Vineyard Yield Prediction Dataset

The current agricultural industry faces challenges such as climate change, pests, and difficulty in yield prediction. Traditional yield prediction methods rely on human experience, resulting in poor accuracy and inefficiency. Existing solutions often lack comprehensive analysis of field environment and visual data, leading to inaccurate predictions. This dataset aims to help AI models provide more accurate yield predictions by integrating vineyard visual data and field environment data. The dataset includes vineyard images captured by high-resolution cameras, with environments including different weather conditions and growth stages. We employ quality control measures like multi-round annotation and consistency checks to ensure annotation accuracy and consistency. Data is stored in JPEG format, with a clear organizational structure for easy and quick retrieval and usage. The core advantage of this dataset is its high annotation accuracy and good consistency, with annotation accuracy reaching 95%. We have introduced new data augmentation techniques to improve model generalization, which is expected to increase yield prediction accuracy by at least 15%.