基于python下的 车牌识别代码

import cv2 import os from cardsetting import Cardsetting from cardtrainmodel import Trainsvm import cardpredictor as cp def run_video_predict(video): cardsetting = Cardsetting() trainsvm = Trainsvm(cardsetting) # 检测是否有训练好的数据存档 if os.path.exists(cardsetting.char_letterAnumber_traindata_path) and os.path.exists(cardsetting.char_chinese_traindata_path) and os.path.exists(cardsetting.char_one_letter_traindata_path): print('检索到数据存档') trainsvm.model_letterAnumber.load(cardsetting.char_letterAnumber_traindata_path) trainsvm.model_one_letter.load(cardsetting.char_one_letter_traindata_path) trainsvm.model_chinese.load(cardsetting.char_chinese_traindata_path) else: # 没有则用原始数据训练模型并保存 trainsvm.train_svm() cot1 = 0 #车道1 this_card1 = False all_cards1 = [] this_card_nunbers1 = [] this_card_number_str1 = [] all_predict_results_to_chars1 = [[], [], [], [], [], [], [], [], []] cot2 = 0 #车道2 this_card2 = False all_cards2 = [] this_card_nunbers2 = [] this_card_number_str2 = [] all_predict_results_to_chars2 = [[], [], [], [], [], [], [], [], []] cot3 = 0 #车道3 this_card3 = False all_cards3 = [] this_card_nunbers3 = [] this_card_number_str3 = [] all_predict_results_to_chars3 = [[], [], [], [], [], [], [], [], []] cot4 = 0 #车道4 this_card4 = False all_cards4 = [] this_card_nunbers4 = [] this_card_number_str4 = [] all_predict_results_to_chars4 = [[], [], [], [], [], [], [], [], []] cot = 0 fps = video.get(cv2.CAP_PROP_FPS) print('\nFPS:', fps) success,frame = video.read() while success: frame1, frame2, frame3, frame4 = vehicle_lane(frame, cardsetting) if cardsetting.vehicle_Lane_number >= 1: #第一车道 card_screenshots1, predict_results1 = cp.run_picture_predict(frame1, trainsvm, cardsetting) if len(predict_results1) != 0: cot1 = 0 this_card1 = True for i, char in enumerate(predict_results1[0]): all_predict_results_to_chars1[i].append(char) #车牌各个字符集合 else: if cot1 == cardsetting.fps_pass: this_card1 = False cot1 += 1 # 统计帧经过 if this_card1 == False and len(all_predict_results_to_chars1[0]) != 0: for chars1 in all_predict_results_to_chars1: #各个字符去重统计数量 card_cot1 = {} chars_sets1 = set(chars1) #转为集合去重 for chars_set1 in chars_sets1: #统计数量 加到字典 card_cot1[chars_set1] = chars1.count(chars_set1) card_cot1 = sorted(card_cot1.items(), key = lambda x: x[1], reverse = True) #根据数量排序 print(card_cot1) if len(card_cot1) != 0: this_card_nunbers1.append(card_cot1[0][0]) #得到车牌各个字符串列表 result_str1 = '' for str1 in this_card_nunbers1: #转化为字符串 result_str1 += str1 this_card_number_str1.append(result_str1) if this_card_number_str1 not in all_cards1: all_cards1.append(this_card_number_str1) this_card_nunbers1 = [] this_card_number_str1 = [] all_predict_results_to_chars1 = [[], [], [], [], [], [], [], [], []] if cardsetting.vehicle_Lane_number >= 2: #第二车道 card_screenshots2, predict_results2 = cp.run_picture_predict(frame2, trainsvm, cardsetting) if len(predict_results2) != 0: cot2 = 0 this_card2 = True for i, char in enumerate(predict_results2[0]): all_predict_results_to_chars2[i].append(char) #车牌各个字符集合 else: if cot2 == cardsetting.fps_pass: this_card2 = False cot2 += 1 # 统计帧经过 if this_card2 == False and len(all_predict_results_to_chars2[0]) != 0: for chars2 in all_predict_results_to_chars2: #各个字符去重统计数量 card_cot2 = {} chars_sets2 = set(chars2) #转为集合去重 for