环境搭建 pipinstall opencv-python numpy目录结构 chkimg.py models/ ├── efficientdet_d2/# EfficientDet-D2 模型文件 │ ├── saved_model.pb# 模型文件 │ ├── variables/# 模型变量 │ └── assets/# 模型资源 ├── labels/# 标签文件 │ └── coco_labels.txt# COCO数据集标签 └── README.md# 本文件 模型下载说明 EfficientDet-D2 模型 从 TensorFlow Hub 下载：https://tfhub.dev/tensorflow/efficientdet/d2/1 下载完成后，解压到models/efficientdet_d2/目录。 COCO 标签文件 从 GitHub 下载：https://raw.githubusercontent.com/tensorflow/models/master/research/object_detection/data/mscoco_label_map.pbtxt 下载完成后，保存为models/labels/coco_labels.txt。 代码 #!/usr/bin/env python3 # -*- coding: utf-8 -*- """ 物体检测脚本 - 完全本地化版本 使用OpenCV的DNN模块和预训练模型，无需TensorFlow Hub """ import osimport sysimport cv2import numpyas npimport argparseclass LocalObjectDetector : def __init__ ( self) : self. load_models( ) def load_models ( self) : """加载本地物体检测模型""" try : print ( "正在加载本地物体检测模型..." ) # 检查本地模型文件 model_dir= os. path. join( os. path. dirname( __file__) , "models" ) dnn_model_path= os. path. join( model_dir, "yolov3.weights" ) dnn_config_path= os. path. join( model_dir, "yolov3.cfg" ) labels_path= os. path. join( model_dir, "labels" , "coco_labels.txt" ) # 如果本地模型不存在，使用OpenCV内置的DNN模型 if os. path. exists( dnn_model_path) and os. path. exists( dnn_config_path) : print ( "使用本地YOLO模型..." ) self. net= cv2. dnn. readNet( dnn_model_path, dnn_config_path) self. use_yolo= True else : print ( "本地模型不存在，使用OpenCV内置模型..." ) print ( "提示：您可以下载YOLO模型到 models/ 目录以获得更好的检测效果" ) self. use_yolo= False # 使用OpenCV的HOG行人检测器作为备选 self. hog= cv2. HOGDescriptor( ) self. hog. setSVMDetector( cv2. HOGDescriptor_getDefaultPeopleDetector( ) ) # 加载COCO标签 if os. path. exists( labels_path) : print ( "使用本地标签文件..." ) self. labels= self. load_coco_labels( labels_path) else : print ( "使用内置标签..." ) self. labels= self. get_builtin_labels( ) print ( "模型加载完成！" ) except Exceptionas e: print ( f"模型加载失败: { e} " ) print ( "请确保已安装OpenCV" ) sys. exit( 1 ) def load_coco_labels ( self, labels_path) : """从文件加载COCO标签""" labels= { } try : with open ( labels_path, 'r' , encoding= 'utf-8' ) as f: for linein f: if 'id:' in lineand 'name:' in line: # 解析标签文件 parts= line. strip( ) . split( ) if len ( parts) >= 4 : id_part= parts[ 1 ] name_part= parts[ 3 ] . strip( '"' ) try : label_id= int ( id_part) labels[ label_id] = name_partexcept ValueError: continue except Exceptionas e: print ( f"加载标签文件失败: { e} " ) return self. get_builtin_labels( ) return labelsdef get_builtin_labels ( self) : """获取内置标签""" return { 0 : 'person' , 1 : 'bicycle' , 2 : 'car' , 3 : 'motorcycle' , 4 : 'airplane' , 5 : 'bus' , 6 : 'train' , 7 : 'truck' , 8 : 'boat' , 9 : 'traffic light' , 10 : 'fire hydrant' , 11 : 'stop sign' , 12 : 'parking meter' , 13 : 'bench' , 14 : 'bird' , 15 : 'cat' , 16 : 'dog' , 17 : 'horse' , 18 : 'sheep' , 19 : 'cow' , 20 : 'elephant' , 21 : 'bear' , 22 : 'zebra' , 23 : 'giraffe' , 24 : 'backpack' , 25 : 'umbrella' , 26 : 'handbag' , 27 : 'tie' , 28 : 'suitcase' , 29 : 'frisbee' , 30 : 'skis' , 31 : 'snowboard' , 32 : 'sports ball' , 33 : 'kite' , 34 : 'baseball bat' , 35 : 'baseball glove' , 36 : 'skateboard' , 37 : 'surfboard' , 