You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

286 lines
9.6 KiB

## License: Apache 2.0. See LICENSE file in root directory.
## Copyright(c) 2021 Intel Corporation. All Rights Reserved.
#####################################################
## auto calibration ##
#####################################################
import argparse
import json
import sys
import time
import pyrealsense2 as rs
__desc__ = """
This script demonstrates usage of Self-Calibration (UCAL) APIs
"""
# mappings
occ_speed_map = {
'very_fast': 0,
'fast': 1,
'medium': 2,
'slow': 3,
'wall': 4,
}
tare_accuracy_map = {
'very_high': 0,
'high': 1,
'medium': 2,
'low': 3,
}
scan_map = {
'intrinsic': 0,
'extrinsic': 1,
}
fl_adjust_map = {
'right_only': 0,
'both_sides': 1
}
ctx = rs.context()
def main(arguments=None):
args = parse_arguments(arguments)
try:
device = ctx.query_devices()[0]
except IndexError:
print('Device is not connected')
sys.exit(1)
# Verify Preconditions:
# 1. The script is applicable for D400-series devices only
cam_name = device.get_info(rs.camera_info.name) if device.supports(rs.camera_info.name) else "Unrecognized camera"
if device.supports(rs.camera_info.product_line):
device_product_line = str(device.get_info(rs.camera_info.product_line))
if device_product_line != 'D400':
print(f'The example is intended for RealSense D400 Depth cameras, and is not', end =" ")
print(f'applicable with {cam_name}')
sys.exit(1)
# 2. The routine assumes USB3 connection type
# In case of USB2 connection, the streaming profiles should be readjusted
if device.supports(rs.camera_info.usb_type_descriptor):
usb_type = device.get_info(rs.camera_info.usb_type_descriptor)
if not usb_type.startswith('3.'):
print('The script is designed to run with USB3 connection type.')
print('In order to enable it with USB2.1 mode the fps rates for the Focal Length and Ground Truth calculation stages should be re-adjusted')
sys.exit(1)
# prepare device
depth_sensor = device.first_depth_sensor()
depth_sensor.set_option(rs.option.emitter_enabled, 0)
if depth_sensor.supports(rs.option.thermal_compensation):
depth_sensor.set_option(rs.option.thermal_compensation, 0)
if args.exposure == 'auto':
depth_sensor.set_option(rs.option.enable_auto_exposure, 1)
else:
depth_sensor.set_option(rs.option.enable_auto_exposure, 0)
depth_sensor.set_option(rs.option.exposure, int(args.exposure))
print("Starting UCAL...")
try:
# The recomended sequence of procedures: On-Chip -> Focal Length -> Tare Calibration
run_on_chip_calibration(args.onchip_speed, args.onchip_scan)
run_focal_length_calibration((args.target_width, args.target_height), args.focal_adjustment)
run_tare_calibration(args.tare_accuracy, args.tare_scan, args.tare_gt, (args.target_width, args.target_height))
finally:
if depth_sensor.supports(rs.option.thermal_compensation):
depth_sensor.set_option(rs.option.thermal_compensation, 1)
print("UCAL finished successfully")
def progress_callback(progress):
print(f'\rProgress {progress}% ... ', end ="\r")
def run_on_chip_calibration(speed, scan):
data = {
'calib type': 0,
'speed': occ_speed_map[speed],
'scan parameter': scan_map[scan],
'white_wall_mode': 1 if speed == 'wall' else 0,
}
args = json.dumps(data)
cfg = rs.config()
cfg.enable_stream(rs.stream.depth, 256, 144, rs.format.z16, 90)
pipe = rs.pipeline(ctx)
pp = pipe.start(cfg)
dev = pp.get_device()
try:
print('Starting On-Chip calibration...')
print(f'\tSpeed:\t{speed}')
print(f'\tScan:\t{scan}')
adev = dev.as_auto_calibrated_device()
table, health = adev.run_on_chip_calibration(args, progress_callback, 30000)
print('On-Chip calibration finished')
print(f'\tHealth: {health}')
adev.set_calibration_table(table)
adev.write_calibration()
finally:
pipe.stop()
def run_focal_length_calibration(target_size, adjust_side):
number_of_images = 25
timeout_s = 30
cfg = rs.config()
cfg.enable_stream(rs.stream.infrared, 1, 1280, 720, rs.format.y8, 30)
cfg.enable_stream(rs.stream.infrared, 2, 1280, 720, rs.format.y8, 30)
lq = rs.frame_queue(capacity=number_of_images, keep_frames=True)
rq = rs.frame_queue(capacity=number_of_images, keep_frames=True)
counter = 0
flags = [False, False]
def cb(frame):
nonlocal counter, flags
if counter > number_of_images:
return
for f in frame.as_frameset():
p = f.get_profile()
if p.stream_index() == 1:
lq.enqueue(f)
flags[0] = True
if p.stream_index() == 2:
rq.enqueue(f)
flags[1] = True
if all(flags):
counter += 1
flags = [False, False]
pipe = rs.pipeline(ctx)
pp = pipe.start(cfg, cb)
dev = pp.get_device()
try:
print('Starting Focal-Length calibration...')
