# otherwise load video
else:
camera = cv2.VideoCapture(args["video"])
# keep looping over frames
while True:
# grab frame
(grabbed, frame) = camera.read()
# check for last frame
if not grabbed:
break
# detect barcode
box = simple_barcode_detection.detect(frame)
# draw bounding box
cv2.drawContours(frame, box, -1, (0, 255, 0), 3)
# show frame, record if key pressed
cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF
# if q, then quit
if key == ord("q"):
break
# cleanup and close windows
camera.release()
cv2.destroyAllWindows()
time.sleep(2.0)
else:
vs = cv2.VideoCapture(args["video"])
while True:
frame = vs.read()
frame = frame[1] if args.get("video", False) else frame
if frame is None:
break
# detect the barcode in the image
box = detect(frame)
#croped=croping(frame,box)
#frame=read_barcodes(box)
if box is not None:
cv2.drawContours(frame, [box], -1, (0, 255, 0), 2)
cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF
if key == ord("q"):
break
camera = cv2.VideoCapture(0)
# otherwise, load the video
else:
camera = cv2.VideoCapture(args["video"])
# keep looping over the frames
while True:
# grab the current frame
(grabbed, frame) = camera.read()
# check to see if we have reached the end of the video
if not grabbed:
break
# detect the barcode in the image
box = simple_barcode_detection.detect(frame)
# if a barcode was found, draw a bounding box on the frame
cv2.drawContours(frame, [box], -1, (0,255,0), 2)
# show the frame and record if the user presses a key
cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF
# if the 'q' key is pressed, stop the loop
if key == ord("q"):
break
# cleanup the camera and close any open windows
camera.release()
cv2.destroyAllWindows()
def train(do_train):
# with tf.name_scope("input"):
x = tf.placeholder(tf.float32, [None, 100, 100], name="x_input")
_y = tf.placeholder(tf.float32, [None, 1], name="y_input")
_x = tf.placeholder(tf.float32, [None, 1], name="predtion")
_pred = _CNN(x)['output']
# 计算最终输出与标准之间的loss
# 把均方误差也加入到集合里
# tf.add_to_collection("losses", tf.reduce_mean(tf.square(_pred - _y)))
# cost = tf.add_n(tf.get_collection("losses"),name="loss")
cost = tf.reduce_mean(tf.square(_pred - _y))
#学习率呈指数下降
# learning_rate = tf.train.exponential_decay(learning_rate=0.1,decay_rate=0.9,global_step = tf.Variable(0),decay_steps=10)
#随机梯度下降算法
# optm = tf.train.GradientDescentOptimizer(learning_rate=0.1).minimize(cost)
#Adam算法
optm = tf.train.AdamOptimizer(learning_rate=0.001).minimize(cost)
_corr = tf.equal(_x, _y)
accr = tf.reduce_mean(tf.cast(_corr, tf.float32))
sess = tf.Session()
sess.run(tf.global_variables_initializer())
# 训练次数
train_epochs = 100
# 保存model
saver = tf.train.Saver(max_to_keep=1)
# # P处理大小
# batch_size = 50
if do_train == 0:
loss = 1
for epoch in range(train_epochs):
x_data, y_data = init_data(epoch, 35)
batch_x = x_data
batch_y = y_data.reshape(len(y_data), 1)
sess.run(optm, feed_dict={x: batch_x, _y: batch_y})
avg_cost = sess.run(cost, feed_dict={x: batch_x, _y: batch_y})
# print(pred)
print("Epoch %03d/%03d" % (epoch, train_epochs))
print("Cost:" + str(avg_cost))
if avg_cost < loss:
now = time.strftime('%Y-%m-%d-%H')
loss = avg_cost
saver.save(
sess,
'/home/zx/qrcode/model/model-' + now + '-' + str(epoch))
if do_train == 1:
saver.restore(sess, '/home/zx/qrcode/model/model-2018-04-19-18-64')
print("Read Model Ready!")
train_acc = 0
test_epochs = 3
#p处理张数
pnum = 50
for i in range(test_epochs):
x_data, y_data = test_data(i, pnum)
y_data = y_data.reshape(len(y_data), 1)
prediction = sess.run(_pred, feed_dict={x: x_data})
for i in range(len(prediction)):
if prediction[i, ] > 0.5:
prediction[i, ] = 1
else:
prediction[i, ] = 0
train_acc = train_acc + sess.run(accr,
feed_dict={
_x: prediction,
_y: y_data
})
train_acc = train_acc / test_epochs
print("Accuracy: " + str(train_acc))
if do_train == 2:
saver.restore(sess, '/home/zx/qrcode/model/model-2018-04-19-18-64')
print("Read Model Ready!")
img = cv2.imread("/home/zx/qrcode/data/_test/no2.jpg")
# img = cv2.imread("/home/zx/qrcode/data/_test/no3.jpg", cv2.IMREAD_GRAYSCALE)
if img is not None:
img = cv2.resize(img, (399, 260))
cropImg = detect(img)
# cv2.imshow("cropImg",cropImg)
# cv2.waitKey(0)
cropImg = cv2.resize(cropImg, (100, 100))
cv2.imwrite("/home/zx/qrcode/data/_test/demo/0.jpg", cropImg)
cropImg = cropImg.reshape(1, 100, 100) // 255
print("Load Image Ready!")
prediction = sess.run(_pred, feed_dict={x: cropImg})
print(prediction[0, ][0])
if prediction[0, ] > 0.5:
print("是")
else:
print("不是")