Abstract: Algae cell detection is of great significance for ecosystem monitoring, and currently algal cell detection still relies on manual labor, which is not only time-consuming and laborious, but also consumes a large amount of cost. In this study, computer-aided detection is combined with the Transformer target detection model and used for the microalgae cell detection task. To address the problem of blurred microalgae cell features, we design a feature-enhanced convolution module incorporating RFA and CA, and select three different backbones, namely, ResNet50, ResNet34, and ResNet18, for the experiments. The experimental results show that the improved model obtains the highest accuracy of 87.6%, 73.3%, 85.7%, 95.2%, 97.7% and 99.0% for six microalgae, namely, Platymonas, Chlorella, Dunaliella, Effrenium, Porphyridium, and Haematococcus in Kaggle’s publicly available microalgae image dataset, respectively, which proves the feasibility of the Transformer model in the field of microalgae target detection and the scientificity, providing new ideas and methods for artificial intelligence in the field of ecological research and microalgae detection.