Coverage for python/lsst/meas/algorithms/accumulator_mean_stack.py: 9%

81 statements  

« prev     ^ index     » next       coverage.py v6.5.0, created at 2023-01-14 03:09 -0800

1# This file is part of meas_algorithms. 

2# 

3# LSST Data Management System 

4# This product includes software developed by the 

5# LSST Project (http://www.lsst.org/). 

6# See COPYRIGHT file at the top of the source tree. 

7# 

8# This program is free software: you can redistribute it and/or modify 

9# it under the terms of the GNU General Public License as published by 

10# the Free Software Foundation, either version 3 of the License, or 

11# (at your option) any later version. 

12# 

13# This program is distributed in the hope that it will be useful, 

14# but WITHOUT ANY WARRANTY; without even the implied warranty of 

15# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 

16# GNU General Public License for more details. 

17# 

18# You should have received a copy of the LSST License Statement and 

19# the GNU General Public License along with this program. If not, 

20# see <https://www.lsstcorp.org/LegalNotices/>. 

21# 

22import numpy as np 

23 

24 

25__all__ = ['AccumulatorMeanStack'] 

26 

27 

28class AccumulatorMeanStack(object): 

29 """Stack masked images. 

30 

31 Parameters 

32 ---------- 

33 shape : `tuple` 

34 Shape of the input and output images. 

35 bit_mask_value : `int` 

36 Bit mask to flag for "bad" inputs that should not be stacked. 

37 mask_threshold_dict : `dict` [`int`: `float`], optional 

38 Dictionary of mapping from bit number to threshold for flagging. 

39 Only bad bits (in bit_mask_value) which mask fractional weight 

40 greater than this threshold will be flagged in the output image. 

41 mask_map : `list` [`tuple`], optional 

42 Mapping from input image bits to aggregated coadd bits. 

43 no_good_pixels_mask : `int`, optional 

44 Bit mask to set when there are no good pixels in the stack. 

45 If not set then will set coadd masked image 'NO_DATA' bit. 

46 calc_error_from_input_variance : `bool`, optional 

47 Calculate the error from the input variance? 

48 compute_n_image : `bool`, optional 

49 Calculate the n_image map as well as stack? 

50 """ 

51 def __init__(self, shape, 

52 bit_mask_value, mask_threshold_dict={}, 

53 mask_map=[], no_good_pixels_mask=None, 

54 calc_error_from_input_variance=True, 

55 compute_n_image=False): 

56 self.shape = shape 

57 self.bit_mask_value = bit_mask_value 

58 self.mask_map = mask_map 

59 self.no_good_pixels_mask = no_good_pixels_mask 

60 self.calc_error_from_input_variance = calc_error_from_input_variance 

61 self.compute_n_image = compute_n_image 

62 

63 # Only track threshold bits that are in the bad bit_mask_value. 

64 self.mask_threshold_dict = {} 

65 for bit in mask_threshold_dict: 

66 if (self.bit_mask_value & 2**bit) > 0: 

67 self.mask_threshold_dict[bit] = mask_threshold_dict[bit] 

68 

69 # sum_weight holds the sum of weights for each pixel. 

70 self.sum_weight = np.zeros(shape, dtype=np.float64) 

71 # sum_wdata holds the sum of weight*data for each pixel. 

72 self.sum_wdata = np.zeros(shape, dtype=np.float64) 

73 

74 if calc_error_from_input_variance: 

75 # sum_w2var holds the sum of weight**2 * variance for each pixel. 

76 self.sum_w2var = np.zeros(shape, dtype=np.float64) 

77 else: 

78 # sum_weight2 holds the sum of weight**2 for each pixel. 

79 self.sum_weight2 = np.zeros(shape, dtype=np.float64) 

80 # sum_wdata2 holds the sum of weight * data**2 for each pixel. 

81 self.sum_wdata2 = np.zeros(shape, dtype=np.float64) 

82 

83 self.or_mask = np.zeros(shape, dtype=np.int64) 

84 self.rejected_weights_by_bit = {} 

85 for bit in self.mask_threshold_dict: 

86 self.rejected_weights_by_bit[bit] = np.zeros(shape, dtype=np.float64) 

87 

88 self.masked_pixels_mask = np.zeros(shape, dtype=np.int64) 

89 

90 if self.compute_n_image: 

91 self.n_image = np.zeros(shape, dtype=np.int32) 

92 

93 def add_masked_image(self, masked_image, weight=1.0): 

94 """Add a masked image to the stack. 

95 

96 Parameters 

97 ---------- 

98 masked_image : `lsst.afw.image.MaskedImage` 

99 Masked image to add to the stack. 

100 weight : `float` or `np.ndarray`, optional 

101 Weight to apply for weighted mean. If an array, 

102 must be same size and shape as input masked_image. 

103 """ 

104 good_pixels = np.where(((masked_image.mask.array & self.bit_mask_value) == 0) 

105 & np.isfinite(masked_image.mask.array)) 

106 

107 self.sum_weight[good_pixels] += weight 

108 self.sum_wdata[good_pixels] += weight*masked_image.image.array[good_pixels] 

109 

110 if self.compute_n_image: 

111 self.n_image[good_pixels] += 1 

112 

113 if self.calc_error_from_input_variance: 

114 self.sum_w2var[good_pixels] += (weight**2.)*masked_image.variance.array[good_pixels] 

115 else: 

116 self.sum_weight2[good_pixels] += weight**2. 

