shivamshrirao / number_plate_face_recognition_iot_project

Training Number Plate detector using SSD mobilenet v2, Tensorflow Object Detection and face recognition.

Number Plate Reader and Face recognintion.

Mini IOT project I made for number plate detection, reading and face recognition for authorization using RaspberryPi. Firstly it uses an Ultrasonic sensor to see if there's a car. If there is it starts the camera. I have then trained an SingleShotDetector mobilenetV2 deep learning Object detection network to detect the position of Number Plate. The image is then preprocessed with openCV and uses pytesseract to extract text from image. If the number plate is verified it will do face recognition. For that it uses dlib library which is pretrained on 3 million images to generate encodings of face. Then compares that to the ones trained in database. If verified the garage door will open.

Applications

• Home security and automation.
• At trafic lights to automatically read number plate and face of rule violaters and ticket them.
• At toll.
• At offices and parking spaces. Keep track of vehicles and owners and authorization.

https://www.youtube.com/watch?v=F_sAMYax1J0

System Diagram

Click Below for Demo

Face Recognition backend link: https://github.com/ShivamShrirao/face_recognition

number_plate_face_recognition_iot_project's Issues

where pb file?

ssdnet = cv2.dnn.readNetFromTensorflow('trained_model/frozen_inference_graph.pb','trained_model/graph.pbtxt')

Cv2 error can't read .pb file

Traceback (most recent call last):
File "/home/iot/detect.py", line 9, in
ssdnet = cv2.dnn.readNetFromTensorflow('frozen_inference_graph.pb','graph.pbtxt')
cv2.error: OpenCV(4.6.0) /tmp/pip-wheel-u79916uk/opencv-python_ea2489746b3a43bfb3f2b5331b7ab47a/opencv/modules/dnn/src/tensorflow/tf_graph_simplifier.cpp:1052: error: (-215:Assertion failed) permIds.size() == net.node_size() in function 'sortByExecutionOrder'