北太平洋远东拟沙丁鱼年龄鉴定方法的构建

doi:10.13304/j.nykjdb.2022.0636

中国农业科技导报 ›› 2023, Vol. 25 ›› Issue (4): 225-233.DOI: 10.13304/j.nykjdb.2022.0636

• 海洋农业淡水渔业 • 上一篇

北太平洋远东拟沙丁鱼年龄鉴定方法的构建

杨超¹^,²(), 韩海斌¹^,², 韦波¹, 张衡¹(), 商宸¹^,², 苏冰³, 刘思源², 蒋沛雯⁴, 相德龙²

^1.中国水产科学研究院东海水产研究所，农业农村部远洋与极地渔业创新重点实验室，上海 200090
^2.上海海洋大学海洋科学学院，上海 201306
^3.大连海洋大学船舶与工程学院，辽宁大连 116000
^4.安徽师范大学生态与环境学院，安徽芜湖 241000

收稿日期:2022-08-02 接受日期:2022-10-12 出版日期:2023-04-01 发布日期:2023-06-26
通讯作者: 张衡
作者简介:杨超 E-mail：243353707@qq.com；
基金资助:
国家重点研发计划项目(2018YFC1406802);中央级公益性科研院所基本科研业务费专项(2021M06);上海市野外科学观测研究站开放基金资助项目(K202001);浙江省远洋渔业资源探捕项目(CTZB-2022080076)

Construction of Method for Age Identification of Sardinops sagax in the North Pacific Ocean

Chao YANG¹^,²(), Haibin HAN¹^,², Bo WEI¹, Heng ZHANG¹(), Chen SHANG¹^,², Bing SU³, Siyuan LIU², Peiwei JIANG⁴, Delong XIANG²

^1.Key Laboratory of East China Sea & Oceanic Fishery Resources Exploitation and Utilization，Ministry of Agriculture of China； East China Sea Fisheries Research Institute，Chinese Academy of Fishery Sciences，Shanghai 200090，China
^2.College of Marine Sciences，Shanghai Ocean University，Shanghai 201306，China
^3.College of Navigation and Ship Engineering，Dalian Ocean University，Liaoning Dalian 116023，China
^4.School of Ecology and Environment，Anhui Normal University，Anhui Wuhu 241000，China

Received:2022-08-02 Accepted:2022-10-12 Online:2023-04-01 Published:2023-06-26
Contact: Heng ZHANG

摘要/Abstract

摘要：

为使鱼类年龄鉴定工作更加准确高效，基于2020—2021年的远东拟沙丁鱼（Sardinops sagax）948尾样本测量的叉长、体质量与耳石质量等生物学数据，结合传统耳石年龄轮纹观察法构建远东拟沙丁鱼年龄鉴定新方法。首先将其矢耳石包埋、研磨和拍照，通过轮纹观察法初步判读其准确年龄，之后构建基于一元线性回归方程、多元线性回归方程以及深度学习模型共3种年龄鉴定方法，阐述了远东拟沙丁鱼年龄鉴定方法构建和准确度提升过程。通过分析单个生物学特征所拟合的一元线性回归方程，对比其R²得出，耳石重量与轮纹观察法所得年龄的相关系数最高。据耳石重量拟合出一元线性回归方程进行年龄鉴定，与多元回归线性方程与深度学习模型所得的年龄鉴定结果进行比较可知，一元回归线性方程、多元线性回归方程及深度学习模型的鉴定结果与轮纹观察法相比在无误差的标准下其准确率分别为43.6%、54.0%和71.6%。该方法将极大提高远东拟沙丁鱼年龄鉴定的准确率及效率，为鱼类年龄鉴定提供新的方法与思路。

关键词: 远东拟沙丁鱼, 耳石重量, 年龄鉴定, 深度学习

Abstract:

To make fish age identification more accurate and efficient， based on the biological data such as fork length， weight， and otolith weight measured by 948 samples of Sardinops sagax from 2020 to 2021， and combined with the traditional otolith age wheel observation method， a new method for age identification of Sardinops sagax was constructed. Firstly， the otolith was buried， ground， and photographed. Its accurate age was preliminarily interpreted by the wheel pattern observation method scenario. Three ways were newly built based on the univariate linear regression equation， the multiple linear regression equation， and the deep learning model. The construction and accuracy improvement process of the age identification method of Sardinops sagax were elaborated. By analyzing the R²of the univariate linear regression equation fitted by a single biological feature， the correlation coefficient between the weight of the otolith and the age observed by the rotting method was the highest. According to the age identification of the unary linear regression equation fitted by the weight of the otolith and compared with the age identification results obtained by the multiple regression linear equation and the deep learning model， it could be seen that the identification results of the univariate regression linear equation， the multiple linear regression equation， and the deep learning model were 43.6%， 54.0%， and 71.6% respectively under the standard of no error compared with the wheel pattern observation method. This method would greatly improve the accuracy and efficiency of the age identification of Sardinops sagax， and provide new methods and ideas for fish age identification.

