基于YOLOv3的红绿灯检测识别（Python源码可直接运行）

在实习的期间为公司写的红绿灯检测，基于YOLOv3的训练好的权重，不需要自己重新训练，只需要调用yolov3.weights，可以做到视频或图片中红绿灯的检测识别。

自动检测识别效果

1.红灯检测

2.绿灯检测

python源码

""" Class definition of YOLO_v3 style detection model on image and video """import colorsys import os from timeit import default_timer as timer import cv2 import numpy as np from keras import backend as K from keras.models import load_model from keras.layers import Input from PIL import Image, ImageFont, ImageDraw from yolo3.model import yolo_eval, yolo_body, tiny_yolo_body from yolo3.utils import letterbox_image import os from keras.utils import multi_gpu_model import collections class YOLO(object): _defaults = {"model_path":''model_data/yolo.h5'',"anchors_path":''model_data/yolo_anchors.txt'',"classes_path":''model_data/coco_classes.txt'',"score":0.3,"iou":0.35,"model_image_size": (416,416),"gpu_num":1, } @classmethod def get_defaults(cls, n):ifn in cls._defaults:returncls._defaults[n]else:return"Unrecognized attribute name ''"+ n +"''"def __init__(self, **kwargs): self.__dict__.update(self._defaults)# set up default valuesself.__dict__.update(kwargs)# and update with user overridesself.class_names = self._get_class() self.anchors = self._get_anchors() self.sess = K.get_session() self.boxes, self.scores, self.classes = self.generate() def _get_class(self): classes_path = os.path.expanduser(self.classes_path) withopen(classes_path) as f: class_names = f.readlines() class_names = [c.strip()forc in class_names]returnclass_names def _get_anchors(self): anchors_path = os.path.expanduser(self.anchors_path) withopen(anchors_path) as f: anchors = f.readline() anchors = [float(x)forxin anchors.split('','')]returnnp.array(anchors).reshape(-1,2) def generate(self): model_path = os.path.expanduser(self.model_path) assert model_path.endswith(''.h5''),''Keras modelorweights must be a .h5 file.''# Load model, or construct model and load weights.num_anchors = len(self.anchors) num_classes = len(self.class_names) is_tiny_version = num_anchors==6# default settingtry: self.yolo_model = load_model(model_path, compile=False) except: self.yolo_model = tiny_yolo_body(Input(shape=(None,None,3)), num_anchors//2, num_classes)ifis_tiny_versionelseyolo_body(Input(shape=(None,None,3)), num_anchors//3, num_classes) self.yolo_model.load_weights(self.model_path)# make sure model, anchors and classes matchelse: assert self.yolo_model.layers[-1].output_shape[-1] == num_anchors/len(self.yolo_model.output) * (num_classes +5),''Mismatch between modelandgivenanchorandclass sizes''print(''{} model, anchors,andclasses loaded.''.format(model_path))# Generate colors for drawing bounding boxes.hsv_tuples = [(x/ len(self.class_names),1.,1.)forxin range(len(self.class_names))] self.colors = list(map(lambdax: colorsys.hsv_to_rgb(*x), hsv_tuples)) self.colors = list(map(lambdax: (int(x[0] *255),int(x[1] *255),int(x[2] *255)), self.colors)) np.random.seed(10101)# Fixed seed for consistent colors across runs.np.random.shuffle(self.colors)# Shuffle colors to decorrelate adjacent classes.np.random.seed(None)# Reset seed to default.# Generate output tensor targets for filtered bounding boxes.self.input_image_shape = K.placeholder(shape=(2, ))ifself.gpu_num>=2: self.yolo_model = multi_gpu_model(self.yolo_model, gpus=self.gpu_num) boxes, scores, classes = yolo_, self.input_image_shape, score_threshold=self.score, iou_threshold=self.iou)returnboxes, scores, classes def getColorList(self): dict = collections.defaultdict(list)# 红色lower_red = np.array([156,43,46]) upper_red = np.array([180,255,255]) color_list = [] color_list.append(lower_red) color_list.append(upper_red) dict[''red''] = color_list# 红色2lower_red = np.array([0,43,46]) upper_red = np.array([10,255,255]) color_list = [] color_list.append(lower_red) color_list.append(upper_red) dict[''red2''] = color_list# 橙色lower_orange = np.array([11,43,46]) upper_orange = np.array([25,255,255]) color_list = [] color_list.append(lower_orange) color_list.append(upper_orange) dict[''orange''] = color_list# 黄色lower_yellow = np.array([26,43,46]) upper_yellow = np.array([34,255,255]) color_list = [] color_list.append(lower_yellow) color_list.append(upper_yellow) dict[''yellow''] = color_list# 绿色lower_green = np.array([35,43,46]) upper_green = np.array([77,255,255]) color_list = [] color_list.append(lower_green) color_list.append(upper_green) dict[''green''] = color_listreturndict def get_color(self,frame):print(''go in get_color'') hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV) maxsum = -100color = None color_dict = self.getColorList() score =0type =''black''ford in color_dict: mask = cv2.inRange(hsv, color_dict[d][0], color_dict[d][1])# print(cv2.inRange(hsv, color_dict[d][0], color_dict[d][1]))#cv2.imwrite(''images/triffic/'' + f + d + ''.jpg'', mask)binary = cv2.threshold(mask,127,255, cv2.THRESH_BINARY)[1] binary = cv2.dilate(binary, None, iterations=2) img, cnts, hiera = cv2.findContours(binary.