摘 要
随着人民的物质生活和小轿车的日益流行，我国的机动车销售量也在一年比一年地增长，由此引发的一系列的问题也越来越突出，道路拥堵、事故频发、车位难找。目前，停车场还在建设之中，但速度远远比不上汽车的增长速度，迫切需要一种有效的智能停车场管理系统，以最大限度有效地利用现有停车场资源，从而改善交通状况，提高城市居民生活体验。
文章在对目前国内外智能停车场的发展和发展状况进行了深入的剖析之后，对其各个构成部分进行了深入的分析，并在此基础上对目前的停车场管理体系进行了一些完善和完善，以期对今后的智能化城市交通建设有所裨益。针对 HSI汽车牌照的特点，本文采用 BP神经网络进行车牌识别。首先，利用 Retinex增强算法对所采集到的车辆进行预处理，从而在有灯光或烟尘遮挡的情况下，牌照的特征更加明显，然后利用 HSI彩色空间定位技术进行车牌的位置和位置校正和文字的分离，最后利用 BP神经网络对预先确定的文字进行训练和辨识，最后以更小的延迟和更好的健壮度完成车牌的自动辨识。针对目前停车场管理中存在的停车场管理问题，提出一种基于蚂蚁算法的停车场定位算法。在自动选取泊位方式下，获得了进场点和泊位座标后，利用预先设定的泊位模型，求出最小的最小转弯数的泊车路线；在这种情况下，系统会根据距离电梯入口和出口的位置，计算出更短的距离和更少的弯道。
关键词：智能车库 车牌识别 路径规划
Abstract
With the increasing popularity of people's material life and cars, the sales volume of motor vehicles in China is also increasing year by year. A series of problems caused by this are also increasingly prominent, including road congestion, frequent accidents, and hard to find parking spaces. At present, the parking lot is still under construction, but the speed is far from the growth rate of cars. There is an urgent need for an effective intelligent parking lot management system to make the most effective use of the existing parking lot resources, thus improving the traffic conditions and improving the living experience of urban residents.
After in-depth analysis of the development and development of intelligent parking lots at home and abroad, the article makes in-depth analysis of its various components, and on this basis makes some improvements and improvements to the current parking lot management system, in order to benefit the future intelligent urban traffic construction. According to the characteristics of HSI vehicle license plate, this paper uses BP neural network for vehicle license plate recognition. First, Retinex enhancement algorithm is used to preprocess the collected vehicles, so that the characteristics of the license plate are more obvious in the case of light or smoke. Then HSI color space positioning technology is used to correct the position and position of the license plate and separate the text. Finally, BP neural network is used to train and identify the pre-determined text, Finally, the automatic identification of vehicle license plate is completed with less delay and better robustness. Aiming at the parking lot management problems existing in the current parking lot management, a parking lot location algorithm based on ant algorithm is proposed. Under the mode of automatic selection of parking space, after obtaining the approach point and parking coordinates, the parking route with the minimum number of turns can be obtained by using the preset parking space model; In this case, the system will calculate the shorter distance and fewer curves according to the location of the elevator entrance and exit.
