标注的平面图元素识别数据集识别率97.4%,可识别冰箱洗手间窗户餐桌门窗床等24种元素支持yolococo jsonpascal voc xml格式的模型训练模型训练指标参数模型训练图数据集拆分训练集8522图像验证集1206图像测试集597图像预处理自动定向应用调整大小拉伸到125*125增强每个训练示例的输出2噪点0.1%数据集标签Corner Sofa, Dinner Table, Dishwasher, Door Window, Double Bed, Double Door, Oven Stove, Refrigerator, Shower, Single Bed, Single Door, Sink, Sofa_1P, Sofa_2P, Sofa_3P, Stairs, Study Table, Television, Toilet, Wardrobe, Wash Basin, Washbasin Cabinet, Washing Machine, Window数据集图片和标注信息示例数据集下载yolo26:https://download.csdn.net/download/pbymw8iwm/92773553yolo v12:https://download.csdn.net/download/pbymw8iwm/92773557yolo v11:https://download.csdn.net/download/pbymw8iwm/92773555yolo v9: https://download.csdn.net/download/pbymw8iwm/92773559yolo v8:https://download.csdn.net/download/pbymw8iwm/92773562yolo v7:https://download.csdn.net/download/pbymw8iwm/92773554yolo v5:https://download.csdn.net/download/pbymw8iwm/92773550yolo darknet:https://download.csdn.net/download/pbymw8iwm/92773558coco json:https://download.csdn.net/download/pbymw8iwm/92773556pascal voc xml:https://download.csdn.net/download/pbymw8iwm/92773551YOLO模型训练下载数据集之后解压到当前文件夹然后将 我的仓库 https://gitcode.com/pbymw8iwm/YOLOProject里的训练模型脚本复制到文件夹下假设你使用的是yolov8来训练你就用 python train_yolov8.py注意请根据你的GPU能力来适当调整训练参数比如训练batchpatienceworkers以及模型类型如果你的GPU硬件条件限制可以联系作者进行付费模型训练部分模型只需要一杯奶茶钱模型验证测试情况验证测试代码#需要安装pip install ultralytics from ultralytics import YOLO import cv2 # 加载训练好的 YOLO .pt 模型 model YOLO(best.pt) # 替换为你实际的 .pt 模型文件路径 # 定义要测试的图片路径 image_path ./image.jpg # 替换为你实际的图片文件路径 # 使用模型对图片进行预测 results model(image_path) # 获取预测结果 for result in results: # 获取绘制了检测框的图片 annotated_image result.plot() # 显示图片 cv2.imshow(YOLOv Inference, annotated_image) # 等待按键退出 cv2.waitKey(0) # 关闭所有 OpenCV 窗口 cv2.destroyAllWindows()推理结果{predictions: [{x: 148.5,y: 149,width: 297,height: 296,confidence: 0.978,class: Toilet,class_id: 18,detection_id: e8a5bcfd-f99f-4224-9296-9f147e3fba89}]}推理结果{predictions: [{x: 148,y: 148,width: 296,height: 296,confidence: 0.983,class: Sink,class_id: 11,detection_id: 7eaf4b2e-e891-4af6-a30a-a183ef770e36}]} 
`)
