农民身份识别挑战赛公开数据

赛题背景
机器虽然被大量用到农业生产中，但人还是不可或缺的因素。通过农民身份识别，可以真实客观地记录农民的状态，为农场管理和农产品追溯提供真实的客观数据；较之直接存储视频，可以有效地降低存储空间；自动识别也比人工监管，大幅度提高效率，减少人工成本。
赛事任务
农民身份识别需要对农民进行分类，本次大赛提供了中国农业大学实验室视频制成的图像序列。参赛选手先对图像进行预处理，并制作样本，对图像中的农民进行识别。选手需要自行训练模型，并上传自己训练好的模型和权重。
赛题数据集
本次比赛为参赛选手提供了25名农民身份，每个身份包含10段视频制成的图像序列，选手需要对图像序列进行预处理，打标签，并对农民进行身份识别。

Competition Background:
Although machines are extensively deployed in agricultural production, humans remain an indispensable factor. Farmer identity recognition can objectively and truthfully record farmers' status, providing authentic and objective data for farm management and agricultural product traceability. Compared with directly storing video files, it can effectively reduce storage space. Moreover, automatic recognition significantly improves efficiency and reduces labor costs relative to manual supervision.

Competition Task:
Farmer identity recognition requires classifying farmers. This competition provides image sequences generated from videos made in the China Agricultural University laboratory. Participants need to preprocess the images, prepare samples, and identify farmers in the images. Participants must independently train their models, and upload the trained models and their weights.

Competition Dataset:
This competition provides 25 farmer identities for participants, with each identity containing 10 image sequences generated from videos. Participants need to preprocess the image sequences, label them, and perform farmer identity recognition.

搜集汇总

背景与挑战

背景概述

该数据集来自农民身份识别挑战赛，旨在通过图像序列识别农民身份，以支持农场管理和农产品追溯。数据包含25名农民的身份，每个身份提供10段视频制成的图像序列，参赛者需预处理图像、打标签并训练分类模型。