摘要
至今主要靠目视计数来掌握香鱼(Plecoglossus altivelis)的资源量,这种方法很耗劳动力且精度低。由华盛顿大学研发的声学摄像仪DIDSON(Dual-frequency IDentification SONar)可在浑浊不清的水中摄取接近光学视频画质的图像。其附带软件虽可对体长为几十厘米的大个体鱼计数,但对于体长6cm以下的小个体鱼,由于其回波信号弱,计数显得非常困难。本文提供了一种基于DIDSON的溯河洄游幼香鱼计数的新方法。首先从声学图像中消除静止的背景,提取出运动目标并除去噪音像素,然后用卡尔曼滤波器对剩下的目标进行跟踪,根据目标移动的特性把目标分为香鱼和垃圾2大类,在除去向下游移动的垃圾之后,可正确地计算向上游移动香鱼的数量。通过现场实测验证了上述声学计数方法的实用性和高效性。
In order to comprehend the abundance of ayu, direct visual counting has been mainly used so far which is not only labor-intensive, but also its counting accuracy is poor. Recently, an acoustic camera called DIDSON （Dual-frequency IDentification SONar） ,which is invented by Washington University ,can provide almost-video-quality images to identify objects even in turbid water. Its software＇s fish counting function is useful for counting big fishes which are several dozen centimeters large. As for the small fishes whose length are only 6 centimeters or less,it becomes technically difficult to count them because their echo image are too weak and small. In this paper we present a new method for counting small size juvenile ayu that migrates upriver using a DIDSON. We first subtract the stationary background from the acoustic image and remove noises. Then we track the remained targets by Kalman filter. According to the character of the moving targets,we sort out the targets into two groups--ayu and garbage. After removed the garbage going downstream, we reached the correct counting of the upstream fishes. The above acoustic counting method has been proved to be useful and effective by an experiment.
出处
《渔业现代化》
北大核心
2009年第2期29-33,共5页
Fishery Modernization
关键词
幼香鱼
鱼量计数
DIDSON
渔业资源调查
Plecoglossus altivelis
fish-counting
DIDSON
survey of fishery resources
