Soil Crack Pattern Recognition Dataset

Currently, the agricultural sector faces issues like drought and soil degradation, leading to decreased crop yields. Traditional soil monitoring methods often rely on manual observation, which is inefficient and error-prone, making it difficult to meet the needs of rapidly changing environments. Existing solutions lack standardized data support for detecting soil cracks. This dataset aims to address the research needs for soil mechanics and structural features under drought conditions by collecting and annotating various forms of soil cracks. Data collection involves shooting with high-resolution cameras in different soil environments to ensure diversity and representativeness of the images. We implemented multiple rounds of annotation and consistency checks, with experts invited for review, ensuring the accuracy and consistency of data annotations. The data is stored in JPG format, organized into images and their corresponding annotation information, facilitating subsequent machine learning model training.