Unmanned Weighing System for Catfish Based on Machine Vision and Multi-sensor Collaboration

XIAO Maohua, JI Shuying, ZHU Hong, LI Dongfang, WANG Bingqing

Abstract

The scale of China’s aquaculture industry continues to expand, but the traditional method of weighing catfish exists in low efficiency, large error, serious fish stress damage and other problems. A catfish-unmanned weighing system was proposed based on machine vision and multi-sensor synergy, fusing target detection algorithms and multi-source information measurement technology. Through the improved YOLO 11 model (HY-YOLO 11), deformable convolution (DCNv3) and spatial enhancement attention module (SEAM) were introduced to effectively solve the problem of fish sticking together, cage deformation and light fluctuation interference in dynamic scenes. The system integrated a floating monitoring platform, an industrial camera and a tensile force sensor, combined with an adaptive lifting mechanism and a buoyancy compensation mechanism, to realize the simultaneous and accurate measurement of the total mass and quantity of the fish. The experimental results showed that the average detection accuracy (mAP50) of the HY-YOLO 11 model reached 94.8% in the dense occlusion scenario, which was 3.7 percentage points higher than the baseline model, the average absolute error (MAE) of fish counting was 0.56, the relative error of mean weight calculation was less than 4%, and average time for a single weighing session was 58s. The system provides an efficient and reliable technical solution for the intelligent management of aquaculture.

 

Keywords: catfish, machine vision, contactless weighing system, YOLO 11, DCNv3, SEAM

 

Download Full Text:

PDF


References


MU Yong, SUN Yanhui. Pond health cultivation techniques of channel catfish [J] . Fishery Wealth Guide, 2025(7) ; 24 -27. (in Chinese)

WANG Zhiyong, FENG Shuqing, CHEN Zhixin, et al. Design and application of weighing system for live fish short-distance transfer [J ] . Fishery Modernization, 2019, 46(3): 30 -34. (in Chinese)

ZHENG Yapeng, ZHANG Lu, LIU Zunxu. Fish mass estimation method based on NGBoost under binocular vision [ J]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55(Supp.2) ; 294 -302. (in Chinese)

HU Jinyou, WANG Jingjie, ZHANG Xiaoshuan, et al. Research status and trends of key technologies for aquaculture informatization[ J]. Transactions of the Chinese Society for Agricultural Machinery, 2015, 46(7) ; 251 -263. (in Chinese)

ZHU Ming, ZHANG Zhenfu, HUANG Huang, et al. Research progress on intelligent feeding methods in fish farming [J ]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(7) ; 38 -46. (in Chinese)

LIU Shijing, LI Guodong, LIU Huang, et al. Development status of aquaculture equipment in China [J] . Journal of Fisheries of China, 2023, 47( 11): 190 -203. (in Chinese)

LI Yanjun, HUANG Kangwei, XIANG Ji. Dynamic fish body size measurement based on stereo vision[ J] . Transactions of the CSAE, 2020, 36(21 ): 220 -226. (in Chinese)

PETRELL R J, SHI X,WARD R K,et al. Determining fish size and swimming speed in cages and tanks using simple video techniques [J ]. Aquacultural Engineering, 1997 ,16( 1 -2) ;63 -84.

ZHANG T, YANG Y, LIU Y, et al. Fully automatic system for fish biomass estimation based on deep neural network [ J]. Ecological Informatics, 2024, 79; 102399.

МА Junxiao, GONG Ze, KANG Jiaming, et al. Quantitative sectioning method of fish based on line laser scanning [J ]. Food & Machinery, 2022, 38( 10) ; 87 -92. (in Chinese)

LI Guangmei, WEI Xinhua, LI Fade, et al. Design and implementation of weighing module for integrated fruit sorting machine [J] . Transactions of the CSAE, 2009, 25(2) : 96 - 100. (in Chinese)

ZHOU Xiaoyan, N1 Qi, XL Hao, et al. Report on the development of full-process mechanization in Chinese aquaculture in 2021 i [R]. Beijing: Chinese Academy of Fishery Sciences, 2021. (in Chinese)

HAO Y, YIN H, LI D. A novel method of fish tail fin removal for mass estimation using computer vision [J]. Computers and Electronics in Agriculture, 2022, 193:106601.

ABINAYA N S, SUSAN D, SIDHARTHAN R K. Deep learning-based segmental analysis of fish for biomass estimation in an occulted environment[ J ]. Computers and Electronics in Agriculture, 2022, 197:106985.

KHANAM R, HUSSAUN M. Yolovll; an overview of the key architectural enhancements[ J ]. arXiv Preprint, arXiv;2410. 17725, 2024.

WANG Jianyong. Pond culture technology of Litopenaeus vannamei in northern China[ J]. Scientific Fish Farming, 2020(8) ; 32 -33. (in Chinese)

KEYS R G. Cubic convolution interpolation for digital image processing[ J ] IEEE Transactions on Acoustics, Speech, and Signal Processing, 1981 , 29(6); 1153 -1160.

REDMON J, FARHADI A. YOLO4; optimal speed and accuracy of object detection [ J]. arXiv Preprint, arXiv:2004. 10934, 2020.

SZELISKI R. Computer vision: algorithms and applications M ]. London: Springer, 2022: 157 -162.

YU Z, HUANG H, CHEN W, et al. Yolo-facev2: a scale and occlusion aware face detector [J] . Pattern Recognition, 2024, 155: 110714.

LI IH, ZhANG Y, ZhANG Y, et al. DCNv3: towards next generation deep cross network for CTR prediction [J]. arXiv Preprint, arXiv:2407. 13349, 2024.

DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16 x 16 words; transformers for image recognition at scale [j]. arXiv Preprint, arXiv;2010. 11929, 2020.


Refbacks

  • There are currently no refbacks.