VISL项目在完成了 | 02年
A typical schema for the system: 该系统一个典型的模式： Abstract 摘要 The purpose of this project was to build a real time application which recognizes license plates from cars at a gate, for example at the entrance of a parking area. 这个项目的目的是建立从汽车板在门入口处时，例如A区牌照时停车一个真正的应用程序，它已承认。 The system, based on regular PC with video camera, catches video frames which include a visible car license plate and processes them. 该系统具有视频摄像机的普通PC机，渔获量的视频帧，其中包括一个明显的汽车牌照和处理它们。 Once a license plate is detected, its digits are recognized, displayed on the User Interface or checked against a database. The focus is on the design of algorithms used for extracting the license plate from a single image, isolating the characters of the plate and identifying the individual characters. 一旦发现车牌，它的数字确认，并显示在用户界面或数据库核对一。形象的重点是设计一个单一的算法车牌从用于提取，分离板的特点及识别单个字符。 The background: 背景： There have been similar past projects at the Lab. 目前已在实验室过去类似的项目。 including projects which implemented the whole system. The purpose of this project was first and foremost to improve the accuracy of the program, and whenever possible its time-complexity. 包括项目实施的整个系统。这个项目的目的首先是改善方案的准确度，并尽可能其时间复杂度。 All the past projects at the Lab. 该实验室的所有项目在过去。 had poor accuracy according to the tests we made on the set of 45 images we used in our program and were successful only when particular conditions were satisfied. For this reason, except from very rare cases the entire program was written again. 根据精度不佳的测试中，我们就程序设置的45个影像，我们用我们的成功，并只有在特定的条件感到满意。 出于这个原因，除了再次从非常罕见的情况下，整个程序写。 Brief description of the implementation: 简要说明执行情况： Our license plate recognition system can be roughly broken down into the following block diagram. 我们的车牌识别系统可大致分为以下框图。 Block diagram of the global system. 框图全球系统。 Alternatively this progression could be viewed as the reduction or suppression of unwanted information from the information carrying signal, here a video sequence containing vast amounts of irrelevant information, to abstract symbols in the form of the characters of a license place. 另外这个进程可以被看作是减少或地方的牌照抑制有害信息从携带信息的信号，这里是一个视频序列包含大量无关信息的特点，形式抽象符号的研究。 The Optical Character recognition (OCR) has been made using the Neural Network technique, using a feed-forward network with 3 layers, 200 neurons in the input layer, 20 neurons in the middle layer, and 10 neurons in the output layer. 光学字符识别（OCR）已采用神经网络技术，采用神经元在输出层的前馈网络的3层，200个神经元在20输入层，中间神经元在10层，。 We kept the Neural Network dataset used in a precedent project which includes 238 digit image s. 我们保留了神经网络数据集图像用在项目的先例，其中包括238位第resized The detailed steps of our algorithm are described in the following diagram: 我们的算法的详细步骤说明如下图： Block diagram of the program subsystems. 框图程序的子系统。 Here are described the outputs of the main steps described above on a given captured frame: 这里介绍捕获帧的一个给定的产出上面所述的主要步骤： Example of a captured frame 示例捕获帧 Captured frame with yellow regions filtered 黄区域捕获的帧过滤 Captured frame with yellow regions dilated 捕获帧地区扩张黄 License plate region 车牌区域 Determining the angle of the plate using the Radon transform 确定氡角度的变换板的使用 Improved LP region 改进的LP地区 Adjusting the LP Contours - Columns Sum Graph 调整唱片轮廓-列和图 Adjusting the LP Contours - Lines Sum Graph 调整唱片轮廓-线条和图 LP Crop 唱片作物 Gray scale LP 灰度唱片 LP binarization and Equalization using an adaptive threshold 唱片二值化，均衡使用自适应阈值 Binary LP 二进制唱片 Normalized LP 归唱片 Determining the LP horizontal contours using 确定使用的LP水平轮廓 the sum of the lines of the precedent image 图像总和先决行 Normalized LP with contours adjusted 归唱片轮廓调节 Character Segmentation using the peaks-to-valleys method 字符分割使用的 山峰到山谷 方法 Dilated digit image 扩张型数位影像 Adjusting digit images horizontal contours - Line sum graph 调整数字图像水平轮廓-线和图 Contours adjusted digit image 调整的数字图像轮廓 Resized digit image 调整大小的数字图像 OCR digits recognition using the Neural Network method OCR的数字识别的 神经网络 方法 Tools 工具 The implementation of the program was developed on Matlab. 该方案实施开发了基于Matlab。 