传统方法在水下相机标定应用中的研究

针对加装保护壳体的水下相机光路较空气中复杂,传统成像模型及标定在水下是否满足工程要求的问题,在带畸变的小孔成像模型中利用传统方法进行了大气和水下的相机标定,验证了两种情况下标定偏差的存在;利用神经网络方法对水下相机标定进行了对比实验,结果显示传统方法在工程应用中的水下相机标定中仍具有优势。

Study of Traditional Method for Underwater Camera Calibration

The optical path of underwanter camera was complex. According to the problem that if the traditional calibration method satisfied the accuracy of underwater calibration, the paper calibrated the camera in pin-hole model with aberration. The result proved that there were differences belween the camera calibration in air and that underwater. It calibrated the underwater camera using method based on neural network. This result proved that the traditonal method still had advantages in engineering-oriented underwater camera calibraiton.