# Copyright 2014 The Android Open Source Project # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """Verifies RAW streams are not croppable.""" import logging import math import os.path from mobly import test_runner import numpy as np import its_base_test import camera_properties_utils import capture_request_utils import image_processing_utils import its_session_utils import target_exposure_utils _CROP_FULL_ERROR_THRESHOLD = 3 # pixels _CROP_REGION_ERROR_THRESHOLD = 0.01 # reltol _DIFF_THRESH = 0.05 # reltol _NAME = os.path.splitext(os.path.basename(__file__))[0] class CropRegionRawTest(its_base_test.ItsBaseTest): """Test that RAW streams are not croppable.""" def test_crop_region_raw(self): with its_session_utils.ItsSession( device_id=self.dut.serial, camera_id=self.camera_id, hidden_physical_id=self.hidden_physical_id) as cam: props = cam.get_camera_properties() props = cam.override_with_hidden_physical_camera_props(props) log_path = self.log_path name_with_log_path = os.path.join(log_path, _NAME) # Check SKIP conditions camera_properties_utils.skip_unless( camera_properties_utils.compute_target_exposure(props) and camera_properties_utils.raw16(props) and camera_properties_utils.per_frame_control(props) and not camera_properties_utils.mono_camera(props)) # Load chart for scene its_session_utils.load_scene( cam, props, self.scene, self.tablet, its_session_utils.CHART_DISTANCE_NO_SCALING) # Calculate the active sensor region for a full (non-cropped) image. a = props['android.sensor.info.activeArraySize'] ax, ay = a['left'], a['top'] aw, ah = a['right'] - a['left'], a['bottom'] - a['top'] logging.debug('Active sensor region: (%d,%d %dx%d)', ax, ay, aw, ah) full_region = { 'left': 0, 'top': 0, 'right': aw, 'bottom': ah } # Calculate a center crop region. zoom = min(3.0, camera_properties_utils.get_max_digital_zoom(props)) if zoom < 1: raise AssertionError(f'zoom: {zoom:.2f}') crop_w = aw // zoom crop_h = ah // zoom crop_region = { 'left': aw // 2 - crop_w // 2, 'top': ah // 2 - crop_h // 2, 'right': aw // 2 + crop_w // 2, 'bottom': ah // 2 + crop_h // 2 } # Capture without a crop region. # Use a manual request with a linear tonemap so that the YUV and RAW # should look the same (once converted by image_processing_utils). e, s = target_exposure_utils.get_target_exposure_combos(log_path, cam)[ 'minSensitivity'] req = capture_request_utils.manual_capture_request(s, e, 0.0, True, props) cap1_raw, cap1_yuv = cam.do_capture(req, cam.CAP_RAW_YUV) # Capture with a crop region. req['android.scaler.cropRegion'] = crop_region cap2_raw, cap2_yuv = cam.do_capture(req, cam.CAP_RAW_YUV) # Check the metadata related to crop regions. # When both YUV and RAW are requested, the crop region that's # applied to YUV should be reported. # Note that the crop region returned by the cropped captures doesn't # need to perfectly match the one that was requested. imgs = {} for s, cap, cr_expected, err_delta in [ ('yuv_full', cap1_yuv, full_region, _CROP_FULL_ERROR_THRESHOLD), ('raw_full', cap1_raw, full_region, _CROP_FULL_ERROR_THRESHOLD), ('yuv_crop', cap2_yuv, crop_region, _CROP_REGION_ERROR_THRESHOLD), ('raw_crop', cap2_raw, crop_region, _CROP_REGION_ERROR_THRESHOLD)]: # Convert the capture to RGB and dump to a file. img = image_processing_utils.convert_capture_to_rgb_image(cap, props=props) image_processing_utils.write_image(img, f'{name_with_log_path}_{s}.jpg') imgs[s] = img # Get the crop region that is reported in the capture result. cr_reported = cap['metadata']['android.scaler.cropRegion'] x, y = cr_reported['left'], cr_reported['top'] w = cr_reported['right'] - cr_reported['left'] h = cr_reported['bottom'] - cr_reported['top'] logging.debug('Crop reported on %s: (%d,%d %dx%d)', s, x, y, w, h) # Test that the reported crop region is the same as the expected # one, for a non-cropped capture, and is close to the expected one, # for a cropped capture. ex = _CROP_FULL_ERROR_THRESHOLD ey = _CROP_FULL_ERROR_THRESHOLD if math.isclose(err_delta, _CROP_REGION_ERROR_THRESHOLD, rel_tol=0.01): ex = aw * err_delta ey = ah * err_delta logging.debug('error X, Y: %.2f, %.2f', ex, ey) if not ( (abs(cr_expected['left'] - cr_reported['left']) <= ex) and (abs(cr_expected['right'] - cr_reported['right']) <= ex) and (abs(cr_expected['top'] - cr_reported['top']) <= ey) and (abs(cr_expected['bottom'] - cr_reported['bottom']) <= ey)): raise AssertionError(f'expected: {cr_expected}, reported: ' f'{cr_reported}, ex: {ex:.2f}, ey: {ey:.2f}') # Also check the image content; 3 of the 4 shots should match. # Note that all the shots are RGB below; the variable names correspond # to what was captured. # Shrink the YUV images 2x2 -> 1 to account for the size reduction that # the raw images went through in the RGB conversion. imgs2 = {} for s, img in imgs.items(): h, w, _ = img.shape if s in ['yuv_full', 'yuv_crop']: img = img.reshape(h//2, 2, w//2, 2, 3).mean(3).mean(1) img = img.reshape(h//2, w//2, 3) imgs2[s] = img # Strip any border pixels from the raw shots (since the raw images may # be larger than the YUV images). Assume a symmetric padded border. xpad = (imgs2['raw_full'].shape[1] - imgs2['yuv_full'].shape[1]) // 2 ypad = (imgs2['raw_full'].shape[0] - imgs2['yuv_full'].shape[0]) // 2 wyuv = imgs2['yuv_full'].shape[1] hyuv = imgs2['yuv_full'].shape[0] imgs2['raw_full'] = imgs2['raw_full'][ypad:ypad+hyuv:, xpad:xpad+wyuv:, ::] imgs2['raw_crop'] = imgs2['raw_crop'][ypad:ypad+hyuv:, xpad:xpad+wyuv:, ::] logging.debug('Stripping padding before comparison: %dx%d', xpad, ypad) for s, img in imgs2.items(): image_processing_utils.write_image( img, f'{name_with_log_path}_comp_{s}.jpg') # Compute diffs between images of the same type. # The raw_crop and raw_full shots should be identical (since the crop # doesn't apply to raw images), and the yuv_crop and yuv_full shots # should be different. diff_yuv = np.fabs((imgs2['yuv_full'] - imgs2['yuv_crop'])).mean() diff_raw = np.fabs((imgs2['raw_full'] - imgs2['raw_crop'])).mean() logging.debug('YUV diff (crop vs. non-crop): %.3f', diff_yuv) logging.debug('RAW diff (crop vs. non-crop): %.3f', diff_raw) if diff_yuv <= _DIFF_THRESH: raise AssertionError('YUV diff too small! diff_yuv: ' f'{diff_yuv:.3f}, THRESH: {_DIFF_THRESH}') if diff_raw >= _DIFF_THRESH: raise AssertionError('RAW diff too big! diff_raw: ' f'{diff_raw:.3f}, THRESH: {_DIFF_THRESH}') if __name__ == '__main__': test_runner.main()