基于深度学习的鱼类特征点检测与体征识别方法

doi:10.13733/j.jcam.issn.2095-5553.2024.06.030

摘要/Abstract

摘要：

为确保鱼类养殖过程中生长状况实时监控及科学化养殖管理，需要实现高效化、自动化的鱼类体征识别。基于此，提出基于深度学习的关键特征点检测模型结合双目视觉的鱼类体征识别方法。基于预处理后的单目视觉数据集对融入金字塔分割注意力的高分辨率网络模型展开训练，获得鱼类关键特征点检测模型，在此基础上能够对双目视觉图像中各特征点进行快速检测识别与匹配，从而根据双目视觉系统内部参数计算各特征点真实坐标并计算获得对应体征参数。试验结果表明，建立的关键特征点检测模型对各特征点PCK值均大于0.85，识别得到的体征参数相对误差均小于10%，能够为鱼类体征快速准确识别提供支撑，有效助力鱼类养殖科学化、智能化发展。

关键词: 鱼类, 水产养殖, 深度学习, 关键点检测, 体征识别

Abstract:

In order to ensure realtime monitoring of growth conditions and scientific breeding management in the process of fish farming, it is necessary to realize efficient and automatic fish sign recognition. Based on this, a fish sign recognition method based on deep learning key feature point detection model combined with binocular vision is proposed. Based on the preprocessed monocular vision data set, the highresolution network model integrated into the pyramid segmentation attention is trained to obtain the fish key feature point detection model. On this basis, the binocular vision image can be rapidly detected, recognized and matched with each feature point, and the real coordinates of each feature point and corresponding physical parameters can be calculated according to the internal parameters of the binocular vision system The test results show that the PCK value of the established key feature point detection model for each feature point is greater than 0.85, and the relative error of the identified sign parameters is less than 10%, which can provide support for the rapid identification of fish signs and effectively help the scientific and intelligent development of fish farming.

Key words: fish, aquaculture, deep learning, key point detection, signs identification

中图分类号:

S24
S951.2

崔海朋, 秦朝旭, 马志宇. 基于深度学习的鱼类特征点检测与体征识别方法[J]. 中国农机化学报, 2024, 45(6): 201-207.

Cui Haipeng, Qin Chaoxu, Ma Zhiyu. Fish key feature point detection and sign identification based on deep learning[J]. Journal of Chinese Agricultural Mechanization, 2024, 45(6): 201-207.

参考文献

［1］　Zhao Z, Liu Y, Sun X, et al. Composited FishNet: Fish detection and species recognition from lowquality underwater videos ［J］. IEEE Transactions on Image Processing, 2021, 30: 4719-4734.

［2］　陈文辉, 蔡妹姝, 严松, 等. 基于深度卷积神经网络与迁移学习的鱼类分类识别［J］. 福建师大福清分校学报, 2019(5): 30-38.Chen Wenhui, Cai Meishu, Yan Song, et al. Fish classification based on deep convolutional neural network and transfer learning ［J］. Journal of Fuqing Branch of Fujian Normal University, 2019(5): 30-38.

［3］　蔡卫明, 庞海通, 张一涛, 等. 基于卷积神经网络的养殖鱼类品种识别模型［J］. 水产学报, 2022, 46(8): 1369-1376.

Cai Weiming, Pang Haitong, Zhang Yitao, et al. Recognition model of farmed fish species based on convolutional neural network ［J］. Journal of Fisheries of China, 2022, 46(8): 1369-1376.

［4］　Yu C, Fan X, Hu Z, et al. Segmentation and measurement scheme for fish morphological features based on Mask RCNN ［J］. Information Processing in Agriculture, 2020, 7(4): 523-534.

［5］　丁顺荣, 肖珂. 基于粒子群优化SVM和多特征融合的鱼类分类方法研究［J］. 中国农机化学报, 2020, 41(11): 113-118, 170.