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) sum =0forc in cnts: sum += cv2.contourArea(c)ifsum > maxsum: maxsum = sum color = difsum > score: score = sum type = dreturntype def detect_image(self, image,path):print(''class'',self._get_class()) start = timer()ifself.model_image_size != (None, None): assert self.model_image_size[0]%32 ==0,''Multiples of32required''assert self.model_image_size[1]%32 ==0,''Multiples of32required''boxed_image = letterbox_image(image, tuple(reversed(self.model_image_size)))else: new_image_size = (image.width - (image.width %32), image.height - (image.height %32)) boxed_image = letterbox_image(image, new_image_size) image_data = np.array(boxed_image, dtype=''float32'')print(image_data.shape) image_data /=255. image_data = np.expand_dims(image_data,0)# Add batch dimension.out_boxes, out_scores, out_classes = self.sess.run( [self.boxes, self.scores, self.classes], feed_dict={ self.yolo_model.input: image_data, self.input_image_shape: [image.size[1], image.size[0]], K.learning_phase():0})print(''Found {} boxesfor{}''.format(len(out_boxes),''img'')) font = ImageFont.truetype(font=''font/FiraMono-Medium.otf'', size=np.floor(3e-2* image.size[1] +0.5).astype(''int32'')) thickness = (image.size[0] + image.size[1]) //300thickness =5print(''thickness'',thickness)print(''out_classes'',out_classes) my_class = [''traffic light''] imgcv = cv2.cvtColor(np.asarray(image), cv2.COLOR_RGB2BGR)fori, c in reversed(list(enumerate(out_classes))): predicted_class = self.class_names[c]print(''predicted_class'',predicted_class)ifpredicted_classnotin my_class:continuebox = out_boxes[i] score = out_scores[i] label =''{} {:.2f}''.format(predicted_class, score) draw = ImageDraw.Draw(image) label_size = draw.textsize(label, font) top, left, bottom, right = box top = max(0, np.floor(top +0.5).astype(''int32'')) left = max(0, np.floor(left +0.5).astype(''int32'')) bottom = min(image.size[1], np.floor(bottom +0.5).astype(''int32'')) right = min(image.size[0], np.floor(right +0.5).astype(''int32''))print(label, (left, top), (right, bottom)) img2 = imgcv[top:bottom, left:right] color = self.get_color(img2) cv2.imwrite(''images/triffic/''+path+str(i) +''.jpg'', img2)ifcolor==''red''orcolor ==''red2'': cv2.rectangle(imgcv, (left, top), (right, bottom), color=(0,0,255), lineType=2, thickness=8) cv2.putText(imgcv,''{0}{1:.2f}''.format(color, score), (left, top -15), cv2.FONT_HERSHEY_SIMPLEX,1.2, (0,0,255),4, cv2.LINE_AA) elif color ==''green'': cv2.rectangle(imgcv, (left, top), (right, bottom), color=(0,255,0), lineType=2, thickness=8) cv2.putText(imgcv,''{0}{1:.2f}''.format(color, score), (left, top -15), cv2.FONT_HERSHEY_SIMPLEX,1.2, (0,255,0),4, cv2.LINE_AA)else: cv2.rectangle(imgcv, (left, top), (right, bottom), color=(255,0,0), lineType=2, thickness=8) cv2.putText(imgcv,''{0}{1:.2f}''.format(color, score), (left, top -15), cv2.FONT_HERSHEY_SIMPLEX,1.2, (255,0,0),4, cv2.LINE_AA)print(imgcv.shape) end = timer()print(end - start)returnimgcv def close_session(self): self.sess.close() def detect_img(yolo, img_path,fname): img = Image.open(img_path) importtimet1 = time.time() img = yolo.detect_image(img,fname)print(''time: {}''.format(time.time() - t1))returnimg#yolo.close_session()if__name__==''__main__'': yolo = YOLO() video_full_path =''images/triffic.mp4''output =''images/res.avi''cap = cv2.VideoCapture(video_full_path) cap.set(cv2.CAP_PROP_POS_FRAMES,1)# 设置要获取的帧号fourcc = cv2.VideoWriter_fourcc(*''XVID'') fps = cap.get(cv2.CAP_PROP_FPS) size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)),int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))) out = cv2.VideoWriter(output, fourcc, fps, size) ret = True count =0whileret : count+=1ret, frame = cap.read()ifnotret :print(''结束'')breakimage = Image.fromarray(cv2.cvtColor(frame,cv2.COLOR_BGR2RGB)) image = yolo.detect_image(image,''pic'') out.write(image) cap.release() out.release() cv2.destroyAllWindows()