Key words: Sardinops sagaxs, otolith weight, age identification, deep learning

中图分类号:

S931.1

杨超, 韩海斌, 韦波, 张衡, 商宸, 苏冰, 刘思源, 蒋沛雯, 相德龙. 北太平洋远东拟沙丁鱼年龄鉴定方法的构建[J]. 中国农业科技导报, 2023, 25(4): 225-233.

Chao YANG, Haibin HAN, Bo WEI, Heng ZHANG, Chen SHANG, Bing SU, Siyuan LIU, Peiwei JIANG, Delong XIANG. Construction of Method for Age Identification of Sardinops sagax in the North Pacific Ocean[J]. Journal of Agricultural Science and Technology, 2023, 25(4): 225-233.

图/表 8

表1 远东拟沙丁鱼叉长、体质量和耳石重量信息

Table 1 Information of fork length，weight and otolish weight of Sardinops sagax

生物学特征 Biological feature

体质量Weight/g

耳石质量Otolish weight/mg

5.6～144.6

0.20～3.35

43.7

1.37

图1 远东拟沙丁鱼耳石A：未经处理的耳石；B：经包埋过后耳石；C：包埋后研磨的耳石。图示为叉长165 mm、年龄为2+龄的远东拟沙丁鱼耳石；白色箭头所指位置为耳石年轮

Fig. 1 Otolith image of the Sardinops sagaxA：Untreated otoliths；B：Otolith after embedding and manual grinding；C：Otolith overground after embedding. The picture shows the otoliths of Sardinops sagax with fork length of 165 mm and age of 2+ ； the position indicated by the white arrow is the otoliths growth ring

图2 神经网络整体框架

Fig. 2 Whole framework of self-built neural network

图3 耳石轮纹观察法所得各年龄组成比例

Fig. 3 Age composition of otolith ring observation method

图4 远东拟沙丁鱼耳石质量、叉长和体质量对应年龄分布的箱形图

Fig. 4 Box diagram of otolith weight， fork length and weight to age distribution of Sardinops sagax

图5 实测轮纹观察法年龄与线性拟合回归鉴定的年龄

Fig. 5 Age identified by measured surface observation and linear fitting regression

图6 深度学习模型的精度、召回率、损失率和F1值与训练次数迭代图注：该图为无误差情况下的各项指标参数。

Fig. 6 Iteration diagram of accuracy、recall rate、loss rate、F1 value and training timesNote： The figure shows the parameters of each index without error.

表2 3种方法鉴定结果的准确率

Table 2 Prediction accuracy results of three methods

方法Method	准确率Accuracy/%
一元拟合方程Unary fitting equation	43.6
多元拟合方程 Multivariate fitting equation	54.0
深度神经网络 Deep neural network	71.6

