1# Copyright 2014 The Android Open Source Project
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7#      http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14
15import os.path
16
17import its.caps
18import its.device
19import its.image
20import its.objects
21import matplotlib
22from matplotlib import pylab
23import numpy as np
24
25NAME = os.path.basename(__file__).split('.')[0]
26LOCKED = 3
27LUMA_LOCKED_TOL = 0.05
28THRESH_CONVERGE_FOR_EV = 8  # AE must converge within this num
29YUV_FULL_SCALE = 255.0
30YUV_SATURATION_MIN = 253.0
31YUV_SATURATION_TOL = 1.0
32
33
34def main():
35    """Tests that EV compensation is applied."""
36
37    with its.device.ItsSession() as cam:
38        props = cam.get_camera_properties()
39        its.caps.skip_unless(its.caps.ev_compensation(props) and
40                             its.caps.ae_lock(props))
41
42        debug = its.caps.debug_mode()
43        mono_camera = its.caps.mono_camera(props)
44        largest_yuv = its.objects.get_largest_yuv_format(props)
45        if debug:
46            fmt = largest_yuv
47        else:
48            match_ar = (largest_yuv['width'], largest_yuv['height'])
49            fmt = its.objects.get_smallest_yuv_format(props, match_ar=match_ar)
50
51        ev_per_step = its.objects.rational_to_float(
52            props['android.control.aeCompensationStep'])
53        steps_per_ev = int(1.0 / ev_per_step)
54        evs = range(-2 * steps_per_ev, 2 * steps_per_ev + 1, steps_per_ev)
55        lumas = []
56        reds = []
57        greens = []
58        blues = []
59
60        # Converge 3A, and lock AE once converged. skip AF trigger as
61        # dark/bright scene could make AF convergence fail and this test
62        # doesn't care the image sharpness.
63        cam.do_3a(ev_comp=0, lock_ae=True, do_af=False, mono_camera=mono_camera)
64
65        for ev in evs:
66            # Capture a single shot with the same EV comp and locked AE.
67            req = its.objects.auto_capture_request()
68            req['android.control.aeExposureCompensation'] = ev
69            req['android.control.aeLock'] = True
70            caps = cam.do_capture([req]*THRESH_CONVERGE_FOR_EV, fmt)
71            luma_locked = []
72            for i, cap in enumerate(caps):
73                if cap['metadata']['android.control.aeState'] == LOCKED:
74                    y = its.image.convert_capture_to_planes(cap)[0]
75                    tile = its.image.get_image_patch(y, 0.45, 0.45, 0.1, 0.1)
76                    luma = its.image.compute_image_means(tile)[0]
77                    luma_locked.append(luma)
78                    if i == THRESH_CONVERGE_FOR_EV-1:
79                        lumas.append(luma)
80                        rgb = its.image.convert_capture_to_rgb_image(cap)
81                        rgb_tile = its.image.get_image_patch(rgb,
82                                                             0.45, 0.45,
83                                                             0.1, 0.1)
84                        rgb_means = its.image.compute_image_means(rgb_tile)
85                        reds.append(rgb_means[0])
86                        greens.append(rgb_means[1])
87                        blues.append(rgb_means[2])
88                        print 'lumas in AE locked captures: ', luma_locked
89                        assert np.isclose(min(luma_locked), max(luma_locked),
90                                          rtol=LUMA_LOCKED_TOL)
91            assert caps[THRESH_CONVERGE_FOR_EV-1]['metadata']['android.control.aeState'] == LOCKED
92
93        pylab.plot(evs, lumas, '-ro')
94        pylab.xlabel('EV Compensation')
95        pylab.ylabel('Mean Luma (Normalized)')
96        matplotlib.pyplot.savefig('%s_plot_means.png' % (NAME))
97
98        # Trim extra saturated images
99        while lumas and lumas[-1] >= YUV_SATURATION_MIN/YUV_FULL_SCALE:
100            if (np.isclose(reds[-1], greens[-1],
101                           YUV_SATURATION_TOL/YUV_FULL_SCALE) and
102                    np.isclose(blues[-1], greens[-1],
103                               YUV_SATURATION_TOL/YUV_FULL_SCALE)):
104                lumas.pop(-1)
105                reds.pop(-1)
106                greens.pop(-1)
107                blues.pop(-1)
108                print 'Removed saturated image.'
109            else:
110                break
111        # Only allow positive EVs to give saturated image
112        assert len(lumas) > 2
113        luma_diffs = np.diff(lumas)
114        min_luma_diffs = min(luma_diffs)
115        print 'Min of the luma value difference between adjacent ev comp: ',
116        print min_luma_diffs
117        # All luma brightness should be increasing with increasing ev comp.
118        assert min_luma_diffs > 0
119
120if __name__ == '__main__':
121    main()
122