Development of image sonar simulator for underwater object recognition
Development of image sonar simulator for underwater object recognition
复制标题
水下目标识别图像声纳模拟器的研制
DOI：
10.1109/urai.2018.8441870
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
Son
中科院分区：
文献类型：
--
作者：
Jeonghwe Gu;Hangil Joe;Son
本文介绍了水下目标识别的两个初步基础研究：图像声纳的建模和模拟器的研制。分析了图像声纳的显示机理，提出了一种图像声纳模型。图像声纳向前发射超声波束，然后检测那些撞击物体并从物体表面反射的超声波束。超声波束被建模为一组射线，并且图像声纳被建模为在某个位置和方向上发射射线的对象。待观察物体的表面被建模为由许多小多边形组成，以便任何任意形状都可以由这些多边形表示。超声波束和物体之间的碰撞被描述为射线与多边形相交的点。在此基础上，研制了图像声纳模拟器。将模拟器生成的声纳图像与真实的声纳图像进行对比，验证了模型的有效性。
In this paper, two preliminary basic studies of underwater recognition are reported: modeling an image sonar and developing its simulator. The display mechanism of the image sonar is analyzed, and an image sonar model is proposed. The image sonar emits ultrasound beams in the forward direction and then detects those that strike objects and are reflected from their surfaces. The ultrasound beams are modeled as a set of rays, and the image sonar is modeled as an object emitting the rays at a certain position and orientation. The surface of the objects to be observed is modeled as consisting of many small polygons so that any arbitrary shape can be represented by these polygons. Collisions between the ultrasound beams and objects are described in terms of the points at which the rays intersect the polygons. On the basis of the model, a simulator for the image sonar is developed. The sonar images made by the simulator are compared with real sonar images to confirm the validity of the model.