A single person against didi Android development team?

This article is the contribution of a friend, he spent a lot of energy to achieve a full function of the didi client, very learning value, recommended to everyone ~

This article is long, it is recommended to collect food; Now created an Android development water friends circle, the circle will not regularly update some Android advanced advanced information, welcome everyone with technical problems to discuss, grow together! (Including senior UI engineer, Android underlying development engineer, Android architect, native performance optimization and hybrid optimization, flutter expertise); The hope has the technology big guy to join, the water circle inside solves the problem more obtains the right bigger!

preface

1. RFID identification and verification function: The RFID identification chip is embedded in the driver’s id card or the identification hardware in the car, and the passenger uses the mobile phone to read the chip information, and through the network (OKHTTP3) sent to the travel platform database for verification (I used NDK plus a C language MD5 encryption algorithm to encrypt the identified information). If it is not a compliant “person” or “car”, it cannot complete the order and report its current location to the platform or regulatory authorities. (In order to facilitate readers’ testing, mobile phones can be used to read any encrypted or non-encrypted RFID chip, such as bank card, bus card, etc. I changed the chip information into my own driver information before verification in the code to ensure that readers can receive a reply from the server when testing.
2. Overseas edition function: click to switch the current language.
3. Driver id number recognition function: read the id number on the driver id, and can also be used to verify and compare with the information in the travel platform database.

The development environment

Main interface overview

Project file structure

• Activities folder: Some activities are independent and do not belong to a function module. You can place them in this folder.
• Adapter folder: More complex Adapters can be removed from the recycler view class file and saved separately to the Adapter folder, such as the Adapter from the Recycler View drop-down menu on the left. The simpler Adapter will still be stored in the calling class.
• Bean folder: Store Entity Entity class, such as driver information will be wrapped into a DriverBean, each driver is a class object, use Gson pass is convenient, use get,set can be used.
• DriverLicenseNFC folder: RFID authentication module used by passengers to verify the driver’s identity or vehicle information.
• DriverLicenseRecognition folder: function module that identifies driver ID number OCR.
• Splash folder: App initialization and boot interface.
• TabFragment: A sliding theme bar at the top of the main screen is used to switch between vehicles or services (Tabs). The codes for different vehicles or services are stored in the TabFragment folder.
• Utils folder: For example, DriverRouteOverlay is used to render the planned driving route on the map. MD5JniUtils uses NDK to call MD5 encryption algorithm to protect RFID chip information. NfcUtils is used to manage THE NFC function of the mobile phone. POIOverlay is used to render surrounding points of interest (POIS) on the map.

Guide the interface

<objectAnimator
    xmlns:android="http://schemas.android.com/apk/res/android"
    android:duration="3000"
    android:propertyName="trimPathEnd"
    android:repeatCount="0"
    android:valueFrom="0"
    android:valueTo="1"
    android:valueType="floatType"/>Copy the code