参考文献 36

1	LLUCH-BELDA D, CRAWFORD R J M, KAWASAKI T, et al.. World-wide fluctuations of sardine and anchovy stocks: the regime problem [J]. South African J. Marine Sci., 1989, 8(1):195-205.
2	PAULY D, DAVID N. ELEFAN I, a basic program for the objective extraction of growth parameters from length-frequency data [J]. Meeresforschung, 1981, 28(4): 205-211.
3	COLLINS R A, SPRATT J D. Age determination of northern anchovies, Engraulis mordax, from otoliths [J]. Northern Anchovy, 1969: 1965-1968.
4	CASSELMAN J M. Analysis of hard tissue of pike Esox lucius L. with special reference to age and growth [C]//International Symposium on the Ageing of Fish. Reading (UK),1974.
5	丘古诺娃. 鱼类年龄和生长的研究方法[M]. 北京:科学出版社, 1956:1-98.
6	BOEHLERT G W. Using objective criteria and multiple regression models for age determination in fishes [J]. Fish. Bull. 1985, 83(2):103-117.
7	CAMPANA S E, ERLEND M. Accuracy and precision of age and hatch date estimates from otolith microstructure examination [J]. ICES J. Marine Sci., 1991,48 (3):303-316.
8	NEDREAAS. Age determination of Northeast Atlantic Sebastes species [J]. ICES J. Marine Sci., 1990, 47(2):208-230.
9	DONALD D B, BABALUK J A, CRAIG J F, et al.. Evaluation of the scale and operculum methods to determine age of adult goldeyes with special reference to a dominant year-class [J]. Trans. Am. Fish. Soc., 1992, 121(6):792-796.
10	YAREMKO M L. Age determination in Pacific sardine,Sardinops sagax [S]. NOAA, 1996.
11	PAWSON M G. Using otolith weight to age fish [J]. J. Fish Biol., 2010, 36(4):521-531.
12	EMMETT R L, BRODEUR R D, MILLER T W, et al.. Pacific sardine (Sardinops sagax) abundance, distribution, and ecological relationships in the Pacific Northwest [J]. California Cooperative Oceanic Fish. Investigations Rep., 2005, 46(16): 122-143.
13	TIAN S. Knowledge on the biological and fisheries aspects of the Japanese Sardine, Sardinops melanostictus (Schlegel, 1846) [J/OL]. J. Mar. Sci. Eng., 2021, 9(12):1403 [2022-07-01]..
14	GAULDIE R W. Function, form and time-keeping properties of fish otoliths [J]. Comparative Biochem. Physiol. Part A: Physiol., 1988, 91(2):395-402.
15	SANDEMAN E J. Age determination and growth-rate of redfish (Sebastes Sp.) from selected areas around Newfoundland [D]. Newfoundland: Memorial University, 1964.
16	徐聚臣, 段友健, 骆小年, 等. 达里湖瓦氏雅罗鱼耳石质量与年龄关系[J]. 水产科学, 2021, 40(2):279-284.
	XU J C, DUAN Y J, LUO X N, et al.. Relationship between otolith weight and age in Amur Ide Leuciscus waleckii in Dali lake [J]. Fish. Sci., 2021,40(2) :279-284.
17	陶志英, 肖俊, 阳刚, 等. 匙吻鲟的耳石重量与年龄关系[J]. 南昌大学学报(理科版), 2019, 43(3):280-283.
	TAO Z Y, XIAO J, YANG G,et al..The relationship between otolith-weight and age in Polyodon spathula [J]. J. Nanchang Univ. (Nat. Sci.), 2019, 43(3):280-283.
18	王旭歌, 何勇凤, 李昊成,等. 长湖短颌鲚耳石重量与年龄关系[J]. 水生态学杂志, 2016, 37(6): 84-88.
	WANG X G, HE Y F, LI H C,et al.. Relationship of sagittal otolith weight to the age of Coilia brachygnathus [J]. J. Hydroecol., 2016, 37(6): 84-88.
19	颜云榕, 侯刚, 贺雄军, 等. 基于矢耳石重量与年龄关系对北部湾二长棘犁齿鲷年龄鉴定的研究[J]. 台湾海峡, 2011, 30(1):63-70.
	YAN Y R, HOU G, HE X J, et al.. Age detemination based on the otolish weight of the Cardinal seabream, Evynn is cardinals,in the Beibu Gulf of the South China sea [J]. J. Oceanography Taiwan Strait, 2011, 30(1):63-70.
20	FU Z P, ZHANG Y N, HOU H Y. Survey of deep learning in face recognition [C]//IEEE International Conference on Orange Technologies, 2014.
21	YANG C, ZHOU B. An application of deep learning in remote sensing: automatic change detection in Urban area [J/OL]. J. Phys.: Conf. Ser., 2021, 1769(1):012 [2022-07-01]. .
22	殷名称. 鱼类生态学[M]. 北京:中国农业出版社, 1995:1-293.
	YIN M C. Fish Ecology [M]. Beijing:China Agriculture Press,1995:1-293.
23	MCFARLANE G, SCHWEIGERT J, HODES V, et al.. Preliminary study on the use of polished otoliths in the age determination of Pacific sardine (Sardinops sagax) in British Columbia waters [J]. California Cooperative Oceanic Fish. Investigations, 2010, 51: 162-168.
24	杨国亮, 许楠, 李放, 等. 关于非线性激活函数的深度学习分类方法研究[J]. 江西理工大学学报, 2018, 39(3):76-83 .
	YANG G L, XU N, LI F, et al.. Research on deep learning classification for nonlinear activation function [J]. J. Jiangxi Univ. Sci. Technol., 2018, 39(3):76-83.
25	GOUTTE C, GAUSSIER E. A probabilistic interpretation of precision, recall and F-score, with implication for evaluation [C]// Proceedings of the 27th European conference on Advances in Information Retrieval Research, 2005:952.
26	郭弘艺, 唐文乔. 长江口刀鲚矢耳石重量与年龄的关系及其在年龄鉴定中的作用[J]. 水产学报, 2006, 30(3):347-352.
	GUO H Y, TANG W Q. The relationship between sagittal otolith weight-age and its use in age determination in Coilia nasus from the estuary of Yangtze river [J]. J. Fish. China, 2006, 30(3):347-352.
27	ANDERSON J, MORISON A, RAY D. Age and growth of Murray Cod, Maccullochella peelii (Perciformes: Percichthyidae), in the lower Murray-Darling Basin, Australia, from thin-sectioned otoliths [J/OL]. Mar. Freshwater Res., 1992, 43(5):983 [2022-07-01]. .
28	FERREIRA B P. Age validation and estimation of growth rate of the coral trout, Plectropomus leopardus, (Lacepede 1802) from Lizard island, Northern Great Barrier Reef [J]. Fish. Bulletin, 1994, 92(1):46-57.
29	LABROPOULOU M, PAPACONSTANTINOU C. Comparison of otolith growth and somatic growth in two macrourid fishes [J]. Fish. Res., 2000, 46(1-3):177-188.
30	CARDINALE M, ARRHENIUS F, JOHNSSON B. Potential use of otolith weight for the determination of age-structure of Baltic cod (Gadus morhua) and plaice (Pleuronectes platessa) [J]. Fish. Res., 2000, 45(3):239-252.
31	LUCKHURST B E, DEAN J M, M.Age REICHERT, growth and reproduction of the lane snapper Lutjanus synagris (Pisces : Lutjanidae) at Bermuda [J]. Mar. Ecol. Progress, 2000, 203(1438):255-261.
32	POLITIKOS D V, PETASIS G, CHATZISPYROU A, et al.. Automating fish age estimation combining otolith images and deep learning: the role of multitask learning [J/OL]. Fish. Res., 2021, 242(1):106033 [2022-07-01]..
33	ORDOEZ A, EIKVIL L, SALBERG A B, et al.. Explaining decisions of deep neural networks used for fish age prediction [J/OL]. PLoS ONE, 2020, 15(6):e0235013 [2022-07-01]. .
34	BENZER S, GARABAGHI F H, BENZER R, et al.. Investigation of some machine learning algorithms in fish age classification [J/OL]. Fish. Res., 2022, 245: 106151 [2022-07-01]. .
35	TROADEC H, BENZINOU A, RODIN V, et al.. Use of deformable template for two-dimensional growth ring detection of otoliths by digital image processing: application to plaice (Pleuronectes platessa) otoliths [J]. Fish. Res., 2000, 46(1-3): 155-163.
36	FABLET R, JOSSE N L. Automated fish age estimation from otolith images using statistical learning [J]. Fish. Res., 2005, 72(2-3):279-290.