A demo program on which the user can see all the steps of the different algorithms, set the level of details he wants to get, and the speed of the demo was also written as shown on Figure 29. 一个演示方案，用户可以看到所有的算法步骤的不同，设置水平得到他想要的细节，并演示速度也是29的书面图上所示。 The demo can be started, stopped, or paused. 该演示可以启动，停止或暂停。 In its current version the demo includes 45 images on which the algorithm was successful. 在其最新版本的演示包括45图像上对算法进行了成功。 It is important to notice that the speed of the simulation does not reflect the real speed of the whole algorithm since “pause” commands has been inserted into the program, and the loading of images itself takes time. 重要的是要注意到的仿真速度并不反映“命令真正的速度”停顿了整个算法方案以来已插入，和图像加载本身需要时间。 The Demo Graphical User Interface: 演示图形用户界面： A very useful freeware permitting to test most of the Matlab image processing integrated functions was downloaded from the Mathworks site and a link to its source appear at the end of this page. 一个非常有用的免费软件的大部分功能，允许测试的综合MATLAB图像处理的源下载The MathWorks公司的网站和它的一个链接出现在页面的最后这一点。 The first picture in this page were taken from Hi-Tech Solutions™ site with their authorization. 在这个页面中第一张照片是从高科技解决方案™授权的网站与他们。 Conclusions and future works 结论和未来工作 The first conclusion is that what is trivial for the human eye may appear a very difficult task for the computer, but still computer vision can be very powerful and permit to perform very useful operations as the one we implemented in this project. 第一个结论是，什么是微不足道的项目对于人眼的任务，可能会出现一个非常困难的计算机，计算机视觉，但仍是非常强大的，许可实施本执行非常有用的操作，我们作为一个。 The algorithms used in the program have been tested and proved to be accurate and efficient, but still there are cases when they fail. 该方案在算法中使用已经过测试，证明是准确，高效，但还是有一些情况时，他们失败。 Following are the most important problems we noticed: 以下是最重要的问题，我们注意到： - - The most important problem is the Neural Network dataset size: if enlarged in future implementations, it will largely improve the accuracy of the algorithm. 最重要的问题是神经网络数据集大小：如果在未来扩大的实现，这将在很大程度上提高了算法的准确度。 - - The Candidate selection algorithm in the yellow regions filtered image sometimes fails, and the main improvement would be to refine the statistically fixed parameters used in this algorithm. 黄区域的候选选择算法在过滤图像有时会失败，主要的改善将是固定的算法改进统计此参数用于。 - - In general, all the statistically fixed parameters should be refined by performing more tests. 一般来说，所有的统计应该进一步完善固定参数通过进行更多的测试。 - - The yellow region extraction algorithm sometimes fail, and it would be a good idea in future implementation to join it to the supplementary algorithm which is based on the fact that the lines where the number plate is located in the image have a clear ""signature"" which corresponds to strong grey level variations at somehow ""regular"" intervals which makes it usually possible to distinguish them from other lines in the image, or at least to pre-select some positions where to look further. 黄区域提取算法有时会失败，这将是一个好主意，在今后的执行加入该算法的补充，是“立足于事实，那行中的数字板块位于形象有一个明确的”“签名“这相当于在强大的灰度变化在某种程度上”，“常规”“间隔这使得它的形象通常可以区分他们在从其他线路，或至少预先选择一些位置在哪里看得更远。 - - Generally, the decision algorithms should be improved, and a way to detect error and to make decisions flow circular should be developed, for example, if there are multiple candidates for LP location that satisfies the criterions, testing each one of them according to predefined supplementary criterions, or, in cases of doubt when identifying the digits, that is when the probability of the best guess being correct is below some threshold, the system should refuse to make a decision. 一般来说，决定算法应该得到改善，一种方法来检测错误，并作出决定应制定流动循环，例如，如果有规范，多名候选人为LP的位置，满足他们的每一个测试根据预先定义的补充规范，或怀疑，在情况下，当确定的数字，那是当概率的最好的猜测是正确的决定作出以下是一些门槛时，系统应该拒绝。 Acknowledgment 承认 We are grateful to our project supervisor, the Lab. 我们感谢我们的项目主管，该实验室。 chief engineer Johanan Erez , for his help and guidance throughout this work. 总工程师加利亚的埃雷兹 ，对他的帮助和指导整个工作。 We are also grateful to the Ollendorf Minerva Center Fund for supporting this project. 我们也感谢Ollendorf密涅瓦中心项目资金支持这一点。 Related documentation 相关文档 ∙ Images 图片 ∙ Full documentation 完整的文件 | ||||||
VISL项目在完成了02年 | |||||||
