Manual Weed Removal Behavior Recognition Dataset

The current agricultural sector faces problems of low efficiency in weed removal and high labor costs. Traditional weed removal methods require a large amount of manpower, and existing smart weed removal technologies are still in the early stages, unable to effectively identify and handle different types of weeds. To address these issues, this dataset aims to provide a high-quality manual weed removal behavior recognition dataset to assist researchers and engineers in developing more efficient smart agricultural solutions. High-resolution cameras are used for data collection in real agricultural environments, ensuring the representativeness of the collected images. To ensure data quality, multiple rounds of annotation and expert reviews were conducted to ensure the consistency and accuracy of the annotations. The data is stored in JPG format and organized in folders for easy processing and analysis.