基于图像配准的水力发电站摄像机预置位校正方法

Image Alignment Based Camera Preset Correction Method for Hydroelectric Power Stations

随着可视化技术的发展,水力发电站部署了大量的云台、球型摄像机,由于水力发电站摄像机的安装环境复杂,经常受到水流、震动等因素的干扰,导致预置位校正不准确。因此,本文提出了一种基于图像配准的校正方法,该方法基于D2-Net的图像匹配算法,可根据摄像机拍摄的图像自动校正预置位,并提高预置位的准确度。该方法首先将两幅待匹配图像输入D2-Net网络,分别输出两幅图像的关键点与其对应的特征描述,然后根据特征匹配与映射矩阵计算算法,得到两幅待匹配图像的映射关系。通过图像配准技术将当前拍摄的图像与参考图像进行对比,然后根据对比结果计算校正参数,最后将校正参数应用到预置位中,实现摄像机的自动校正。实验结果表明,该方法能有效提高水力发电站摄像机预置位的校正准确度,具有很高的实用价值。

With the development of visualisation technology,a large number of pan-tilt and dome cameras are deployed in hydroelectric power stations.Due to the complex installation environment of hydroelectric power station cameras,they are often disturbed by factors such as water flow and vibration,resulting in inaccurate preset position correction.Therefore,this paper proposes a correction method based on image alignment,which is based on D2-Net's image matching algorithm and can automatically correct the preset position and improve the accuracy of the preset position based on the images captured by the cameras.The method first inputs two images to be matched into the D2-Net network,outputs the key points of the two images with their corresponding feature descriptions,and then obtains the mapping relationship between the two images to be matched according to the feature matching and mapping matrix calculation algorithm.The image alignment technique compares the currently captured image with the reference image,then calculates the correction parameters based on the comparison results,and finally applies the correction parameters to the preset bit to realise the automatic camera correction.The experimental results show that the method can effectively improve the calibration accuracy of the camera presets in hydroelectric power stations and has high practical value.