38 : 'tennis racket' , 39 : 'bottle' , 40 : 'wine glass' , 41 : 'cup' , 42 : 'fork' , 43 : 'knife' , 44 : 'spoon' , 45 : 'bowl' , 46 : 'banana' , 47 : 'apple' , 48 : 'sandwich' , 49 : 'orange' , 50 : 'broccoli' , 51 : 'carrot' , 52 : 'hot dog' , 53 : 'pizza' , 54 : 'donut' , 55 : 'cake' , 56 : 'chair' , 57 : 'couch' , 58 : 'potted plant' , 59 : 'bed' , 60 : 'dining table' , 61 : 'toilet' , 62 : 'tv' , 63 : 'laptop' , 64 : 'mouse' , 65 : 'remote' , 66 : 'keyboard' , 67 : 'cell phone' , 68 : 'microwave' , 69 : 'oven' , 70 : 'toaster' , 71 : 'sink' , 72 : 'refrigerator' , 73 : 'book' , 74 : 'clock' , 75 : 'vase' , 76 : 'scissors' , 77 : 'teddy bear' , 78 : 'hair drier' , 79 : 'toothbrush' } def detect_objects_yolo ( self, img) : """使用YOLO模型检测物体""" height, width= img. shape[ : 2 ] # 预处理图像 blob= cv2. dnn. blobFromImage( img, 1 / 255.0 , ( 416 , 416 ) , swapRB= True , crop= False ) self. net. setInput( blob) # 前向传播 layer_names= self. net. getLayerNames( ) output_layers= [ layer_names[ i- 1 ] for iin self. net. getUnconnectedOutLayers( ) ] outputs= self. net. forward( output_layers) # 解析检测结果 detections= [ ] for outputin outputs: for detectionin output: scores= detection[ 5 : ] class_id= np. argmax( scores) confidence= scores[ class_id] if confidence> 0.5 : # 置信度阈值 center_x= int ( detection[ 0 ] * width) center_y= int ( detection[ 1 ] * height) w= int ( detection[ 2 ] * width) h= int ( detection[ 3 ] * height) x= int ( center_x- w/ 2 ) y= int ( center_y- h/ 2 ) detection_info= { 'class_id' : class_id, 'class_name' : self. labels. get( class_id, f'Unknown- { class_id} ' ) , 'confidence' : float ( confidence) , 'bbox' : [ x, y, x+ w, y+ h] } detections. append( detection_info) return detectionsdef detect_objects_hog ( self, img) : """使用HOG检测器检测行人""" # 检测行人 boxes, weights= self. hog. detectMultiScale( img, winStride= ( 8 , 8 ) , padding= ( 4 , 4 ) , scale= 1.05 ) detections= [ ] for ( x, y, w, h) , weightin zip ( boxes, weights) : detection_info= { 'class_id' : 0 , 'class_name' : 'person' , 'confidence' : float ( weight) , 'bbox' : [ x, y, x+ w, y+ h] } detections. append( detection_info) return detectionsdef detect_objects ( self, image_path) : """检测图片中的物体""" try : # 读取图像 img= cv2. imread( image_path) if imgis None : print ( f"无法读取图像: { image_path} " ) return None , [ ] print ( "正在进行物体检测..." ) # 根据可用模型选择检测方法 if self. use_yolo: detections= self. detect_objects_yolo( img) else : detections= self. detect_objects_hog( img) return img, detectionsexcept Exceptionas e: print ( f"物体检测失败: { e} " ) return None , [ ] def draw_detections ( self, img, detections) : """在图像上绘制检测结果""" if imgis None or not detections: return imgfor detectionin detections: bbox= detection[ 'bbox' ] class_name= detection[ 'class_name' ] confidence= detection[ 'confidence' ] # 绘制边界框 x1, y1, x2, y2= bbox cv2. rectangle( img, ( x1, y1) , ( x2, y2) , ( 0 , 255 , 0 ) , 2 ) # 绘制标签 label= f" { class_name} : { confidence: .2f } " = cv2. getTextSize( label, cv2. FONT_HERSHEY_SIMPLEX, 0.5 , 2 ) [ 0 ] cv2. rectangle( img, ( x1, y1- label_size[ 1 ] - 10 ) , ( x1+ label_size[ 0 ] , y1) , ( 0 , 255 , 0 ) , - 1 ) cv2. putText( img, label, ( x1, y1- 5 ) , cv2. FONT_HERSHEY_SIMPLEX, 0.5 , ( 0 , 0 , 0 ) , 2 ) return imgdef analyze_image ( self, image_path) : """分析图像中的物体""" if not os. path. exists( image_path) : print ( f"错误: 文件不存在 { image_path} " ) return print ( f"\n { '=' * 50 } " ) print ( f"分析图像: { image_path} " ) print ( f" { '=' * 50 } " ) # 基本信息 try : img= cv2. imread( image_path) if imgis not None : height, width, channels= img. shapeprint ( f"图像尺寸: { width} x { height} " ) print ( f"颜色通道: { channels} " ) file_size= os. path. getsize( image_path) / 1024 print ( f"文件大小: { file_size: .2f } KB" ) else : print ( "无法读取图像" ) return except Exceptionas e: print ( f"读取图像失败: { e} " ) return # 物体检测 print ( f"\n { '=' * 20 } 物体检测 { '=' * 20 } " ) result_img, detections= self. detect_objects( image_path) if detections: print ( f"检测到 { len ( detections) } 个物体:" ) print ( "-" * 40 ) # 按置信度排序 detections. sort( key= lambda x: x[ 'confidence' ] , reverse= True ) for i, detectionin enumerate ( detections, 1 ) : print ( f" { i} . { detection[ 'class_name' ] } " ) print ( f" 置信度: { detection[ 'confidence' ] : .3f } " ) print ( f" 位置: { detection[ 'bbox' ] } " ) print ( ) # 统计物体类型 object_counts= { } for detectionin detections: class_name= detection[ 'class_name' ] object_counts[ class_name] = object_counts. get( class_name, 0 ) + 1 print ( "物体统计:" ) for obj_type, countin object_counts. items( ) : print ( f" { obj_type} : { count} 个" ) else : print ( "未检测到任何物体" ) # 保存结果 if result_imgis not None and detections: result_img= self. draw_detections( result_img, detections) output_path= f"detected_ { os. path. basename( image_path) } " . imwrite( output_path, result_img) print ( f"\n检测结果已保存: { output_path} " ) def batch_process ( self, folder_path) : """批量处理文件夹中的图像""" if not os. path. isdir( folder_path) : print ( f"错误: { folder_path} 不是有效文件夹" ) return extensions= [ '.jpg' , '.jpeg' , '.png' , '.bmp' , '.tiff' ] image_files= [ ] for extin extensions: image_files. extend( [ ffor fin os. listdir( folder_path) if f. lower( ) . endswith( ext) ] ) if not image_files: print ( f"在 { folder_path} 中未找到图像文件" ) return print ( f"找到 { len ( image_files) } 个图像文件" ) for i, image_filein enumerate ( image_files, 1 ) : image_path= os. path. join( folder_path, image_file) print ( f"\n处理 { i} / { len ( image_files) } : { image_file} " ) self. analyze_image( image_path) def main ( ) : parser= argparse. ArgumentParser( description= '本地物体检测工具' ) parser. add_argument( 'image_path' , nargs= '?' , help = '要检测的图片路径' ) parser. add_argument( '--batch' , help = '批量处理文件夹' ) parser. add_argument( '--confidence' , type = float , default= 0.5 , help = '检测置信度阈值 (默认: 0.5)' ) args= parser. parse_args( ) if not args. image_pathand not args. batch: print ( "请提供图像路径或使用 --batch 参数指定文件夹" ) parser. print_help( ) return detector= LocalObjectDetector( ) if args. batch: detector. batch_process( args. batch) else : detector. analyze_image( args. image_path) if __name__== "__main__" : main( ) 代码运行命令 python chkimg.py 图片路径.jpg python chkimg.py--batch 文件夹路径 python chkimg.py-h 识别结果