Bottom slide menu

app:layout_behavior="@string/bottom_sheet_behavior"Copy the code

Left side slide menu

/** * Get the image gradient of driver's id in RGB color space, and then do binarization on this image. Thus, the certificate area was obtained by contour discovery and area size filtering * author:Tantuo 86-1860194917 * @param fileUri * @return*/ public Mat findLicenseContour(Uri fileUri) {// Use openCV Imgcodecs to get the id image from the camera Imgcodecs.imread(fileUri.getPath());if (src.empty()) {
            returnnull; } // get the x gradient and y gradient of the id photo Mat grad_x = new Mat(); Mat grad_y = new Mat(); Mat abs_grad_x = new Mat(); Mat abs_grad_y = new Mat(); // The gradient operation is the derivative operation of pixels in the image. By :Tantuo imgproc. Scharr(SRC, grad_x, cvtype.cv_32f, 1, 0); Imgproc.Scharr(src, grad_y, CvType.CV_32F, 0, 1); // CvType in openCV with 32-bit floating point numbers is used to hold pixel data values that may be negative core.convertScaleabs (grad_x, abs_grad_x); Core.convertScaleAbs(grad_y, abs_grad_y); OpenCV uses release() to release the Mat image and recycle() to release the BitMap image grad_x.release(); grad_y.release(); Mat grad = new Mat(); Mat grad = new Mat(); Mat grad = new Mat(); AddWeighted (abs_grad_x, 0.5, abs_grad_y, 0.5, 0, grad); abs_grad_x.release(); abs_grad_y.release(); Imgproc.cvtcolor binary = new Mat(); // imgproc.cvtcolor binary = new Mat(); Imgproc.cvtColor(grad, binary, Imgproc.COLOR_BGR2GRAY); Imgproc.threshold(binary, binary, 40, 255, imgproc. THRESH_BINARY); grad.release(); // Run the following morphology excution to remove the noise of binary image. First to get the Size of 3 * 3 pixel structure elements Mat kernel = Imgproc. GetStructuringElement (Imgproc MORPH_RECT, new Size (3, 3)); // Then perform the Morph_open operation on the structure elements. Imgproc.morphologyEx(binary, binary, imgproc. MORPH_OPEN, kernel); // Next use openCV's imgproc.findContours () method to find the contour of the driver's id in the image; Mat hierarchy = new Mat(); Imgproc.findContours(binary, contours, hierarchy, Imgproc.RETR_EXTERNAL, Imgproc.CHAIN_APPROX_SIMPLE, new Point(0, 0)); int hw = binary.cols() / 2; int hh = binary.rows() / 2; Rect roi = new Rect();for(int i = 0; i < contours.size(); i++) { Rect rect = Imgproc.boundingRect(contours.get(i)); // If the contour width of a certain ROI interest area is found to be more than half of the picture, it can be considered as the contour of driver's certificateif (rect.width > hw) {
                roi.x = rect.x;
                roi.y = rect.y;
                roi.width = rect.width;
                roi.height = rect.height;
                break; }} // Return if not foundif (roi.width == 0 || roi.height == 0) {
            returnnull; } Mat card = new Mat(); src.submat(roi).copyTo(card); // Remember to release the resource 0 src.release(); binary.release();return card;
    }Copy the code

Public Mat findCardNumBlock(Mat card) {// Initialize HSV color space (H: Hue,S:saturation,V:value brightness, brightness); Mat binary = new Mat(); Imgproc.COLOR_BGR2HSV imgproc. cvtColor(card, HSV, imgproc. COLOR_BGR2HSV); //inThe Range function will filter the root threshold of the HSV color image, which will be used to filter out colors that are not helpful in identifying the marked area in the upper left corner // and save the filtered image in binary // Scalar () is a structure with three parameters, which represent the hue, saturation of HSV, Core. InRange (HSV, New Scalar(30, 40, 45), New Scalar(180, 255, 255), binary); // The above will get a binary image of the driver's id, However, there is a lot of noise. // A morphology excution is carried out on the binary image to remove the small 5*5 structural elements (noise) Mat kernel = Imgproc.getStructuringElement(Imgproc.MORPH_RECT, new Size(5, 5)); Imgproc.morphologyEx(binary, binary, Imgproc.MORPH_OPEN, kernel); List<MatOfPoint> contours = new ArrayList<>(); Mat hierarchy = new Mat(); Imgproc.findContours(binary, contours, hierarchy, Imgproc.RETR_EXTERNAL, Imgproc.CHAIN_APPROX_SIMPLE, new Point(0, 0)); int offsetx = binary.cols() / 3; int offsety = binary.rows() / 3; Rect numberROI = new Rect(); // Find the enclosing rectangle contourArea for each ROIfor(int i = 0; i < contours.size(); i++) { Rect roi = Imgproc.boundingRect(contours.get(i)); // If the identified rectangular area is too small (less than 200 pixels), ignore itif (Imgproc.contourArea(contours.get(i)) < 200) {
                continue; } // Once you have found the mark area, base on the mark area, the id number should be located at the mark x *2 and the width should be about bine.cols () -roi.x -100 pixels // The id number should be about 0.7 times the height of the mark (base) *0.7; // If the outline of the marker in the upper left corner is less than one third of the length and width of the document, the marker will be used as the standard for the numbered areaif(roi.x < offsetx && roi.y < offsety) { numberROI.x = 3* roi.x + 120; numberROI.y = roi.y + 4 * roi.height - 65; numberROI.width = binary.cols() - roi.x - 390; Height = (int) (ROI. Height * 0.9);break; }} // Return null if not foundif (numberROI.width == 0 || numberROI.height == 0) {
            returnnull; } textimage = new Mat(); card.submat(numberROI).copyTo(textImage); Release mat resource card.release(); hsv.release(); binary.release();return textImage;
    }Copy the code

