| 基于机器视觉的养殖鱼摄食行为识别方法 |
| |
| 引用本文： | 李济泽,位威,张凯凯.基于机器视觉的养殖鱼摄食行为识别方法[J].福建工程学院学报,2022,0(4):378-382. |
| |
| 作者姓名： | 李济泽 位威 张凯凯 |
| |
| 作者单位： | 福建工程学院机械与汽车工程学院 |
| |
| 摘 要： | 利用机器视觉系统获取养殖鱼摄食图像纹理特征来识别养殖鱼群的摄食行为。 从养殖鱼摄食图像上直接提取能表征鱼群摄食行为的20 维特征，通过归一化、PCA 降维和支持向量机训练获得养殖鱼摄食行为识别模型，以实现对养殖鱼摄食行为的识别。 结果表明，提出方法的平均精确度为92.3%、假负率7.34%、假正率4.15%，为指导养殖鱼智能投饵提供了参考。
|
| 关 键 词： | 养殖鱼 摄食行为 机器视觉 图像纹理 支持向量机 |
Recognition method of fish feeding behavior based on machine vision |
| |
| LI Jize,WEI Wei,ZHANG Kaikai.Recognition method of fish feeding behavior based on machine vision[J].Journal of Fujian University of Technology,2022,0(4):378-382. |
| |
| Authors: | LI Jize WEI Wei ZHANG Kaikai |
| |
| Institution: | School of Mechanical and Automotive Engineering, Fujian University of Technology |
| |
| Abstract: | The texture features of the feeding images of farmed fish acquired by the machine vision system were used to identify the feeding behavior of farmed fish. Twenty-dimensional features that can characterize the fish feeding behavior were extracted from the fish feeding images. The fish feeding behavior recognition model was obtained from normalization, PCA dimensionality reduction and support vector machine training, so as to realize the identification of the fish feeding behavior. Experimental results show that the average precision of the proposed method is 92.3%, the false negative rate is 7.34%, and the false positive rate is 4.15%, which provides reference for instructing fish swarm intelligent feeding. |
| |
| Keywords: | farmed fish feeding behavior machine vision images texture support vector machine |
|
| 点击此处可从《福建工程学院学报》浏览原始摘要信息 |
| 点击此处可从《福建工程学院学报》下载免费的PDF全文 |
