Untitled

from flask import Flask
from flask import request, jsonify
from flask_cors import CORS, cross_origin
import numpy as np
import cv2
import base64

# Khai bao cong cua server
my_port = '8000'
scale = 0.00392
conf_threshold = 0.5
nms_threshold = 0.4

# Doan ma khoi tao server
app = Flask(__name__)
CORS(app)

# Cac ham ho tro chay YOLO

def get_output_layers(net):
    layer_names = net.getLayerNames()
    output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
    return output_layers


def build_return(class_id, x, y, x_plus_w, y_plus_h):
    return str(class_id) + "," + str(x) + "," + str(y) + "," + str(x_plus_w) + "," + str(y_plus_h)


# Khoi tao model YOLO
net = cv2.dnn.readNet("yolov3.weights", "yolov3.cfg")


# Khai bao ham xu ly request index
@app.route('/')
@cross_origin()
def index():
    return "Welcome to flask API!"

# Khai bao ham xu ly request hello_word
@app.route('/hello_world', methods=['GET'])
@cross_origin()
def hello_world():
    # Lay staff id cua client gui len
    staff_id = request.args.get('staff_id')
    # Tra ve cau chao Hello
    return "Hello "  + str(staff_id)

# Khai bao ham xu ly request detect
@app.route('/detect', methods=['POST'])
@cross_origin()
def detect():

    # Lay du lieu image B64 gui len va chuyen thanh image
    image_b64 = request.form.get('image')
    image = np.fromstring(base64.b64decode(image_b64), dtype=np.uint8)
    image = cv2.imdecode(image, cv2.IMREAD_ANYCOLOR)

    # Lay kich thuoc anh gui len
    Width = image.shape[1]
    Height = image.shape[0]

    # Nhan dien bang YOLO

    blob = cv2.dnn.blobFromImage(image, scale, (416, 416), (0, 0, 0), True, crop=False)
    net.setInput(blob)
    outs = net.forward(get_output_layers(net))

    class_ids = []
    confidences = []
    boxes = []

    for out in outs:
        for detection in out:
            scores = detection[5:]
            class_id = np.argmax(scores)
            confidence = scores[class_id]
            if confidence > conf_threshold:
                center_x = int(detection[0] * Width)
                center_y = int(detection[1] * Height)
                w = int(detection[2] * Width)
                h = int(detection[3] * Height)
                x = center_x - w / 2
                y = center_y - h / 2
                class_ids.append(class_id)
                confidences.append(float(confidence))
                boxes.append([x, y, w, h])

    indices = cv2.dnn.NMSBoxes(boxes, confidences, conf_threshold, nms_threshold)
    retString = ""

    for i in indices:
        i = i[0]
        box = boxes[i]
        x = box[0]
        y = box[1]
        w = box[2]
        h = box[3]
        # Xay dung chuoi tra ve client
        retString += build_return(class_ids[i], round(x), round(y), round( w), round( h)) + "|"

    return retString;


# Thuc thi server
if __name__ == '__main__':
    app.run(debug=True, host='0.0.0.0',port=my_port)