玉米田间杂草数据集

该数据集是在新疆昌吉华兴农场玉米试验田采集，共有4130张图片、10521个边界框，用于玉米田真实场景杂草检测算法研究。反枝苋1304个实例、灰黎717个实例、龙葵290个实例、骆驼刺328个实例、马齿苋1390个实例、田旋花1227个实例、狗尾草891个实例、合被苋859个实例和玉米幼苗3515个实例。yolo类别标签如下：Amaranthus': 0, 'Chenopodium': 1, 'Solanum nigrum': 2, 'Alhagi sparsifolia': 3, 'Purslane': 4, 'Convolvulus': 5, 'Setaria': 6, 'Amaranthus polygonoides': 7, 'Corn': 8。现有的数据集多为单目标的图片，本次的数据集以多目标为主，弥补了杂草检测领域的数据集短缺的不足。This dataset was collected from a corn experimental field at Huaxing Farm, Changji, Xinjiang, China. It contains 4,130 images and 10,521 bounding boxes, and is designed for research on real-scenario weed detection algorithms in corn fields. Specifically, it includes 1,304 instances of Amaranthus, 717 instances of Chenopodium, 290 instances of Solanum nigrum, 328 instances of Alhagi sparsifolia, 1,390 instances of Purslane, 1,227 instances of Convolvulus, 891 instances of Setaria, 859 instances of Amaranthus polygonoides, and 3,515 instances of corn seedlings. The YOLO category labels are defined as follows: "Amaranthus": 0, "Chenopodium": 1, "Solanum nigrum": 2, "Alhagi sparsifolia": 3, "Purslane": 4, "Convolvulus": 5, "Setaria": 6, "Amaranthus polygonoides": 7, "Corn": 8. Most existing datasets primarily feature single-target images, while this dataset focuses on multi-target scenarios, which addresses the shortage of available datasets in the field of weed detection.

搜集汇总

背景与挑战

背景概述

该数据集包含3970张图片和6998个边界框，采集自新疆昌吉华兴农场玉米试验田，用于玉米田真实场景杂草检测算法研究。数据集涵盖了8种杂草类别，以多目标为主，弥补了杂草检测领域的数据集短缺。