chars_set2 in chars_sets2: #统计数量 加到字典 card_cot2[chars_set2] = chars2.count(chars_set2) card_cot2 = sorted(card_cot2.items(), key = lambda x: x[1], reverse = True) # 根据数量排序 print(card_cot2) if len(card_cot2) != 0: this_card_nunbers2.append(card_cot2[0][0]) #得到车牌各个字符串列表 result_str2 = '' for str2 in this_card_nunbers2:#转化为字符串 result_str2 += str2 this_card_number_str2.append(result_str2) if this_card_number_str2 not in all_cards2: all_cards2.append(this_card_number_str2) this_card_nunbers2 = [] this_card_number_str2 = [] all_predict_results_to_chars2 = [[], [], [], [], [], [], [], [], []] if cardsetting.vehicle_Lane_number >= 3: #第三车道 card_screenshots3, predict_results3 = cp.run_picture_predict(frame3, trainsvm, cardsetting) if len(predict_results3) != 0: cot3 = 0 this_card3 = True for i, char in enumerate(predict_results3[0]): all_predict_results_to_chars3[i].append(char) #车牌各个字符集合 else: if cot3 == cardsetting.fps_pass: this_card3 = False cot3 += 1 #统计帧经过 if this_card3 == False and len(all_predict_results_to_chars3[0]) != 0: for chars3 in all_predict_results_to_chars3: #各个字符去重统计数量 card_cot3 = {} chars_sets3 = set(chars3) #转为集合去重 for chars_set3 in chars_sets3: #统计数量 加到字典 card_cot3[chars_set3] = chars3.count(chars_set3) card_cot3 = sorted(card_cot3.items(), key=lambda x: x[1], reverse=True) #根据数量排序 print(card_cot3) if len(card_cot3) != 0: this_card_nunbers3.append(card_cot3[0][0]) #得到车牌各个字符串列表 result_str3 = '' for str3 in this_card_nunbers3: #转化为字符串 result_str3 += str3 this_card_number_str3.append(result_str3) if this_card_number_str3 not in all_cards3: all_cards3.append(this_card_number_str3) this_card_nunbers3 = [] this_card_number_str3=[] all_predict_results_to_chars3 = [[], [], [], [], [], [], [], [], []] if cardsetting.vehicle_Lane_number >= 4: #第四车道 card_screenshots4, predict_results4 = cp.run_picture_predict(frame4, trainsvm, cardsetting) if len(predict_results4) != 0: cot4 = 0 this_card4 = True for i, char in enumerate(predict_results4[0]): all_predict_results_to_chars4[i].append(char) #车牌各个字符集合 else: if cot4 == cardsetting.fps_pass: this_card4 = False cot4 += 1 #统计帧跳过 if this_card4 == False and len(all_predict_results_to_chars4[0]) != 0: for chars4 in all_predict_results_to_chars4: #各个字符去重统计数量 card_cot4 = {} chars_sets4 = set(chars4) #转为集合去重 for chars_set4 in chars_sets4: #统计数量 加到字典 card_cot4[chars_set4] = chars4.count(chars_set4) card_cot4 = sorted(card_cot4.items(), key = lambda x: x[1], reverse = True) # 根据数量排序 print(card_cot4) if len(card_cot4) != 0: this_card_nunbers4.append(card_cot4[0][0]) #得到车牌各个字符串列表 result_str4 = '' for str4 in this_card_nunbers4:#转化为字符串 result_str4 += str4 this_card_number_str4.append(result_str4) if this_card_number_str4 not in all_cards4: all_cards4.append(this_card_number_str4) this_card_nunbers4 = [] this_card_number_str4 = [] all_predict_results_to_chars4 = [[], [], [], [], [], [], [], [], []] cot += 1 print('Frame:', cot) cv2.imshow('Frame', frame) if cv2.waitKey(1) & 0xFF == ord('q'): break success, frame = video.read() return(all_cards1, all_cards2, all_cards3, all_cards4) def vehicle_lane(frame, cardsetting): """车道分割""" if cardsetting.vehicle_Lane_number == 1: frame1=frame[0:, cardsetting.Lane1:cardsetting.Lane2] frame2=None frame3=None frame4=None elif cardsetting.vehicle_Lane_nu