# TensorFlow-ALPR
The project developed using TensorFlow to detect the License Plate from a car and uses the Tesseract Engine to recognize the charactes from the detected plate.
### Software Packs Needed
* <a href='https://www.anaconda.com/download/'>Anaconda 3</a> (**Tool comes with most of the required python packages along with python3 & spyder IDE**)<br>
* <a href='https://github.com/tesseract-ocr/tesseract'>Tesseract Engine</a> (**Must need to be installed**)<br>
### Python Packages Needed
* <a href='https://github.com/tensorflow/tensorflow'>Tensorflow</a><br>
* <a href='https://github.com/skvark/opencv-python'>openCV</a><br>
* <a href='https://github.com/madmaze/pytesseract'>pytesseract</a><br>
* <a href='https://github.com/tzutalin/labelImg'>labelImg</a><br>
### ABOUT PROJECT
* TensorFlow is an open-source software library **(Deep learning)** for dataflow programming across a range of tasks. It is a symbolic math library, and also used for machine learning applications such as neural networks. So we have planned to use it for number plate detection.
#### TRAINING PHASE -- IMAGE LABELING
* Collected the set of 100 images (Cars along with number plate) from the sources such as Google Images and Flickr. Then annotated the set of images by drawing the boundary box over the number plates to send it for the training phase.
* The Annoation gives the co-ordinates of license plates such as **(xmin, ymin, xmax, ymax)**
* Then the co-ordinates are saved into a **XML** file
* All the XML files are grouped and the Co-ordinates are saved in **CSV** file.
* Then the CSV file is converted into **TensorFlow record format**.
* The set of other separate 10 images also gone through the above steps and saved as **Test Record file**
<p align="center">
<img src="custom_plate/image_readme/labelImg.png" width=676 height=450>
</p>
#### GPU TRAINING
* By using the **Tensorflow-gpu** version, the set of annotated images were sent into the Convolutional neural network called as **ssd-mobilenet** where the metrics such as model learning rate, batch of images sent into the network and evaluation configurations were set. The training phase of the model took several days. At last the model came around with the positive result and detected the number plate over the input images.
<p align="center">
<img src="custom_plate/image_readme/test.png" width=676 height=450>
</p>
#### OCR PART
* Then the detected number plate is cropped using Tensorflow, By using the Google **Tesseract-OCR** (Package originally developed to scan hard copy documents to filter out the characters from it) the picture undergoes some coversions using **computer vision** package then the charcters are filtered out.
#### CROP
<p align="left">
<img src="custom_plate/image_readme/crop.png" width=300 height=100>
</p>
#### CONVERSION
<p align="left">
<img src="custom_plate/image_readme/conversion.png" width=300 height=100>
</p>
<p align="center">
<img src="custom_plate/image_readme/char_recog.png" width=900 height=600>
</p>
#### MOTION DETECTION PART
* The basic motion capturing has been implemented to capture the picture of moving vehicle by using the **openCV** where the threshold of the camera is fixed (threshold value changes in according to frame's boundary area). If the vehicle touches the boundary the picture is captured. **(In progress)**
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