该系统一个典型的模式： A typical schema for the system: 摘要 Abstract 这个项目的目的是建立从汽车板在门入口处时，例如A区牌照时停车一个真正的应用程序，它已承认。 The purpose of this project was to build a real time application which recognizes license plates from cars at a gate, for example at the entrance of a parking area. 该系统具有视频摄像机的普通PC机，渔获量的视频帧，其中包括一个明显的汽车牌照和处理它们。 The system, based on regular PC with video camera, catches video frames which include a visible car license plate and processes them. 一旦发现车牌，它的数字确认，并显示在用户界面或数据库核对一。形象的重点是设计一个单一的算法车牌从用于提取，分离板的特点及识别单个字符。 Once a license plate is detected, its digits are recognized, displayed on the User Interface or checked against a database. The focus is on the design of algorithms used for extracting the license plate from a single image, isolating the characters of the plate and identifying the individual characters. 背景： The background: 目前已在实验室过去类似的项目。 There have been similar past projects at the Lab. 包括项目实施的整个系统。这个项目的目的首先是改善方案的准确度，并尽可能其时间复杂度。 including projects which implemented the whole system. The purpose of this project was first and foremost to improve the accuracy of the program, and whenever possible its time-complexity. 该实验室的所有项目在过去。 All the past projects at the Lab. 根据精度不佳的测试中，我们就程序设置的45个影像，我们用我们的成功，并只有在特定的条件感到满意。 出于这个原因，除了再次从非常罕见的情况下，整个程序写。 had poor accuracy according to the tests we made on the set of 45 images we used in our program and were successful only when particular conditions were satisfied. For this reason, except from very rare cases the entire program was written again. 简要说明执行情况： Brief description of the implementation: 我们的车牌识别系统可大致分为以下框图。 Our license plate recognition system can be roughly broken down into the following block diagram. 框图全球系统。 Block diagram of the global system. 另外这个进程可以被看作是减少或地方的牌照抑制有害信息从携带信息的信号，这里是一个视频序列包含大量无关信息的特点，形式抽象符号的研究。 Alternatively this progression could be viewed as the reduction or suppression of unwanted information from the information carrying signal, here a video sequence containing vast amounts of irrelevant information, to abstract symbols in the form of the characters of a license place. 光学字符识别（OCR）已采用神经网络技术，采用神经元在输出层的前馈网络的3层，200个神经元在20输入层，中间神经元在10层，。 The Optical Character recognition (OCR) has been made using the Neural Network technique, using a feed-forward network with 3 layers, 200 neurons in the input layer, 20 neurons in the middle layer, and 10 neurons in the output layer. 我们保留了神经网络数据集图像用在项目的先例，其中包括238位第 We kept the Neural Network dataset used in a precedent project which includes 238 digit image s. 我们的算法的详细步骤说明如下图： The detailed steps of our algorithm are described in the following diagram: 框图程序的子系统。 Block diagram of the program subsystems. 这里介绍捕获帧的一个给定的产出上面所述的主要步骤： Here are described the outputs of the main steps described above on a given captured frame: 示例捕获帧 Example of a captured frame 黄区域捕获的帧过滤 Captured frame with yellow regions filtered 捕获帧地区扩张黄 Captured frame with yellow regions dilated 车牌区域 License plate region 确定氡角度的变换板的使用 Determining the angle of the plate using the Radon transform 改进的LP地区 Improved LP region 调整唱片轮廓-列和图 Adjusting the LP Contours - Columns Sum Graph 调整唱片轮廓-线条和图 Adjusting the LP Contours - Lines Sum Graph 唱片作物 LP Crop 灰度唱片 Gray scale LP 唱片二值化，均衡使用自适应阈值 LP binarization and Equalization using an adaptive threshold 二进制唱片 Binary LP 归唱片 Normalized LP 确定使用的LP水平轮廓 Determining the LP horizontal contours using 图像总和先决行 the sum of the lines of the precedent image 归唱片轮廓调节 Normalized LP with contours adjusted 字符分割使用的 山峰到山谷 方法 Character Segmentation using the peaks-to-valleys method 扩张型数位影像 Dilated digit image 调整数字图像水平轮廓-线和图 Adjusting digit images horizontal contours - Line sum graph 调整的数字图像轮廓 Contours adjusted digit image 调整大小的数字图像 Resized digit image OCR的数字识别的 神经网络 方法 OCR digits recognition using the Neural Network method 工具 Tools 该方案实施开发了基于Matlab。 The implementation of the program was developed on Matlab. 一个演示方案，用户可以看到所有的算法步骤的不同，设置水平得到他想要的细节，并演示速度也是29的书面图上所示。 