智慧工业检测车辆损伤评估分割数据集3079张yolovoccoco标注方式数据集统计信息图像尺寸:896*896类别数量: 21 类训练集图像:2260; 验证集图像:630; 测试集图像:189总图像数: 3079类别分布:类别名称 | 包含该类别的图像数 | 总标注数damage-crack-裂纹损伤618, 930damage-dent-凹陷损伤1737, 2526damage-scratch-划痕损伤2108, 3573damage-tear-撕裂损伤60, 73part-back left door-左后门23, 23part-back right door-右后门31, 32part-bedside panel-床边板/侧围板4, 4part-fender-挡泥板/翼子板124, 131part-front bumper-前保险杠129, 134part-front left door-左前门28, 29part-front right door-右前门35, 35part-hood-引擎盖96, 98part-quarter panel-后侧围板77, 78part-rear bumper-后保险杠73, 74part-rear panel-后围板1, 1part-rocker panel-门槛板/侧裙62, 62part-roof-车顶23, 23part-roof pillar-车顶立柱55, 60part-tailgate-尾门7, 7part-trunk lid-行李箱盖41, 41part-wheel arc panel-轮拱板2, 2一、车辆损伤评估分割数据集详情表项目详细参数数据集名称车辆损伤评估实例分割数据集总图像数量3079张图像尺寸896×896标注格式YOLO、VOC、COCO类别总数21类训练集2260张验证集630张测试集189张damage-crack-裂纹损伤图像数618标注数930damage-dent-凹陷损伤图像数1737标注数2526damage-scratch-划痕损伤图像数2108标注数3573damage-tear-撕裂损伤图像数60标注数73part-back left door-左后门图像数23标注数23part-back right door-右后门图像数31标注数32part-bedside panel-侧围板图像数4标注数4part-fender-翼子板/挡泥板图像数124标注数131part-front bumper-前保险杠图像数129标注数134part-front left door-左前门图像数28标注数29part-front right door-右前门图像数35标注数35part-hood-引擎盖图像数96标注数98part-quarter panel-后侧围板图像数77标注数78part-rear bumper-后保险杠图像数73标注数74part-rear panel-后围板图像数1标注数1part-rocker panel-门槛板/侧裙图像数62标注数62part-roof-车顶图像数23标注数23part-roof pillar-车顶立柱图像数55标注数60part-tailgate-尾门图像数7标注数7part-trunk lid-行李箱盖图像数41标注数41part-wheel arc panel-轮拱板图像数2标注数2二、YOLOv11 实例分割训练代码1. 数据集配置文件car_damage_seg.yamlpath:./datasets/car_damagetrain:images/trainval:images/valtest:images/testnc:21names:0:damage-crack1:damage-dent2:damage-scratch3:damage-tear4:part-back left door5:part-back right door6:part-bedside panel7:part-fender8:part-front bumper9:part-front left door10:part-front right door11:part-hood12:part-quarter panel13:part-rear bumper14:part-rear panel15:part-rocker panel16:part-roof17:part-roof pillar18:part-tailgate19:part-trunk lid20:part-wheel arc panel2. 训练、验证、推理完整代码fromultralyticsimportYOLOimportwarnings warnings.filterwarnings(ignore)deftrain_car_damage():# 加载YOLOv11分割模型modelYOLO(yolo11n-seg.pt)model.train(datacar_damage_seg.yaml,epochs100,imgsz896,batch4,device0,workers2,patience15,projectruns/segment,namecar_damage_eval,saveTrue,valTrue)print(车辆损伤分割模型训练完成)defval_model():# 模型评估modelYOLO(runs/segment/car_damage_eval/weights/best.pt)metricsmodel.val(datacar_damage_seg.yaml,imgsz896)print(f检测mAP0.5:{metrics.box.map50:.4f})print(f分割mAP0.5:{metrics.seg.map50:.4f})defpredict_damage(source):# 图像/视频推理modelYOLO(runs/segment/car_damage_eval/weights/best.pt)resultsmodel.predict(sourcesource,imgsz896,conf0.4,saveTrue,show_masksTrue)returnresultsif__name____main__:train_car_damage()val_model()# predict_damage(test.jpg)三、应用场景车险定损保险公司车辆出险自动定损识别损伤类型、位置与面积提升理赔效率汽修行业维修门店自动检测车身损伤辅助制定维修方案与报价车辆质检车企、二手车市场外观缺陷筛查、车况智能评估智能驾后系统停车场、4S店摄像头自动记录车辆初始状态规避纠纷学术与实训实例分割、工业视觉、自动驾驶方向毕设、课程设计、算法研究算法竞赛车辆缺陷检测、实例分割类赛事标准基线数据集
智慧工业检测车辆损伤评估分割数据集，21类车身缺陷检测数据集、汽车划痕凹陷裂纹分割数据YOLOv11车辆实例分割数据集车险智能定损数据集、二手车车况检测数据集、车身零部件+损伤联合识别数据集工业视觉
