无人机航拍图像拼接技术研究与实现

Research and implementation of UAV aerial image mosaic technology

图像拼接是从局部视图序列中获得更宽视野的过程,在过去的二十年中,相关专家学者们已经提出了许多图像拼接算法和技术。在无人机航拍中,受限于无人机稳定性差、像素低等局限,所获得的图像往往需要拼接降噪后方能进行实际应用。然而,传统的图像拼接直接应用于无人机航拍图像中,存在运算复杂、占用内存大、图像精准性较差等问题。因此,迫切需要适用于无人机航拍图像的拼接算法和技术。文章据此出发,首先分析无人机航拍图像拼接的必要和技术现状,指出一般的图像拼接需要准备和测量GCP的位置,且需要消耗较多的时间和精力。在此基础上,提出无人机航拍图像拼接的关键技术与实现,指出HARRIS算法更加适用于无人机航拍图像的拼接,并且详细论述了图像拼接的流程。

Image mosaic is a process to obtain a wider field of vision from the partial view sequence.In the past two decades, relevant experts and scholars have proposed many image mosaic algorithms and technologies. In UAV aerial photography, due to the limitations of poor stability and low pixels of UAV,the obtained images often need to be spliced and denoised before they can be applied in practice. However,traditional image stitching is directly applied to UAV aerial images, which has some problems, such as complex operation, large memory occupation, poor image accuracy and so on. Therefore, there is an urgent need for stitching algorithms and technologies suitable for UAV aerial images. Based on this, this paper first analyzes the necessity and technical status of UAV aerial image mosaic, and points out that the general image mosaic needs to prepare and measure the position of GCP, and needs to consume more time and energy. On this basis, this paper puts forward the key technology and implementation of UAV aerial image mosaic, points out that Harris algorithm is more suitable for UAV aerial image mosaic, and discusses the process of image mosaic in detail.