Tensorflow openCV pytesseract labelImg TensorFlow是一个开源软件库(深度学习),用于跨一系列任务进行数据流编程。它是一个符号数学库,也用于机器学习应用,如神经网络。所以我们计划用它来检测车牌。 训练阶段——图像标记 收集了100张图片(汽车和车牌)。然后通过在车牌上绘制边界框来对图像集进行注释。 注释给出了车牌的坐标,例如(xmin,ymin,xmax,ymax) 然后坐标被保存到一个XML文件中 所有XML文件都被分组,坐标保存在CSV文件中。 然后将CSV文件转换为TensorFlow记录格式。 其他10个单独的图像集也完成了上述步骤,并保存为测试记录文件 注释的图像集发送到称为ssd mobilenet的卷积神经网络,其中设置了模型学习率、发送到网络的图像批次和评估配置等指标。 最后,该模型得出了肯定的结果,并检测到输入图像上的车牌。
资源详情
资源评论
资源推荐
收起资源包目录
基于TensorFlow实现的汽车牌照识别 (219个子文件)
checkpoint 77B
ssd_mobilenet_v1_pets.config 4KB
faster_rcnn_inception_resnet_v2_atrous_oid.config 4KB
model.ckpt.data-00000-of-00001 21.15MB
model.ckpt.index 9KB
object_detection_tutorial.ipynb 28KB
image2.jpg 1.76MB
image13.jpg 1.52MB
image8.jpg 1.52MB
image10.jpg 1.52MB
image14.jpg 1.52MB
image9.jpg 1.52MB
image11.jpg 1.36MB
image2.jpg 1.35MB
image12.jpg 1.26MB
image6.jpg 1.24MB
image7.jpg 1.16MB
image8.jpg 1.14MB
image13.jpg 1.14MB
image14.jpg 947KB
image1.jpg 485KB
20180402_13h18m22s008695.jpg 462KB
20180402_13h28m31s060635.jpg 455KB
20180402_13h17m47s037806.jpg 455KB
20180402_13h15m40s033267.jpg 455KB
20180402_13h34m21s087296.jpg 455KB
20180402_12h54m37s090834.jpg 454KB
image6.jpg 454KB
20180402_13h17m39s097610.jpg 454KB
20180402_12h56m19s094361.jpg 453KB
20180402_13h34m16s011650.jpg 452KB
20180402_13h28m34s076485.jpg 452KB
20180402_12h56m15s084131.jpg 451KB
20180402_12h56m17s019230.jpg 451KB
20180402_13h16m38s035289.jpg 450KB
20180402_12h56m06s054257.jpg 450KB
20180402_12h56m43s065112.jpg 449KB
20180402_12h56m21s028982.jpg 449KB
20180402_12h55m29s009454.jpg 448KB
20180402_12h56m39s019916.jpg 447KB
20180402_12h56m29s095204.jpg 446KB
20180402_12h56m01s043774.jpg 445KB
20180402_12h56m13s134098.jpg 445KB
20180402_12h55m59s103734.jpg 444KB
20180402_13h33m57s056565.jpg 444KB
20180402_12h56m37s065293.jpg 443KB
20180402_12h55m55s033465.jpg 443KB
20180402_12h55m53s074149.jpg 443KB
20180402_12h56m27s074703.jpg 442KB
20180402_13h34m01s066902.jpg 442KB
20180402_12h55m27s016885.jpg 441KB
20180402_13h33m16s010841.jpg 440KB
20180402_12h53m40s069144.jpg 439KB
image5.jpg 438KB
20180402_13h16m18s064384.jpg 438KB
20180402_12h56m23s084891.jpg 437KB
20180402_13h33m19s085851.jpg 437KB
20180402_12h56m35s089660.jpg 437KB
20180402_12h56m24s094565.jpg 436KB
20180402_13h33m42s086397.jpg 435KB
20180402_13h16m58s071473.jpg 434KB
20180402_12h56m25s069289.jpg 434KB
20180402_13h16m05s033579.jpg 434KB
20180402_12h56m11s074086.jpg 434KB
20180402_13h28m36s067081.jpg 433KB
20180402_12h56m07s073800.jpg 431KB
20180402_13h28m39s045072.jpg 430KB
20180402_12h55m25s056709.jpg 430KB
20180402_12h56m33s065276.jpg 428KB
20180402_13h30m05s058100.jpg 427KB
20180402_13h16m19s034540.jpg 426KB
20180402_13h16m21s069874.jpg 425KB
image2.jpg 422KB
20180402_12h59m40s081553.jpg 422KB
20180402_12h59m13s024683.jpg 420KB
20180402_12h59m21s004355.jpg 419KB
20180402_13h16m20s065528.jpg 418KB
20180402_12h56m09s124105.jpg 417KB
20180402_13h16m23s059532.jpg 417KB
20180402_12h59m23s074095.jpg 416KB
20180402_13h28m37s094933.jpg 415KB
20180402_12h58m16s097857.jpg 415KB
20180402_13h16m10s099137.jpg 414KB
20180402_12h58m28s048099.jpg 412KB
20180402_12h59m09s014751.jpg 410KB
20180402_12h58m18s128476.jpg 410KB
20180402_12h58m15s097713.jpg 408KB
20180402_12h59m11s015031.jpg 408KB
20180402_13h28m35s095124.jpg 408KB
20180402_12h58m22s013402.jpg 406KB
20180402_12h58m17s083247.jpg 406KB
20180402_12h58m26s063321.jpg 406KB
20180402_12h58m20s048994.jpg 406KB
20180402_12h59m17s051471.jpg 404KB
20180402_13h17m40s072984.jpg 404KB
20180402_12h58m19s097643.jpg 403KB
20180402_12h58m23s128823.jpg 397KB
20180402_13h16m22s095458.jpg 397KB
20180402_12h58m25s063350.jpg 396KB
20180402_12h59m06s060140.jpg 395KB
共 219 条
- 1
- 2
- 3
自不量力的A同学
- 粉丝: 47
- 资源: 2711
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功
评论0