print(f'\tTarget Size:\t{target_size}')
print(f'\tSide Adjustment:\t{adjust_side}')
stime = time.time()
while counter < number_of_images:
time.sleep(0.5)
if timeout_s < (time.time() - stime):
raise RuntimeError(f"Failed to capture {number_of_images} frames in {timeout_s} seconds, got only {counter} frames")
adev = dev.as_auto_calibrated_device()
table, ratio, angle = adev.run_focal_length_calibration(lq, rq, target_size[0], target_size[1],
fl_adjust_map[adjust_side],progress_callback)
print('Focal-Length calibration finished')
print(f'\tRatio:\t{ratio}')
print(f'\tAngle:\t{angle}')
adev.set_calibration_table(table)
adev.write_calibration()
finally:
pipe.stop()
def run_tare_calibration(accuracy, scan, gt, target_size):
data = {'scan parameter': scan_map[scan],
'accuracy': tare_accuracy_map[accuracy],
}
args = json.dumps(data)
print('Starting Tare calibration...')
if gt == 'auto':
target_z = calculate_target_z(target_size)
else:
target_z = float(gt)
cfg = rs.config()
cfg.enable_stream(rs.stream.depth, 256, 144, rs.format.z16, 90)
pipe = rs.pipeline(ctx)
pp = pipe.start(cfg)
dev = pp.get_device()
try:
print(f'\tGround Truth:\t{target_z}')
print(f'\tAccuracy:\t{accuracy}')
print(f'\tScan:\t{scan}')
adev = dev.as_auto_calibrated_device()
table = adev.run_tare_calibration(target_z, args, progress_callback, 30000)
print('Tare calibration finished')
adev.set_calibration_table(table)
adev.write_calibration()
finally:
pipe.stop()
def calculate_target_z(target_size):
number_of_images = 50 # The required number of frames is 10+
timeout_s = 30
cfg = rs.config()
cfg.enable_stream(rs.stream.infrared, 1, 1280, 720, rs.format.y8, 30)
q = rs.frame_queue(capacity=number_of_images, keep_frames=True)
# Frame queues q2, q3 should be left empty. Provision for future enhancements.
q2 = rs.frame_queue(capacity=number_of_images, keep_frames=True)
q3 = rs.frame_queue(capacity=number_of_images, keep_frames=True)
counter = 0
def cb(frame):
nonlocal counter
if counter > number_of_images:
return
for f in frame.as_frameset():
q.enqueue(f)
counter += 1
pipe = rs.pipeline(ctx)
pp = pipe.start(cfg, cb)
dev = pp.get_device()
try:
stime = time.time()
while counter < number_of_images:
time.sleep(0.5)
if timeout_s < (time.time() - stime):
raise RuntimeError(f"Failed to capture {number_of_images} frames in {timeout_s} seconds, got only {counter} frames")
adev = dev.as_auto_calibrated_device()
print('Calculating distance to target...')
print(f'\tTarget Size:\t{target_size}')
target_z = adev.calculate_target_z(q, q2, q3, target_size[0], target_size[1])
print(f'Calculated distance to target is {target_z}')
finally:
pipe.stop()
return target_z
def parse_arguments(args):
parser = argparse.ArgumentParser(description=__desc__)
parser.add_argument('--exposure', default='auto', help="Exposure value or 'auto' to use auto exposure")
parser.add_argument('--target-width', default=175, type=int, help='The target width')
parser.add_argument('--target-height', default=100, type=int, help='The target height')
parser.add_argument('--onchip-speed', default='medium', choices=occ_speed_map.keys(), help='On-Chip speed')
parser.add_argument('--onchip-scan', choices=scan_map.keys(), default='intrinsic', help='On-Chip scan')
parser.add_argument('--focal-adjustment', choices=fl_adjust_map.keys(), default='right_only',
help='Focal-Length adjustment')
parser.add_argument('--tare-gt', default='auto',
help="Target ground truth, set 'auto' to calculate using target size"
"or the distance to target in mm to use a custom value")
parser.add_argument('--tare-accuracy', choices=tare_accuracy_map.keys(), default='medium', help='Tare accuracy')
parser.add_argument('--tare-scan', choices=scan_map.keys(), default='intrinsic', help='Tare scan')
return parser.parse_args(args)
if __name__ == '__main__':
main()