[1]	赵越, 卫勇, 单慧勇, 穆志民, 张健欣, 吴海云, 赵辉, 胡建龙. 基于深度学习的高分辨率麦穗图像检测方法[J]. 中国农业科技导报, 2022, 24(9): 96-105.
[2]	刘海涛, 韩鑫, 兰玉彬, 伊丽丽, 王宝聚, 崔立华. 基于YOLOv4网络的棉花顶芽精准识别方法[J]. 中国农业科技导报, 2022, 24(8): 99-108.
[3]	吴建伟, 黄杰, 熊晓菲, 高晗, 秦向阳. 基于AI的桃树病害智能识别方法研究与应用[J]. 中国农业科技导报, 2022, 24(5): 111-118.
[4]	高姻燕, 孙义, 李葆春. 基于无人机RGB影像估测田间小麦穗数[J]. 中国农业科技导报, 2022, 24(3): 103-110.
[5]	黄昊, 谢圣桥, 陈度, 王恒. 深度学习在苹果产业链中的应用与研究进展[J]. 中国农业科技导报, 2022, 24(10): 79-89.
[6]	周惠汝, 吴波明. 深度学习在作物病害图像识别方面应用的研究进展[J]. 中国农业科技导报, 2021, 23(5): 61-68.
[7]	吕纯阳, 刘升平, 郭秀明, 肖顺夫, 刘大众, 杨菲菲, 李路华. 基于SSD模型的巢门蜜蜂检测[J]. 中国农业科技导报, 2021, 23(5): 98-107.
[8]	许贝贝1,王文生1,2*,郭雷风1,陈桂鹏3. 基于非接触式的牛只身份识别研究进展与展望[J]. 中国农业科技导报, 2020, 22(7): 79-89.
[9]	董伟1,钱蓉1,张洁2*,张立平1,陈红波2,张萌1,朱静波1,卜英乔3. 基于深度学习的蔬菜鳞翅目害虫自动识别与检测计数[J]. 中国农业科技导报, 2019, 21(12): 76-84.

北太平洋远东拟沙丁鱼年龄鉴定方法的构建

Construction of Method for Age Identification of Sardinops sagax in the North Pacific Ocean

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 36

相关文章 9

编辑推荐

Metrics