Academic Journal of Engineering and Technology Science, 2026, 9(2); doi: 10.25236/AJETS.2026.090204.
Xijun Gao1, Dairan Li1, Yutong Li1, Jiahong Yin2
1School of Railway Intelligent Engineering, Dalian Jiaotong University, Dalian, China
2School of Materials Science and Engineering, Dalian Jiaotong University, Dalian, China
As an important component of aquatic ecosystems, the population distribution and activity status of microalgae are key indicators for water quality monitoring and ecological research. However, traditional microalgae detection methods mainly rely on manual microscopic counting or flow cytometry, which suffer from limitations such as expensive equipment, cumbersome operations, and excessive dependence on laboratory environments and professional personnel, making it difficult to meet the needs of real-time on-site monitoring in the field. Addressing the aforementioned issues, this study proposes a low-cost, portable, automated microalgae detection system featuring software-hardware synergy. In terms of hardware, a visible light and fluorescence dual-mode microscopic imaging device based on 3D printing technology was designed, achieving synchronized acquisition of microalgae morphology and activity as well as device miniaturization. In terms of algorithms, an improved GIFF-AlgaeDet deep learning object detection network is proposed. By introducing the CAGS attention mechanism, a lightweight re-parameterized dual detection head, and the SIoU loss function, it effectively solves the problems of missed detection and false detection caused by the minute size of microalgae targets and complex backgrounds, while significantly reducing model parameters. Experimental results show that the average detection precision of the system on a self-built microalgae dataset reaches 96.2%, with a single-frame inference time of only 0.02 seconds. It is significantly superior to current mainstream detection algorithms and possesses the advantages of high precision, low cost, and portability, providing an efficient and intelligent solution for water environment ecological monitoring and aquaculture management.
Microalgae detection, Deep learning, Portable microscopic imaging, Attention mechanism, Lightweight network
Xijun Gao, Dairan Li, Yutong Li, Jiahong Yin. Research on Real-Time Microalgae Detection Based on Deep Learning and a Portable Dual-Mode Imaging System. Academic Journal of Engineering and Technology Science (2026), Vol. 9, Issue 2: 27-36. https://doi.org/10.25236/AJETS.2026.090204.
[1] L. Junfei, C. Ying, S. Yu, and C. Ting, “An approach to microalgae identification based on joint optimization model of slicing and detection,” Expert Systems with Applications, vol. 280, no. Compendex, 2025, doi: 10.1016/j.eswa.2025.127565.
[2] J. Liu, Y. Chen, C. Wang, W. Li, and D. Zhu, “Motif field combined with two-stream feature fusion network and double detection head for identification and prediction of microalgae in seawater,” Water Research, vol. 288, no. Compendex, 2026, doi: 10.1016/j.watres.2025.124574.
[3] J. Scarfe, D. Kosmutzky, and R. E. R. Nisbet, “A game of tag: A review of protein tags for the successful detection, purification and fluorescence labelling of proteins expressed in microalgae,” Plant Journal, vol. 122, no. Compendex, 2025, doi: 10.1111/tpj.70272.
[4] J. Borisova, I. V. Morshchinin, V. I. Nazarova, N. Molodkina, and N. O. Nikitin, “Low-Cost Microalgae Cell Concentration Estimation in Hydrochemistry Applications Using Computer Vision,” Sensors, vol. 25, no. Compendex, 2025, doi: 10.3390/s25154651.
[5] Zhou S , Chen T , Fu E S ,et al. A microfluidic microalgae detection system for cellular physiological response based on an object detection algorithm. Lab on a Chip, 2024, 24(10):12. DOI:10.1039/D3LC00941F.
[6] T. Eerola et al., “Survey of automatic plankton image recognition: challenges, existing solutions and future perspectives,” Artificial Intelligence Review, vol. 57, no. Compendex, 2024, doi: 10.1007/s10462-024-10745-y.
[7] J. Luo, “Principal Component Analysis–Based Feature Recognition Technology for Marine Plankton Images in IoT Systems,” Journal of Testing and Evaluation, vol. 52, no. Compendex, 2024, doi: 10.1520/JTE20230079.
[8] Q. Ye, Y. Shi, and S. Guo, “CAGs-Net: A Novel Adjacent-Context Network With Channel Attention Gate for 3D Brain Tumor Image Segmentation,” International Journal of Biomedical Imaging, vol. 2025, no. Compendex, 2025, doi: 10.1155/ijbi/6656059.
[9] G. Li et al., “LR-Net: Lossless Feature Fusion and Revised SIoU for Small Object Detection,” Computers, Materials and Continua, vol. 85, no. Compendex, pp. 3267–3288, 2025, doi: 10.32604/cmc.2025.067763.
[10] F. Ren, J. Fei, H. Li, and B. T. Doma, “Steel Surface Defect Detection Using Improved Deep Learning Algorithm: ECA-SimSPPF-SIoU-Yolov5,” IEEE Access, vol. 12, no. Compendex, pp. 32545–32553, 2024, doi: 10.1109/ACCESS.2024.3371584.