Key words: Intelligent garage license plate recognition path planning
目 录
摘 要 I
Abstract II
1 前言 1
1.1 研究的背景和意义 1
1.2 国内外研究 2
1.2.1 国内研究 2
1.2.2 国外研究 2
2 车库管理系统中关键技术 3
2.1 车牌图像处理方法 3
2.1.1 我国小型汽车车牌现行规范 3
2.1.2 我国小型汽车车牌特征 4
2.1.3 Retinex 图像增强方法 5
2.2 字符识别方法 5
2.2.1 支持向量机(SVM)字符识别方法 5
2.2.2 人工神经网络(ANN)字符识别方法 6
2.3 路径规划算法 6
2.3.1 空间表示 6
2.3.2 搜索算法 7
3 车牌识别系统设计研究 7
3.1 系统总体结构 8
3.2 车牌图像处理 8
3.2.1 车牌图像获取 8
3.2.2 车牌定位 8
3.2.3 车牌预处理 9
4 系统集成及软件设计 10
4.1 系统结构设计及功能分析 10
4.1.1 系统结构设计 10
4.1.2 系统功能分析 11
4.2 系统活动图分析与设计 12
4.2.2 操作员活动图分析设计 13
5 结论 14
参考文献 16
致 谢 18
随着人民的物质生活和小轿车的日益流行，我国的机动车销售量也在一年比一年地增长，由此引发的一系列的问题也越来越突出，道路拥堵、事故频发、车位难找。目前，停车场还在建设之中，但速度远远比不上汽车的增长速度，迫切需要一种有效的智能停车场管理系统，以最大限度有效地利用现有停车场资源，从而改善交通状况，提高城市居民生活体验。
文章在对目前国内外智能停车场的发展和发展状况进行了深入的剖析之后，对其各个构成部分进行了深入的分析，并在此基础上对目前的停车场管理体系进行了一些完善和完善，以期对今后的智能化城市交通建设有所裨益。针对 HSI汽车牌照的特点，本文采用 BP神经网络进行车牌识别。首先，利用 Retinex增强算法对所采集到的车辆进行预处理，从而在有灯光或烟尘遮挡的情况下，牌照的特征更加明显，然后利用 HSI彩色空间定位技术进行车牌的位置和位置校正和文字的分离，最后利用 BP神经网络对预先确定的文字进行训练和辨识，最后以更小的延迟和更好的健壮度完成车牌的自动辨识。针对目前停车场管理中存在的停车场管理问题，提出一种基于蚂蚁算法的停车场定位算法。在自动选取泊位方式下，获得了进场点和泊位座标后，利用预先设定的泊位模型，求出最小的最小转弯数的泊车路线；在这种情况下，系统会根据距离电梯入口和出口的位置，计算出更短的距离和更少的弯道。
关键词：智能车库 车牌识别 路径规划
Abstract
With the increasing popularity of people's material life and cars, the sales volume of motor vehicles in China is also increasing year by year. A series of problems caused by this are also increasingly prominent, including road congestion, frequent accidents, and hard to find parking spaces. At present, the parking lot is still under construction, but the speed is far from the growth rate of cars. There is an urgent need for an effective intelligent parking lot management system to make the most effective use of the existing parking lot resources, thus improving the traffic conditions and improving the living experience of urban residents.
After in-depth analysis of the development and development of intelligent parking lots at home and abroad, the article makes in-depth analysis of its various components, and on this basis makes some improvements and improvements to the current parking lot management system, in order to benefit the future intelligent urban traffic construction. According to the characteristics of HSI vehicle license plate, this paper uses BP neural network for vehicle license plate recognition. First, Retinex enhancement algorithm is used to preprocess the collected vehicles, so that the characteristics of the license plate are more obvious in the case of light or smoke. Then HSI color space positioning technology is used to correct the position and position of the license plate and separate the text. Finally, BP neural network is used to train and identify the pre-determined text, Finally, the automatic identification of vehicle license plate is completed with less delay and better robustness. Aiming at the parking lot management problems existing in the current parking lot management, a parking lot location algorithm based on ant algorithm is proposed. Under the mode of automatic selection of parking space, after obtaining the approach point and parking coordinates, the parking route with the minimum number of turns can be obtained by using the preset parking space model; In this case, the system will calculate the shorter distance and fewer curves according to the location of the elevator entrance and exit.
Key words: Intelligent garage license plate recognition path planning
目 录
摘 要 I
Abstract II
1 前言 1
1.1 研究的背景和意义 1
1.2 国内外研究 2
1.2.1 国内研究 2
1.2.2 国外研究 2
2 车库管理系统中关键技术 3
2.1 车牌图像处理方法 3
2.1.1 我国小型汽车车牌现行规范 3
2.1.2 我国小型汽车车牌特征 4
2.1.3 Retinex 图像增强方法 5
2.2 字符识别方法 5
2.2.1 支持向量机(SVM)字符识别方法 5
2.2.2 人工神经网络(ANN)字符识别方法 6
2.3 路径规划算法 6
2.3.1 空间表示 6
2.3.2 搜索算法 7
3 车牌识别系统设计研究 7
3.1 系统总体结构 8
3.2 车牌图像处理 8
3.2.1 车牌图像获取 8
3.2.2 车牌定位 8
3.2.3 车牌预处理 9
4 系统集成及软件设计 10
4.1 系统结构设计及功能分析 10
4.1.1 系统结构设计 10
4.1.2 系统功能分析 11
4.2 系统活动图分析与设计 12
4.2.2 操作员活动图分析设计 13
5 结论 14
参考文献 16
致 谢 18