Imgproc.cvtColor(card, hsv, Imgproc.COLOR_BGR2HSV);Copy the code

Core.inRange(hsv, new Scalar(30, 40, 45), new Scalar(180, 255, 255), binary);Copy the code

// The following morphology excution is carried out on the binarized image to remove the small 5*5 structural elements (noise) Mat kernel = Imgproc.getStructuringElement(Imgproc.MORPH_RECT, new Size(5, 5)); Imgproc.morphologyEx(binary, binary, Imgproc.MORPH_OPEN, kernel); After noise processing, start to find the number area in the ID card area to find the Contour operation: // Find the outer rectangle contourArea for each ROIfor(int i = 0; i < contours.size(); i++) { Rect roi = Imgproc.boundingRect(contours.get(i)); // If the identified rectangular area is too small (less than 200 pixels), ignore itif (Imgproc.contourArea(contours.get(i)) < 200) {
                continue; } // Once you have found the mark area, base on the mark area, the id number should be located at the mark x *2 and the width should be about bine.cols () -roi.x -100 pixels // The id number should be about 0.7 times the height of the mark (base) *0.7; // If the outline of the marker in the upper left corner is less than one third of the length and width of the document, the marker will be used as the standard for the numbered areaif(roi.x < offsetx && roi.y < offsety) { numberROI.x = 3* roi.x + 120; numberROI.y = roi.y + 4 * roi.height - 65; numberROI.width = binary.cols() - roi.x - 390; Height = (int) (ROI. Height * 0.9);break; }}Copy the code

Mat textImage = new Mat(); card.submat(numberROI).copyTo(textImage); Release mat resource card.release(); hsv.release(); binary.release();return textImage;Copy the code

• The splitNumberBlock(Mat textImage) method uses binarization to identify character Outlines in numeric regions.
• The getSplitLinePos(Mat Mtexts) method is used to separate two numbers bonded together in the image.
• The function of extractFeatureData(Mat txtImage) method is to extract the feature of card number recognition, and obtain the black pixel feature of each number of card number, which is an important basis for the feature and recognition of each number.
• The dumpFeature(float[] fv, int textNum) method saves the generated feature text file on the phone.
• ReadFeatureVector (File f) is used to read saved feature vectors.
• RecognitionChar (Mat charImage) is used to identify the id number according to the feature vector.

@Override
    protected void onResume() { super.onResume(); // Foreground distribution system, used to ensure the highest capture priority when detecting RFID tags. NfcUtils.mNfcAdapter.enableForegroundDispatch(this, NfcUtils.mPendingIntent, NfcUtils.mIntentFilter, NfcUtils.mTechList);
    }
    @Override
    protected void onPause() { super.onPause(); // Shut down the foreground scheduling system NfcUtils.mNfcAdapter.disableForegroundDispatch(this);
    }Copy the code

The NDK is used to invoke the MD5 encryption algorithm

I need more video and document information of the private message, more advanced and interview information please enter the circle to receive, welcome to rational discussion, have questions welcome to leave a message to discuss.