A demo program on which the user can see all the steps of the different algorithms, set the level of details he wants to get, and the speed of the demo was also written as shown on Figure 29. 该演示可以启动，停止或暂停。 The demo can be started, stopped, or paused. 在其最新版本的演示包括45图像上对算法进行了成功。 In its current version the demo includes 45 images on which the algorithm was successful. 重要的是要注意到的仿真速度并不反映“命令真正的速度”停顿了整个算法方案以来已插入，和图像加载本身需要时间。 It is important to notice that the speed of the simulation does not reflect the real speed of the whole algorithm since “pause” commands has been inserted into the program, and the loading of images itself takes time. 演示图形用户界面： The Demo Graphical User Interface: 一个非常有用的免费软件的大部分功能，允许测试的综合MATLAB图像处理的源下载The MathWorks公司的网站和它的一个链接出现在页面的最后这一点。 A very useful freeware permitting to test most of the Matlab image processing integrated functions was downloaded from the Mathworks site and a link to its source appear at the end of this page. 在这个页面中第一张照片是从高科技解决方案™授权的网站与他们。 The first picture in this page were taken from Hi-Tech Solutions™ site with their authorization. 结论和未来工作 Conclusions and future works 第一个结论是，什么是微不足道的项目对于人眼的任务，可能会出现一个非常困难的计算机，计算机视觉，但仍是非常强大的，许可实施本执行非常有用的操作，我们作为一个。 The first conclusion is that what is trivial for the human eye may appear a very difficult task for the computer, but still computer vision can be very powerful and permit to perform very useful operations as the one we implemented in this project. 该方案在算法中使用已经过测试，证明是准确，高效，但还是有一些情况时，他们失败。 The algorithms used in the program have been tested and proved to be accurate and efficient, but still there are cases when they fail. 以下是最重要的问题，我们注意到： Following are the most important problems we noticed: - - 最重要的问题是神经网络数据集大小：如果在未来扩大的实现，这将在很大程度上提高了算法的准确度。 The most important problem is the Neural Network dataset size: if enlarged in future implementations, it will largely improve the accuracy of the algorithm. - - 黄区域的候选选择算法在过滤图像有时会失败，主要的改善将是固定的算法改进统计此参数用于。 The Candidate selection algorithm in the yellow regions filtered image sometimes fails, and the main improvement would be to refine the statistically fixed parameters used in this algorithm. - - 一般来说，所有的统计应该进一步完善固定参数通过进行更多的测试。 In general, all the statistically fixed parameters should be refined by performing more tests. - - 黄区域提取算法有时会失败，这将是一个好主意，在今后的执行加入该算法的补充，是“立足于事实，那行中的数字板块位于形象有一个明确的”“签名“这相当于在强大的灰度变化在某种程度上”，“常规”“间隔这使得它的形象通常可以区分他们在从其他线路，或至少预先选择一些位置在哪里看得更远。 The yellow region extraction algorithm sometimes fail, and it would be a good idea in future implementation to join it to the supplementary algorithm which is based on the fact that the lines where the number plate is located in the image have a clear ""signature"" which corresponds to strong grey level variations at somehow ""regular"" intervals which makes it usually possible to distinguish them from other lines in the image, or at least to pre-select some positions where to look further. - - 一般来说，决定算法应该得到改善，一种方法来检测错误，并作出决定应制定流动循环，例如，如果有规范，多名候选人为LP的位置，满足他们的每一个测试根据预先定义的补充规范，或怀疑，在情况下，当确定的数字，那是当概率的最好的猜测是正确的决定作出以下是一些门槛时，系统应该拒绝。 Generally, the decision algorithms should be improved, and a way to detect error and to make decisions flow circular should be developed, for example, if there are multiple candidates for LP location that satisfies the criterions, testing each one of them according to predefined supplementary criterions, or, in cases of doubt when identifying the digits, that is when the probability of the best guess being correct is below some threshold, the system should refuse to make a decision. 承认 Acknowledgment 我们感谢我们的项目主管，该实验室。 We are grateful to our project supervisor, the Lab. 总工程师加利亚的埃雷兹 ，对他的帮助和指导整个工作。 chief engineer Johanan Erez , for his help and guidance throughout this work. 我们也感谢Ollendorf密涅瓦中心项目资金支持这一点。 We are also grateful to the Ollendorf Minerva Center Fund for supporting this project. 相关文档 Related documentation ∙ 图片 Images ∙ 完整的文件 Full documentation | |||||||
版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。
发表评论