117 self.sum_wdata2[good_pixels] += weight*(masked_image.image.array[good_pixels]**2.) 

118 

119 # Mask bits are propagated for good pixels 

120 self.or_mask[good_pixels] |= masked_image.mask.array[good_pixels] 

121 

122 # Bad pixels are only tracked if they cross a threshold 

123 for bit in self.mask_threshold_dict: 

124 bad_pixels = ((masked_image.mask.array & 2**bit) > 0) 

125 self.rejected_weights_by_bit[bit][bad_pixels] += weight 

126 self.masked_pixels_mask[bad_pixels] |= 2**bit 

127 

128 def fill_stacked_masked_image(self, stacked_masked_image): 

129 """Fill the stacked mask image after accumulation. 

130 

131 Parameters 

132 ---------- 

133 stacked_masked_image : `lsst.afw.image.MaskedImage` 

134 Total masked image. 

135 """ 

136 with np.warnings.catch_warnings(): 

137 # Let the NaNs through and flag bad pixels below 

138 np.warnings.simplefilter("ignore") 

139 

140 # The image plane is sum(weight*data)/sum(weight) 

141 stacked_masked_image.image.array[:, :] = self.sum_wdata/self.sum_weight 

142 

143 if self.calc_error_from_input_variance: 

144 mean_var = self.sum_w2var/(self.sum_weight**2.) 

145 else: 

146 # Compute the biased estimator 

147 variance = self.sum_wdata2/self.sum_weight - stacked_masked_image.image.array[:, :]**2. 

148 # De-bias 

149 variance *= (self.sum_weight**2.)/(self.sum_weight**2. - self.sum_weight2) 

150 

151 # Compute the mean variance 

152 mean_var = variance*self.sum_weight2/(self.sum_weight**2.) 

153 

154 stacked_masked_image.variance.array[:, :] = mean_var 

155 

156 # Propagate bits when they cross the threshold 

157 for bit in self.mask_threshold_dict: 

158 hypothetical_total_weight = self.sum_weight + self.rejected_weights_by_bit[bit] 

159 self.rejected_weights_by_bit[bit] /= hypothetical_total_weight 

160 propagate = np.where(self.rejected_weights_by_bit[bit] > self.mask_threshold_dict[bit]) 

161 self.or_mask[propagate] |= 2**bit 

162 

163 # Map mask planes to new bits for pixels that had at least one 

164 # bad input rejected and are in the mask_map. 

165 for mask_tuple in self.mask_map: 

166 self.or_mask[(self.masked_pixels_mask & mask_tuple[0]) > 0] |= mask_tuple[1] 

167 

168 stacked_masked_image.mask.array[:, :] = self.or_mask 

169 

170 if self.no_good_pixels_mask is None: 

171 mask_dict = stacked_masked_image.mask.getMaskPlaneDict() 

172 no_good_pixels_mask = 2**(mask_dict['NO_DATA']) 

173 else: 

174 no_good_pixels_mask = self.no_good_pixels_mask 

175 

176 bad_pixels = (self.sum_weight <= 0.0) 

177 stacked_masked_image.mask.array[bad_pixels] |= no_good_pixels_mask 

178 

179 def add_image(self, image, weight=1.0): 

180 """Add an image to the stack. 

181 

182 No bit-filtering is performed when adding an image. 

183 

184 Parameters 

185 ---------- 

186 image : `lsst.afw.image.Image` 

187 Image to add to the stack. 

188 weight : `float` or `np.ndarray`, optional 

189 Weight to apply for weighted mean. If an array, 

190 must be same size and shape as input image. 

191 """ 

192 self.sum_weight[:, :] += weight 

193 self.sum_wdata[:, :] += weight*image.array[:] 

194 

195 if self.compute_n_image: 

196 self.n_image[:, :] += 1 

197 

198 def fill_stacked_image(self, stacked_image): 

199 """Fill the image after accumulation. 

200 

201 Parameters 

202 ---------- 

203 stacked_image : `lsst.afw.image.Image` 

204 Total image. 

205 """ 

206 with np.warnings.catch_warnings(): 

207 # Let the NaNs through, this should only happen 

208 # if we're stacking with no inputs. 

209 np.warnings.simplefilter("ignore") 

210 

211 # The image plane is sum(weight*data)/sum(weight) 

212 stacked_image.array[:, :] = self.sum_wdata/self.sum_weight 

213 

214 @staticmethod 

215 def stats_ctrl_to_threshold_dict(stats_ctrl): 

216 """Convert stats control to threshold dict. 

217 

218 Parameters 

219 ---------- 

220 stats_ctrl : `lsst.afw.math.StatisticsControl` 

221 

222 Returns 

223 ------- 

224 threshold_dict : `dict` 

225 Dict mapping from bit to propagation threshold. 

226 """ 

227 threshold_dict = {} 

228 for bit in range(64): 

229 threshold_dict[bit] = stats_ctrl.getMaskPropagationThreshold(bit) 

230 

231 return threshold_dict