Ding Shunrong, Xiao Ke. Improving fish classification method in particle swarm optimization SVM and multifeature fusion ［J］.Journal of Chinese Agricultural Mechanization, 2022, 41(11): 113-118, 170.

［6］　Mathias A, Dhanalakshmi S, Kumar R, et al. Underwater object detection based on Bidimensional empirical mode decomposition and Gaussian Mixture Model approach ［J］. Ecological Informatics, 2021, 66: 101469.

［7］　Zhang J, Chen Z, Tao D. Towards high performance human keypoint detection ［J］. International Journal of Computer Vision, 2021, 129(9): 2639-2662.

［8］　曾文献, 马月, 李伟光. 轻量化二维人体骨骼关键点检测算法综述［J］. 科学技术与工程, 2022, 22(16): 6377-6392.

Zeng Wenxian, Ma Yue, Li Weiguang. A survey of lightweight twodimensional human skeleton key point detection algorithms ［J］. Science Technology and Engineering, 2022, 22(16): 6377-6392.

［9］　赵建敏, 关晓鹏. 基于双目深度估计的牛体尺测量方法设计［J］. 光电子·激光, 2022, 33(4): 429-435.

Zhao Jianmin, Guan Xiaopeng. Design of the measurement method of cow body size based on binocular depth estimation ［J］. Journal of Optoelectronics Laser, 2022, 33(4): 429-435.

［10］　董鹏, 周烽, 赵悰悰, 等. 基于双目视觉的水下海参尺寸自动测量方法［J］. 计算机工程与应用, 2021, 57(8): 271-278.

Dong Peng, Zhou Feng, Zhao Congcong, et al. Automatic measurement of underwater sea cucumber size based on binocular vision ［J］. Computer Engineering and Applications, 2021, 57(8): 271-278.

［11］　郭卜瑜, 于佳, 王姣姣, 等. 双目视觉用于鱼苗尺寸测量［J］. 光学技术, 2017, 43(2): 153-157.Guo Buyu, Yu Jia, Wang Jiaojiao, et al. Stereo vision for measuring the size of fish ［J］. Optical Technique, 2017, 43(2): 153-157.

［12］　Shi C, Wang Q, He X, et al. An automatic method of fish length estimation using underwater stereo system based on LabVIEW ［J］. Computers and Electronics in Agriculture, 2020, 173: 105419.

［13］　李艳君, 黄康为, 项基. 基于立体视觉的动态鱼体尺寸测量［J］. 农业工程学报, 2020, 36(21): 220-226.

Li Yanjun, Huang Kangwei, Xiang Ji. Measurement of dynamic fish dimension based on stereoscopic vision ［J］. Transactions of the Chinese Society of Agricultural Engineering, 2020, 36(21): 220-226.

［14］　Wu X, Sahoo D, Hoi S C H. Recent advances in deep learning for object detection ［J］. Neurocomputing, 2020, 396: 39-64.

［15］　Lei W, Luo J, Hou F, et al. Underground cylindrical objects detection and diameter identification in GPR Bscans via the CNNLSTM framework ［J］. Electronics, 2020, 9(11): 1804.

［16］　Wang C, Pang C. Object detection method of power equipment based on mask RCNN ［J］. Academic Journal of Science and Technology, 2022, 1(2): 60-62.

［17］　Sun K, Xiao B, Liu D, et al. Deep highresolution representation learning for human pose estimation ［C］. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 5693-5703.

［18］　马皖宜，张德平. 基于多谱注意力高分辨率网络的人体姿态估计［J］. 计算机辅助设计与图形学学报, 2022, 34(8): 1283-1292.Ma Wanyi, Zhang Deping. Human pose estimation based on multispectral attention and high resolution network ［J］. Journal of ComputerAided Design & Computer Graphics, 2022, 34(8): 1283-1292.

［19］　Zhang H, Zu K, Lu J, et al. EPSANet: An efficient pyramid squeeze attention block on convolutional neural network ［C］. Proceedings of the Asian Conference on Computer Vision. 2022: 1161-1177.

