def save_tf():
STRIDES, ANCHORS, NUM_CLASS, XYSCALE = utils.load_config()
input_layer = tf.keras.layers.Input(
[FLAGS.input_size, FLAGS.input_size, 3])
feature_maps = YOLOv4(input_layer, NUM_CLASS)
bbox_tensors = []
prob_tensors = []
for i, fm in enumerate(feature_maps):
if i == 0:
output_tensors = decode_tf(fm, FLAGS.input_size // 8, NUM_CLASS,
STRIDES, ANCHORS, i, XYSCALE)
elif i == 1:
output_tensors = decode_tf(fm, FLAGS.input_size // 16, NUM_CLASS,
STRIDES, ANCHORS, i, XYSCALE)
else:
output_tensors = decode_tf(fm, FLAGS.input_size // 32, NUM_CLASS,
STRIDES, ANCHORS, i, XYSCALE)
bbox_tensors.append(output_tensors[0])
prob_tensors.append(output_tensors[1])
pred_bbox = tf.concat(bbox_tensors, axis=1)
pred_prob = tf.concat(prob_tensors, axis=1)
boxes, pred_conf = filter_boxes(pred_bbox,
pred_prob,
score_threshold=FLAGS.score_thres,
input_shape=tf.constant(
[FLAGS.input_size, FLAGS.input_size]))
pred = tf.concat([boxes, pred_conf], axis=-1)
model = tf.keras.Model(input_layer, pred)
utils.load_weights(model, FLAGS.weights)
model.summary()
model.save(FLAGS.output)
def main(_argv):
input_layer = tf.keras.layers.Input([FLAGS.size, FLAGS.size, 3])
feature_maps = YOLOv4(input_layer, NUM_CLASS)
bbox_tensors = []
for i, fm in enumerate(feature_maps):
bbox_tensor = decode(fm, NUM_CLASS, i)
bbox_tensors.append(bbox_tensor)
model = tf.keras.Model(input_layer, bbox_tensors)
utils.load_weights(model, 'data/YOLOv4-obj_1000.weights')
vid = cv2.VideoCapture(FLAGS.input) # Reading input
return_value, frame = vid.read()
fourcc = cv2.VideoWriter_fourcc('F', 'M', 'P', '4')
out = cv2.VideoWriter(FLAGS.output, fourcc, 10.0,
(frame.shape[1], frame.shape[0]), True)
plates = []
n = 0
Sum = 0
while True:
start = time.time()
n += 1
return_value, frame = vid.read()
if frame is None:
continue
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
bboxes = plateDetect(frame, FLAGS.size,
model) # License plate detection
for i in range(len(bboxes)):
img = frame[int(bboxes[i][1]):int(bboxes[i][3]),
int(bboxes[i][0]):int(bboxes[i][2])]
prediction_groups = pipeline.recognize(
[img]) # Text detection and recognition on license plate
string = ''
for j in range(len(prediction_groups[0])):
string = string + prediction_groups[0][j][0].upper()
if platePattern(string) == True and string not in plates:
plates.append(string)
if len(plates) > 0:
drawText(frame, plates)
frame = utils.draw_bbox(
frame, bboxes) # Draws bounding box around license plate
frame = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
Sum += time.time() - start
print('Avg fps:- ', Sum / n)
out.write(frame)
cv2.imshow("result", frame)
if cv2.waitKey(1) == 27: break
out.release()
cv2.destroyAllWindows()
def main(_argv):
print('main')
input_layer = tf.keras.layers.Input([FLAGS.size, FLAGS.size, 3])
feature_maps = YOLOv4(input_layer, NUM_CLASS)
bbox_tensors = []
for i, fm in enumerate(feature_maps):
bbox_tensor = decode(fm, NUM_CLASS, i)
bbox_tensors.append(bbox_tensor)
model = tf.keras.Model(input_layer, bbox_tensors)
utils.load_weights(model, 'data/YOLOv4-obj_1000.weights')
img = cv2.imread(FLAGS.input) # Reading input
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
bboxes = plateDetect(img, FLAGS.size, model) # License plate detection
plates = []
for i in range(len(bboxes)):
plate_img = img[int(bboxes[i][1]):int(bboxes[i][3]),
int(bboxes[i][0]):int(bboxes[i][2])]
prediction_groups = pipeline.recognize(
[plate_img]) # Text detection and recognition on license plate
string = ''
for j in range(len(prediction_groups[0])):
string = string + prediction_groups[0][j][0].upper()
if platePattern(string) == True and string not in plates:
plates.append(string)
if len(plates) > 0:
drawText(img, plates)
img = utils.draw_bbox(img,
bboxes) # Draws bounding box around license plate
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
cv2.imwrite(FLAGS.output, img) # Saving output
print('Output saved to ', FLAGS.output)
cv2.imshow("result", img)
cv2.waitKey(0)