def detect(signal_queue, key, sensors_queue):
detector = detection.Detector()
while True:
time.sleep(0.001)
# update sensors data
while True:
try:
(
detector.time,
# detector.gpsRaw_position,
# detector.gpsRaw_speed,
detector.compassRaw,
detector.camerasRaw), frame = sensors_queue.get(
block=True, timeout=0.001)
timer = time.time()
detector.process()
# send all signals except object detection to decider
signal_queue.put((detector.signals, frame))
# detectedInfo(' '.join(
# ['Detector frame:', str(frame), 'stage1: spend', str(int((time.time() - timer) * 1000)), 'ms']
# ))
timer = time.time()
detector.object_detection()
# object detection updated
signal_queue.put((detector.signals, frame))
# detectedInfo(' '.join(
# ['Detector frame:', str(frame), 'stage2: spend', str(int((time.time() - timer) * 1000)), 'ms']
# ))
except queue.Empty:
pass
def __init__(self, detector=None, recognizer=None, scale=2, max_size=2048):
if detector is None:
detector = detection.Detector()
if recognizer is None:
recognizer = recognition.Recognizer()
self.scale = scale
self.detector = detector
self.recognizer = recognizer
self.max_size = max_size
import psycopg2
import sys
import detection
conn = psycopg2.connect('dbname=gis user=gis')
run = "iaicoco"
sensitives = ("car", "person", "truck")
cur = conn.cursor()
cur.execute(
"SELECT id FROM panoramas WHERE authorised=0 AND id >= 9686 ORDER BY id")
results = cur.fetchall()
if (len(results) == 0):
print("No Results")
else:
prob = 25
if (len(sys.argv) > 1):
prob = int(sys.argv[1])
detector = detection.Detector(prob, run)
for result in results:
detector.detect(result[0])
def output(img_path, pipe, out_path):
img_path = img_path[0]
img = cv2.imread(img_path)
# img = cv2.resize(img, (img.shape[1] // 2, img.shape[0] // 2))
predictions = pipe.recognize(images=[img])[0]
drawn = tools.drawBoxes(
image=img, boxes=predictions, boxes_format='predictions'
)
print(
'Predicted:', [text for text, box in predictions]
)
cv2.imwrite(out_path, drawn)
validation = get_ocr_detector_dataset('test')
generator_kwargs = {'width': 640, 'height': 640}
validation_image_generator = datasets.get_detector_image_generator(
labels=validation,
**generator_kwargs
)
detector = detection.Detector()
detector.model.load_weights(f'model/{args.model}.h5')
pipe = pipeline.Pipeline(detector=detector)
for i, img in enumerate(validation):
print('sdfsdf')
print(img[0])
output(img, pipe, f'output_{i}.png')