发布时间：2026/7/20 10:11:42
智慧工业检测车辆损伤评估分割数据集3079张yolovoccoco标注方式数据集统计信息图像尺寸:896*896类别数量: 21 类训练集图像:2260; 验证集图像:630; 测试集图像:189总图像数: 3079类别分布:类别名称 | 包含该类别的图像数 | 总标注数damage-crack-裂纹损伤618, 930damage-dent-凹陷损伤1737, 2526damage-scratch-划痕损伤2108, 3573damage-tear-撕裂损伤60, 73part-back left door-左后门23, 23part-back right door-右后门31, 32part-bedside panel-床边板/侧围板4, 4part-fender-挡泥板/翼子板124, 131part-front bumper-前保险杠129, 134part-front left door-左前门28, 29part-front right door-右前门35, 35part-hood-引擎盖96, 98part-quarter panel-后侧围板77, 78part-rear bumper-后保险杠73, 74part-rear panel-后围板1, 1part-rocker panel-门槛板/侧裙62, 62part-roof-车顶23, 23part-roof pillar-车顶立柱55, 60part-tailgate-尾门7, 7part-trunk lid-行李箱盖41, 41part-wheel arc panel-轮拱板2, 2一、车辆损伤评估分割数据集详情表项目详细参数数据集名称车辆损伤评估实例分割数据集总图像数量3079张图像尺寸896×896标注格式YOLO、VOC、COCO类别总数21类训练集2260张验证集630张测试集189张damage-crack-裂纹损伤图像数618标注数930damage-dent-凹陷损伤图像数1737标注数2526damage-scratch-划痕损伤图像数2108标注数3573damage-tear-撕裂损伤图像数60标注数73part-back left door-左后门图像数23标注数23part-back right door-右后门图像数31标注数32part-bedside panel-侧围板图像数4标注数4part-fender-翼子板/挡泥板图像数124标注数131part-front bumper-前保险杠图像数129标注数134part-front left door-左前门图像数28标注数29part-front right door-右前门图像数35标注数35part-hood-引擎盖图像数96标注数98part-quarter panel-后侧围板图像数77标注数78part-rear bumper-后保险杠图像数73标注数74part-rear panel-后围板图像数1标注数1part-rocker panel-门槛板/侧裙图像数62标注数62part-roof-车顶图像数23标注数23part-roof pillar-车顶立柱图像数55标注数60part-tailgate-尾门图像数7标注数7part-trunk lid-行李箱盖图像数41标注数41part-wheel arc panel-轮拱板图像数2标注数2二、YOLOv11 实例分割训练代码1. 数据集配置文件car_damage_seg.yamlpath:./datasets/car_damagetrain:images/trainval:images/valtest:images/testnc:21names:0:damage-crack1:damage-dent2:damage-scratch3:damage-tear4:part-back left door5:part-back right door6:part-bedside panel7:part-fender8:part-front bumper9:part-front left door10:part-front right door11:part-hood12:part-quarter panel13:part-rear bumper14:part-rear panel15:part-rocker panel16:part-roof17:part-roof pillar18:part-tailgate19:part-trunk lid20:part-wheel arc panel2. 训练、验证、推理完整代码fromultralyticsimportYOLOimportwarnings warnings.filterwarnings(ignore)deftrain_car_damage():# 加载YOLOv11分割模型modelYOLO(yolo11n-seg.pt)model.train(datacar_damage_seg.yaml,epochs100,imgsz896,batch4,device0,workers2,patience15,projectruns/segment,namecar_damage_eval,saveTrue,valTrue)print(车辆损伤分割模型训练完成)defval_model():# 模型评估modelYOLO(runs/segment/car_damage_eval/weights/best.pt)metricsmodel.val(datacar_damage_seg.yaml,imgsz896)print(f检测mAP0.5:{metrics.box.map50:.4f})print(f分割mAP0.5:{metrics.seg.map50:.4f})defpredict_damage(source):# 图像/视频推理modelYOLO(runs/segment/car_damage_eval/weights/best.pt)resultsmodel.predict(sourcesource,imgsz896,conf0.4,saveTrue,show_masksTrue)returnresultsif__name____main__:train_car_damage()val_model()# predict_damage(test.jpg)三、应用场景车险定损保险公司车辆出险自动定损识别损伤类型、位置与面积提升理赔效率汽修行业维修门店自动检测车身损伤辅助制定维修方案与报价车辆质检车企、二手车市场外观缺陷筛查、车况智能评估智能驾后系统停车场、4S店摄像头自动记录车辆初始状态规避纠纷学术与实训实例分割、工业视觉、自动驾驶方向毕设、课程设计、算法研究算法竞赛车辆缺陷检测、实例分割类赛事标准基线数据集
