基于TensorFlow实现的汽车牌照识别

# 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)**