| 基于鱼类运动行为的水质异常评价因子研究 |
| |
| 引用本文： | 程淑红,刘洁,李雷华. 基于鱼类运动行为的水质异常评价因子研究[J]. 仪器仪表学报, 2015, 36(8): 1759-1766 |
| |
| 作者姓名： | 程淑红 刘洁 李雷华 |
| |
| 作者单位： | 燕山大学电气工程学院 河北省测试计量技术及仪器重点实验室秦皇岛066004 |
| |
| 基金项目： | 河北省自然科学青年基金项目(F2012203031)、河北省专家出国培训项目（2012002）资助 |
| |
| 摘 要： | 针对利用鱼类运动行为进行异常水质评价的问题,本文对如何选取具有普遍性、全面性、代表性等特点且能够反映水环境与鱼类运动特征参数之间的主要相关性的因子进行了研究。首先通过研究水质变化对红鲫鱼运动轨迹的影响,引入基于鱼类运动轨迹的曲率、邻近特征这两个特征参数;其次分别提取正常和异常水质中红鲫鱼的速度、加速度、曲率、邻近特征4个特征参数,建立水质异常评价因子数据库;最后将不同评价因子输入支持向量机(SVM)进行分类识别,且采用ROC曲线考察鱼类运动特征参数和水质安全定性关系的准确性。实验结果表明,曲率、邻近特征作为水质评价因子明显优于速度、加速度,不但异常水质识别率达到90%以上,而且通过ROC曲线可知其建立的模型具有稳健性、普适性、高效性。
|
| 关 键 词： | 鱼类运动行为；水质异常评价因子；支持向量机；ROC曲线 |
Study on anomaly water quality assessment factorbased on fish movement behavior |
| |
| Cheng Shuhong,Liu Jie,Li Leihua. Study on anomaly water quality assessment factorbased on fish movement behavior[J]. Chinese Journal of Scientific Instrument, 2015, 36(8): 1759-1766 |
| |
| Authors: | Cheng Shuhong Liu Jie Li Leihua |
| |
| Affiliation: | Institute of Electrical Engineering, Yanshan University, Measurement Technologyand Instrumentation Key Laboratory of Hebei Province, Qinhuangdao 066004, China |
| |
| Abstract: | Aiming at the problem of using fish movement behavior for anomaly water quality monitoring, this paper studies how to select the general, comprehensive and representative factors that can reflect the main relationship between water environment and fish movement behavior. Firstly, through studying the effect of water quality change on the movement trajectory of red crucian carp, two feature parameters, the curvature and vicinity feature calculated based on the fish movement trajectory are introduced. Secondly, four feature parameters, the velocity, acceleration, curvature and vicinity feature of the red crucian carp are extracted in normal water quality and the water with dichlorvos added, respectively; then the abnormal water quality evaluation index database is established for water quality detection. Finally, different evaluation indexes are fed to the support vector machine to conduct classification identification, and the ROC curve is adopted to investigate the accuracy of the relationship between water environment safety and fish movement feature parameters. The comparison experiment results show that curvature and vicinity feature as water quality assessment factors are obviously better than velocity and acceleration; the abnormal water quality recognition rate reaches above 90%. Moreover, from the ROC curve, it is known that the established models based curvature and vicinity feature are steady, universal and effective. |
| |
| Keywords: | fish movement behavior anomaly water quality assessment factor SVM ROC curve |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《仪器仪表学报》浏览原始摘要信息 |
|
点击此处可从《仪器仪表学报》下载免费的PDF全文 |
|
