Oats are a worldly important crop for their richness in nutrition and economic importance. However, pest infestations are a threat to oats yields and cause significant economic losses. Effective crop management requires timely and accurate pest detection, but traditional methods are too slow and reliant on humans. Using the Pestopia dataset from Kaggle, and leveraged deep learning, this study intends to automate pest detection in oat crops. We evaluated the effectiveness of four deep learning models: VGG19, ResNet152, MobileNetV3, and EfficientNetV3. The best-performing model, MobileNetV3, was found to achieve the highest accuracy, while preserving lightweight architecture and quicker inference time, with the latter being particularly useful for real-time applications in precision agriculture. This research adds to pest detection automation that improves oat crop management and minimizes losses to contribute to more sustainable practices of agriculture. Overall, the findings pave the path for integrating AI-based solutions in precision agriculture to achieve high productivity and efficiency. The contribution of this research is to automate the detection of pests in oat crop management and minimize cereal losses. The findings also allow combining in precision agriculture AI-based solutions to improve productivity and efficiency.