可见光遥感影像中地物目标边界清晰度较低,容易导致地物目标与背景之间的区分度降低,进而难以提取地物目标。为此,提出无人机可见光遥感影像地物目标提取方法。从光谱特征、纹理特征和边缘特征三个方面分析无人机可见光遥感影像特征。结合三种影像特征对无人机可见光遥感影像数据集实行增广处理。对完成增广后的数据集定义影像编码标签,以此确定地物目标增强权重,通过参量化处理地物目标光谱特征,计算光谱吸收指数,获取地物目标提取表达式,从而实现无人机可见光遥感影像地物目标提取。实验结果表明,所提方法能够保证地物目标边界的清晰度,具有较强的地物目标提取能力。

Abstract:

The low definition of ground object boundaries in unmanned aerial vehicle visible light remote sensing images can easily lead to a decrease in the differentiation between ground object targets and the background, making it difficult to extract ground object targets. To this end, a method for extracting ground objects from unmanned aerial vehicle visible light remote sensing images is proposed. Analyze the features of unmanned aerial vehicle visible light remote sensing images from three aspects: spectral features, texture features, and edge features. Augmenting the unmanned aerial vehicle visible light remote sensing image dataset by combining three image features. Define image encoding labels for the expanded dataset to determine the weight of ground object enhancement. By parameterizing the spectral features of ground object targets, calculate the spectral absorption index, and obtain the expression for ground object extraction, the unmanned aerial vehicle visible light remote sensing image ground object extraction can be achieved. The experimental results show that the proposed method can ensure the clarity of the boundary of ground objects and has strong ability to extract ground objects.