supervised_classification_utils.py

from sklearn import *
from unsupervised_classification_utils import *


def fitSupervised(target, features):
    # choose the model you want
    model = input('Choose your supervised model : SVM, NeuralNetwork')

    if model == 'SVM':
        # model fit with target and features data
        clf = svm.SVC(kernel='linear', C=1).fit(features, target)
        # score of a training prediction
        score = clf.score(features, target)
        print('score : ', score)

    if model == 'NeuralNetwork':
        # model fit with target and features data
        clf = neural_network.MLPClassifier(solver='lbfgs', alpha=1e-5, hidden_layer_sizes=(5, 2), random_state=1)
        clf.fit(features, target)
        # score of a training prediction
        score = clf.score(features, target)
        print('score : ', score)

    return clf


def supervised_classification_Results(outputDirectory,shapeFileDirectory,csvPath):
    dictionary = imageDictionary(outputDirectory, shapeFileDirectory)

    target, features = target_features(outputDirectory, csvPath, dictionary)

    model = fitSupervised(target, features)

    date, shapeFile = selectParameters(dictionary)

    dataset = fillDataset(shapeFile, date, dictionary, outputDirectory)

    separation = earthAndSea(dataset)

    sea = seaDataset(dataset, separation, invert=True) * 65535

    prediction = model.predict(sea)

    # Recover the full dataset : earth,sea,class
    # add a column prediction
    sea.insert(sea.shape[1], "Prediction", prediction, allow_duplicates=False)
    # add all rows corresponding to earth by reindexing with 0 value
    index = np.arange(0, dataset.shape[0], 1)
    data = sea.reindex(index, fill_value=0)

    source = outputDirectory + "/" + date + "/" + shapeFile + "/" + dictionary[date][shapeFile][0]

    with rasterio.open(source) as src:
        img = src.read(1)
        shape = img.shape
    pred = data['Prediction'].to_numpy()
    img_classified = pred.reshape(shape[0], shape[1])

    plt.figure()
    plt.imshow(img_classified)
    plt.title("herbier : supervised (SVM)")