How to access the license plate recognition API for iOS development?

IOS develops license plate recognition

I did some character recognition before. Now that I think about it, it should also be effective for car license plate recognition, so write something for beginners to try, license plate recognition.

The first part, call the system camera to take pictures,

The camera here doesn't need too many functions, so it uses UIImagePickerController directly, which provides simple and convenient functions of taking pictures and selecting pictures in the photo library. After turning on the camera, you need to call the method in UIImagePicker Controller Delegate, the method to be executed after shooting and the method to be executed after clicking Cancel.

Write a picture here.

The second part, select the part to be used and correct the picture.

Generally, the pictures taken are crooked. At this time, it is necessary to correct the needed parts of the picture. The technology used at this time is OpenCV source computer vision library. Display the front of the picture on the screen through opencv.

Figure 12

The third part, image capture.

There are two ways to capture images. A popular method is to capture the blue, yellow, black, white and part of the bottom of the license plate with colors. I haven't found a similar method for iOS (someone found it to remind me). My method is to cover the picture with a layer by using rubber technology. The user selects the range of the picture by daubing, and records the minimum X and Y and the maximum height and weight of the operation position by intercepting the picture by gesture. The user manually finds the range of the picture.

Figure 13

The fourth part, OCR recognition.

Character recognition technology, using TesseractSample, recognizes the vehicle license plate number, and the recognition error will be relatively large. After all, it is only a test, so a UITextField is provided, and the recognized results are displayed in the UITextField. If there is an error, you can click Edit at any time. Make corrections.

Figure 14

Github address:

/Wang Jianchao 1990/ Gourmet