munhra

import firebase from "firebase";

const config = {
  apiKey: "Minha Chave Tira o Zóio",
  authDomain: "iplantacao-72b97.firebaseapp.com",
  databaseURL: "https://iplantacao-72b97.firebaseio.com",
  projectId: "iplantacao-72b97",
  storageBucket: "iplantacao-72b97.appspot.com",
  messagingSenderId: "Tira o Zoio"
};

munhra

private void savePrediction(Prediction prediction) {
        Log.d(TAG,"savePrediction");
        FirebaseDatabase firebaseDatabase = FirebaseDatabase.getInstance();
        DatabaseReference databaseReference = firebaseDatabase.getReference();
        databaseReference.child("prediction").child(prediction.getId()).setValue(prediction);
}

private void uploadImage(String absolutePath, final Prediction prediction) {

        Uri file = Uri.fromFile(new File(absolutePath));

munhra

 public void runModel(String absolutePath) {
        if (mInterpreter == null) {
            Log.e(TAG, "Image classifier has not been initialized; Skipped.");
            return;
        }

        float[][][][] imgData = cropImage(absolutePath);

        try {
            FirebaseModelInputs inputs = new FirebaseModelInputs.Builder().add(imgData).build();

munhra

private float[][][][] cropImage(String absolutePath) {

        Mat imageRaw = Imgcodecs.imread(absolutePath);


        float[][][][] imgData = new float [DIM_BATCH_SIZE] [DIM_IMG_SIZE_X] [DIM_IMG_SIZE_Y] [DIM_PIXEL_SIZE];

        int square_qtd_x = 20;
        int square_qtd_y = 20;

munhra

from __future__ import print_function
import keras
#from keras.datasets import mnist
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras import backend as K
import numpy as np
import ml_bullgreen_dataset_handler
from sklearn import cross_validation

munhra

model = Sequential()
    model.add(Conv2D(32, kernel_size=(3, 3),
                 activation='relu',
                 input_shape=input_shape))
    model.add(Conv2D(64, (3, 3), activation='relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.40)) #0.4 removed as it is bad of tflite # 0.25
    model.add(Flatten())
    model.add(Dense(128, activation='relu'))
    model.add(Dropout(0.25)) #0.25 removed as it is bad of tflite # 0.5

munhra

def __create_feature_vector_raw(self, file_name, piquete_id, score, flatten=False):

        sample_size = (self.CROP_WIDTH, self.CROP_HEIGHT, 3)
        image_path = self.__image_files_root_folder+ str(piquete_id)+'/'+file_name
        print(image_path)
        image_raw = cv2.imread(image_path)

        if image_raw is not None:

            if not self.__check_for_bad_images(file_name, piquete_id, score):

munhra

    def runCrossValidation(self, farm_dataset):
        print('start cross validation')
        x_train, x_test, y_train, y_test = cross_validation.train_test_split(
            farm_dataset.feature_vector,farm_dataset.target, test_size=0.15)
        
        #x_train_scaled = preprocessing.scale(x_train)
        #y_train_scaled = preprocessing.scale(y_train)
        #x_test_scaled = preprocessing.scale(x_test)
        #y_test_scaled = preprocessing.scale(y_test)
        

munhra

def __create_feature_vector_mean(self, file_name, piquete_id, score, height):
        rgb_mean = [0] * 3
        image_path = self.__image_files_root_folder+ str(piquete_id)+'/'+file_name
        image = cv2.imread(image_path)
        print(image_path)
        means = cv2.mean(image)
        
        if means is not None:
            #raw = image.flatten()
            print(str(means[:3])+'\n')

munhra

p{font-size: 14px}

article ~ p {font-size: 14px}

article ~ p {color:#00F;}

p+p {text-indent:20px;} /*pega o proximo igual*/

li:first-child {font-size: 20px;} /*pega o primeiro filho*/