［20］　罗桂娥. 双目立体视觉深度感知与三维重建若干问题研究［D］. 长沙: 中南大学, 2012.Luo Guie. Some issues of depth perception and three dimension reconstruction from binocular stereo vision ［D］. Changsha: Central South University, 2012.

［21］　周燕, 刘紫琴, 曾凡智, 等. 深度学习的二维人体姿态估计综述［J］. 计算机科学与探索, 2021, 15(4): 641-657.

Zhou Yan, Liu Ziqin, Zeng Fanzhi, et al. Survey on twodimensional human pose estimation of deep learning ［J］. Journal of Frontiers of Computer Science and Technology, 2021, 15(4): 641-657.

[1]	赵晴晴, 田云臣, 崔智明, 李建波, 马国宁, 隋金柱. 陆基工厂化水产养殖池自动清洗机器人设计与试验[J]. 中国农机化学报, 2024, 45(7): 68-74.
[2]	吴利刚, 吕媛媛, 周倩, 陈乐, 张梁, 史建华. 基于改进YOLOv5的轻量级黄花成熟检测方法[J]. 中国农机化学报, 2024, 45(7): 235-242.
[3]	张继成, 黄向党. 基于改进YOLOv3的玉米病害识别方法[J]. 中国农机化学报, 2024, 45(7): 269-275.
[4]	雷浩, 苑迎春, 何振学. 基于多尺度卷积和注意力机制的枣品种识别[J]. 中国农机化学报, 2024, 45(6): 135-141.
[5]	黄尧, 何敬, 付饶, 刘刚, , 林远杨. YOLOv5s-CBAM算法在福寿螺虫卵识别中的应用分析[J]. 中国农机化学报, 2024, 45(6): 223-228.
[6]	陶旭, , 余富强, , 蔡金金, 么炜, , 刘博, . 面向激光雷达点云数据的多结构树种识别[J]. 中国农机化学报, 2024, 45(5): 168-175.
[7]	漆海霞, , , 李承杰, 黄桂珍. 基于轻量化YOLOv4的死淘鸡目标检测算法[J]. 中国农机化学报, 2024, 45(5): 195-201.
[8]	罗鑫, 李加强, 何超. 基于改进YOLOv4的澳洲坚果视觉监测方法[J]. 中国农机化学报, 2024, 45(5): 217-222.
[9]	廖舒怀, , 王凯, 宋健, 解福祥, 王名声, , 龚中良. 基于移动机器人的树木识别与冠层信息测量[J]. 中国农机化学报, 2024, 45(5): 232-238.
[10]	王化佳, 顾寄南, 王梦妮, 夏子林. 基于深度学习的茶嫩芽分割与采摘点定位方法研究[J]. 中国农机化学报, 2024, 45(5): 246-252.
[11]	陈天兄, 范俊杰, 张日喜, 刘建龙, , 柏宗春, , 孙建国. 水产养殖投喂装备发展现状研究[J]. 中国农机化学报, 2024, 45(4): 45-50.
[12]	陈鲁威, , 曾锦, , 袁全春, , 夏烨, , 潘健, 吕晓兰, , . 基于改进DeepLabV3+的梨树冠层分割方法[J]. 中国农机化学报, 2024, 45(4): 155-161.
[13]	张慧蒙, 何超, 徐嘉雯, 罗鑫, 荣剑, 刘学渊. 基于SCGYOLOv5n的收获期澳洲坚果检测算法[J]. 中国农机化学报, 2024, 45(4): 214-221.
[14]	邱金凯, 许秀英, 康烨, 臧浩, 马锴, 郭志鹏. 基于语义分割的无人机图像西瓜果实提取研究[J]. 中国农机化学报, 2024, 45(3): 182-188.
[15]	邱畅, 田光兆, 赵嘉威, 谢尚杰, 郑奎. 基于改进YOLOv5的草莓病害识别[J]. 中国农机化学报, 2024, 45(3): 198-204.