Number Plate Detection

Number Plate Detection

For Number Plate Detection, the pre-trained models used are

  • Pytesseract
  • Trained Neural Network

(info) Notes:

  • Pytesseract is an independent library inside Python for text detection in images. It can detect alphanumeric as well as special characters from the text.
  • Trained Neural Network is pre-trained using textual data. You can feed an image to the model and attempt to derive text from it. The detection of alphanumeric and special characters depends on the purpose for which the neural networks are trained.|

The available properties of Number Plate Detection are as shown in the figure given below.


Using the Properties pane, you can select

  • A suitable name for the task
  • Vehicle Image files for analysis from the Input Image Column (X)
  • Number Plate Labels for analysis from the Input Label Column (Y)
  • Model to be used for analysis

(info)

Notes:

  • If you select only X and not Y, you get only the predicted number plates. In that case, the results might not be accurate.
  • If you select both X and Y, you get more detailed results. However, you have to select an equal number of features in both X and Y in this case.
In the output, you get
    • An image selected for number plate detection and the predicted number on its plate
    • Performance Metrics for
    • The displayed image

    • Applied model

(info) Note:

For the vehicle image displayed on the Result page, each corresponding character in the Actual Number Plate and Predicted Number Plate is compared to detect the Correctly Predicted Characters. Character Error Rate value gives the rate of the wrongly predicted characters.

For the applied model, the results include the Total Number Plates and the Correctly Predicted Number Plates out of them. They also include the Overall Accuracy and the Overall Character Error.
In the results, you can see that,

  • For the displayed image 1 (out of 3), all ten characters of the number plate are correctly predicted, and the character error rate is zero.
  • For the applied model,
  • Three (3) number plates are analyzed for detection.
  • Out of them, two (2) number plates are correctly predicted.
  • The overall accuracy is 66.6667%
  • The overall character error is 3.7033.

    • Related Articles

    • Number Plate Detection

      For Number Plate Detection, the pre-trained models used are Pytesseract Trained Neural Network Notes: Pytesseract is an independent library inside Python for text detection in images. It can detect alphanumeric as well as special characters from the ...

    • Face Detection

      For Face Detection, the algorithm libraries used are MTCNN (Multi-task Cascaded Convolutional Neural Networks) RetinaFace Notes: MTCNN is a faster algorithm compared to RetinaFace. It can be used in the case of images with many faces and where high ...

    • Face Detection

      For Face Detection, the algorithm libraries used are MTCNN (Multi-task Cascaded Convolutional Neural Networks) RetinaFace Notes: MTCNN is a faster algorithm compared to RetinaFace. It can be used in the case of images with many faces and where high ...

    • Outlier Detection

      Outlier Detection is located under Model Studio in Data Preparation, in the task pane on the left. Use drag-and-drop method to use algorithm in the canvas. Click the algorithm to view and select different properties for analysis. Refer to Properties ...

    • Outlier Detection

      Outlier Detection is located under Model Studio ( ) in Data Preparation, in the task pane on the left. Use drag-and-drop method to use algorithm in the canvas. Click the algorithm to view and select different properties for analysis. Refer to ...