Sensing of Objects in Water by Computer Vision Techniques

在本研究中，我们提出了三种测量水中物体三维形状的新方法：1)用激光测距仪测量水箱内物体的三维形状；2)用带机械手的激光测距仪测量水中物体的三维形状；3)用立体摄像系统对水下环境中气泡噪声的鲁棒观测。针对1)，我们提出了用射灯和狭缝光分别测量液体中物体形状的三维测量方法。将视觉传感器应用于液体中物体的测量时，会遇到图像失真的问题。它是由光在空气和液体之间的边界上的折射引起的，畸变的图像给测量距离的三角测量带来误差。该方法可以精确测量液体中物体表面的三维坐标，用于计算折射效应。对于2)，我们构建了一个由6DOF机械手和激光测距仪组成的新型传感系统。在移动机械手的同时，可以测量水中物体的三维形状、位置和方向。对于3)，在水下环境下的观测遇到了场景有时会被气泡干扰的困难。它还有一个问题，当相机被放置在水密玻璃板后面的空气中时，空气和水的边界表面会发生折射，导致图像的几何畸变。因此，我们提出了一种水下环境观测的鲁棒观测方法，该方法通过图像处理去除气泡图像和考虑折射的三维测量来实现。实验结果验证了所提方法的有效性。

In this study, we proposed three new methods that can measure 3-D shapes of objects in water as follows:1)3-D Measurement of Objects in Water Tank with a Laser Range Finder, 2)3-D Measurement of Objects in Water with a Laser Range Finder Equipped with a Manipulator, and 3)Robust Observation in Underwater Environments against Bubble Noises with a Stereo Camera System.As to 1), we proposed 3-D measurement method of objects' shapes in liquid by using a spot light and a slit light, respectively. When applying vision sensors to measuring objects in liquid, we meet the problem of an image distortion. It is caused by the refraction of the light on the boundary between the air and the liquid, and the distorted image brings errors in a triangulation for the range measurement. Our proposed method can measure the accurate 3-D coordinates of object surfaces in liquid taken for calculating the refraction effect.As to 2), we constructed a new sensing system that consisted of 6DOF manipulator and a laser range finder. 3-D shapes, positions, and orientations of objects in water could be measured while moving the manipulator.As to 3), observation in underwater environments meets the difficulty that scenes are sometimes disturbed by air bubbles. It has also the problem that, when a camera is set in air behind a watertight glass plate, refraction occurs at the boundary surface between air and water, resulting in geometrical distortion of images. Therefore, we proposed a robust observation method for underwater environments observation, which is realized by image processing for bubble image removal and three-dimensional measurement where refraction is taken into account.The effectiveness of our proposed methods was verified through experimental results.