1# Copyright 2015 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 its.image 16import its.caps 17import its.device 18import its.objects 19import its.target 20import matplotlib 21import matplotlib.pyplot 22import numpy 23import os.path 24from matplotlib import pylab 25 26NR_MODES = [0, 1, 2, 3, 4] 27 28 29def main(): 30 """Test that the android.noiseReduction.mode param is applied when set for 31 reprocessing requests. 32 33 Capture reprocessed images with the camera dimly lit. Uses a high analog 34 gain to ensure the captured image is noisy. 35 36 Captures three reprocessed images, for NR off, "fast", and "high quality". 37 Also captures a reprocessed image with low gain and NR off, and uses the 38 variance of this as the baseline. 39 """ 40 41 NAME = os.path.basename(__file__).split(".")[0] 42 43 NUM_SAMPLES_PER_MODE = 4 44 SNR_TOLERANCE = 3 # unit in db 45 46 with its.device.ItsSession() as cam: 47 props = cam.get_camera_properties() 48 49 its.caps.skip_unless(its.caps.compute_target_exposure(props) and 50 its.caps.per_frame_control(props) and 51 its.caps.noise_reduction_mode(props, 0) and 52 (its.caps.yuv_reprocess(props) or 53 its.caps.private_reprocess(props))) 54 55 # If reprocessing is supported, ZSL NR mode must be avaiable. 56 assert(its.caps.noise_reduction_mode(props, 4)) 57 58 reprocess_formats = [] 59 if (its.caps.yuv_reprocess(props)): 60 reprocess_formats.append("yuv") 61 if (its.caps.private_reprocess(props)): 62 reprocess_formats.append("private") 63 64 for reprocess_format in reprocess_formats: 65 # List of variances for R, G, B. 66 snrs = [[], [], []] 67 nr_modes_reported = [] 68 69 # NR mode 0 with low gain 70 e, s = its.target.get_target_exposure_combos(cam)["minSensitivity"] 71 req = its.objects.manual_capture_request(s, e) 72 req["android.noiseReduction.mode"] = 0 73 74 # Test reprocess_format->JPEG reprocessing 75 # TODO: Switch to reprocess_format->YUV when YUV reprocessing is 76 # supported. 77 size = its.objects.get_available_output_sizes("jpg", props)[0] 78 out_surface = {"width":size[0], "height":size[1], "format":"jpg"} 79 cap = cam.do_capture(req, out_surface, reprocess_format) 80 img = its.image.decompress_jpeg_to_rgb_image(cap["data"]) 81 its.image.write_image(img, "%s_low_gain_fmt=jpg.jpg" % (NAME)) 82 tile = its.image.get_image_patch(img, 0.45, 0.45, 0.1, 0.1) 83 ref_snr = its.image.compute_image_snrs(tile) 84 print "Ref SNRs:", ref_snr 85 86 e, s = its.target.get_target_exposure_combos(cam)["maxSensitivity"] 87 for nr_mode in NR_MODES: 88 # Skip unavailable modes 89 if not its.caps.noise_reduction_mode(props, nr_mode): 90 nr_modes_reported.append(nr_mode) 91 for channel in range(3): 92 snrs[channel].append(0) 93 continue 94 95 rgb_snr_list = [] 96 # Capture several images to account for per frame noise 97 # variations 98 for n in range(NUM_SAMPLES_PER_MODE): 99 req = its.objects.manual_capture_request(s, e) 100 req["android.noiseReduction.mode"] = nr_mode 101 cap = cam.do_capture(req, out_surface, reprocess_format) 102 103 img = its.image.decompress_jpeg_to_rgb_image(cap["data"]) 104 if n == 0: 105 its.image.write_image( 106 img, 107 "%s_high_gain_nr=%d_fmt=jpg.jpg" 108 %(NAME, nr_mode)) 109 nr_modes_reported.append( 110 cap["metadata"]["android.noiseReduction.mode"]) 111 112 tile = its.image.get_image_patch(img, 0.45, 0.45, 0.1, 0.1) 113 # Get the variances for R, G, and B channels 114 rgb_snrs = its.image.compute_image_snrs(tile) 115 rgb_snr_list.append(rgb_snrs) 116 117 r_snrs = [rgb[0] for rgb in rgb_snr_list] 118 g_snrs = [rgb[1] for rgb in rgb_snr_list] 119 b_snrs = [rgb[2] for rgb in rgb_snr_list] 120 rgb_snrs = [numpy.mean(r_snrs), 121 numpy.mean(g_snrs), 122 numpy.mean(b_snrs)] 123 print "NR mode", nr_mode, "SNRs:" 124 print " R SNR:", rgb_snrs[0],\ 125 "Min:", min(r_snrs), "Max:", max(r_snrs) 126 print " G SNR:", rgb_snrs[1],\ 127 "Min:", min(g_snrs), "Max:", max(g_snrs) 128 print " B SNR:", rgb_snrs[2],\ 129 "Min:", min(b_snrs), "Max:", max(b_snrs) 130 131 for chan in range(3): 132 snrs[chan].append(rgb_snrs[chan]) 133 134 # Draw a plot. 135 pylab.figure(reprocess_format) 136 for channel in range(3): 137 pylab.plot(NR_MODES, snrs[channel], "-"+"rgb"[channel]+"o") 138 139 pylab.xlabel("Noise Reduction Mode") 140 pylab.ylabel("SNR (dB)") 141 pylab.xticks(NR_MODES) 142 matplotlib.pyplot.savefig("%s_plot_%s_SNRs.png" % 143 (NAME, reprocess_format)) 144 145 assert nr_modes_reported == NR_MODES 146 147 for j in range(3): 148 # Larger is better 149 # Verify OFF(0) is not better than FAST(1) 150 assert(snrs[j][0] < 151 snrs[j][1] + SNR_TOLERANCE) 152 # Verify FAST(1) is not better than HQ(2) 153 assert(snrs[j][1] < 154 snrs[j][2] + SNR_TOLERANCE) 155 # Verify HQ(2) is better than OFF(0) 156 assert(snrs[j][0] < snrs[j][2]) 157 if its.caps.noise_reduction_mode(props, 3): 158 # Verify OFF(0) is not better than MINIMAL(3) 159 assert(snrs[j][0] < 160 snrs[j][3] + SNR_TOLERANCE) 161 # Verify MINIMAL(3) is not better than HQ(2) 162 assert(snrs[j][3] < 163 snrs[j][2] + SNR_TOLERANCE) 164 # Verify ZSL(4) is close to MINIMAL(3) 165 assert(numpy.isclose(snrs[j][4], snrs[j][3], 166 atol=SNR_TOLERANCE)) 167 else: 168 # Verify ZSL(4) is close to OFF(0) 169 assert(numpy.isclose(snrs[j][4], snrs[j][0], 170 atol=SNR_TOLERANCE)) 171 172if __name__ == '__main__': 173 main() 174 175