删除重复的行附近并保留得分最高的行

时间:2019-02-14 08:18:53

标签: r hough-transform

我有一个包含三个变量的数据框:thetarhoscore。我想过滤掉“几乎”重复的行。

例如:如果theta的值与另一行的间隔为10 rho的间隔为15,则可以认为它们是重复项。然后,我要删除两个重复项中最低的score行。

我在想的是,首先arrange()将数据帧score(降序),然后创建一个for()循环。然后创建一个向量lines,并在没有重复项的情况下向lines添加一行。

我真的想避免使用for()循环,因为数据帧很大(超过40000行),并且这样的循环可能很慢。我不知道如何使用例如dplyr来实现这一目标。

数据样本:

> dput(head(df %>% arrange(desc(score)), 1000))

structure(list(theta = c(112.727272727111, 294.545454545031, 
298.18181818139, 294.545454545031, 116.363636363469, 116.363636363469, 
116.363636363469, 112.727272727111, 294.545454545031, 298.18181818139, 
112.727272727111, 119.999999999828, 116.363636363469, 112.727272727111, 
294.545454545031, 298.18181818139, 294.545454545031, 301.818181817748, 
294.545454545031, 112.727272727111, 298.18181818139, 116.363636363469, 
290.909090908673, 290.909090908673, 116.363636363469, 116.363636363469, 
112.727272727111, 298.18181818139, 116.363636363469, 294.545454545031, 
119.999999999828, 298.18181818139, 298.18181818139, 119.999999999828, 
116.363636363469, 287.272727272315, 294.545454545031, 116.363636363469, 
112.727272727111, 112.727272727111, 101.818181818036, 298.18181818139, 
290.909090908673, 116.363636363469, 119.999999999828, 298.18181818139, 
298.18181818139, 290.909090908673, 109.090909090752, 290.909090908673, 
294.545454545031, 109.090909090752, 290.909090908673, 119.999999999828, 
290.909090908673, 109.090909090752, 112.727272727111, 301.818181817748, 
290.909090908673, 116.363636363469, 101.818181818036, 109.090909090752, 
109.090909090752, 116.363636363469, 287.272727272315, 87.2727272726019, 
119.999999999828, 112.727272727111, 301.818181817748, 119.999999999828, 
119.999999999828, 83.6363636362435, 109.090909090752, 290.909090908673, 
109.090909090752, 119.999999999828, 301.818181817748, 116.363636363469, 
109.090909090752, 109.090909090752, 283.636363635956, 290.909090908673, 
330.909090908615, 109.090909090752, 298.18181818139, 301.818181817748, 
116.363636363469, 105.454545454394, 29.0909090908673, 101.818181818036, 
105.454545454394, 119.999999999828, 290.909090908673, 116.363636363469, 
112.727272727111, 109.090909090752, 269.090909090522, 105.454545454394, 
109.090909090752, 290.909090908673, 269.090909090522, 301.818181817748, 
301.818181817748, 305.454545454107, 272.727272726881, 283.636363635956, 
87.2727272726019, 290.909090908673, 116.363636363469, 301.818181817748, 
290.909090908673, 294.545454545031, 239.999999999655, 105.454545454394, 
119.999999999828, 301.818181817748, 116.363636363469, 290.909090908673, 
83.6363636362435, 327.272727272257, 283.636363635956, 283.636363635956, 
254.545454545089, 287.272727272315, 258.181818181447, 119.999999999828, 
298.18181818139, 261.818181817806, 87.2727272726019, 290.909090908673, 
210.909090908788, 290.909090908673, 105.454545454394, 327.272727272257, 
319.99999999954, 65.4545454544514, 287.272727272315, 119.999999999828, 
283.636363635956, 287.272727272315, 298.18181818139, 105.454545454394, 
279.999999999598, 272.727272726881, 305.454545454107, 287.272727272315, 
87.2727272726019, 43.6363636363009, 87.2727272726019, 43.6363636363009, 
287.272727272315, 138.18181818162, 119.999999999828, 309.090909090465, 
7.27272727271682, 90.9090909089603, 290.909090908673, 301.818181817748, 
39.9999999999425, 32.7272727272257, 109.090909090752, 116.363636363469, 
279.999999999598, 105.454545454394, 119.999999999828, 276.363636363239, 
305.454545454107, 301.818181817748, 101.818181818036, 218.181818181505, 
272.727272726881, 272.727272726881, 301.818181817748, 309.090909090465, 
279.999999999598, 254.545454545089, 243.636363636014, 229.09090909058, 
32.7272727272257, 199.999999999713, 254.545454545089, 272.727272726881, 
294.545454545031, 301.818181817748, 283.636363635956, 123.636363636186, 
7.27272727271682, 287.272727272315, 87.2727272726019, 254.545454545089, 
283.636363635956, 305.454545454107, 65.4545454544514, 330.909090908615, 
119.999999999828, 123.636363636186, 283.636363635956, 72.7272727271682, 
76.3636363635266, 109.090909090752), rho = c(41.0836002377417, 
-45.0465084042093, -37.9532480936103, -43.044158227946, 30.9877673496997, 
35.9936427903578, 32.9901175259629, 39.0812500614784, -46.0476834923408, 
-29.9438473885574, 40.0824251496101, 23.8936399748773, 26.9830669971733, 
42.0847753258733, -42.0429831398144, -36.9520730054787, -41.0418080516828, 
-26.8536397978266, -44.0453333160777, 43.0859504140049, -35.950897917347, 
28.9854171734365, -50.1354285011879, -51.1366035893195, 34.9924677022261, 
27.9842420853049, 46.0894756783997, -34.9497228292154, 25.9818919090416, 
-48.0500336686041, 13.8818890935612, -38.9544231817418, -40.9567733580051, 
24.8948150630089, 29.9865922615681, -60.2278996099466, -47.0488585804724, 
33.9912926140945, 37.0788998852153, 38.0800749733469, 85.3732907252706, 
-33.9485477410838, -54.1401288537143, 31.9889424378313, 16.885414357956, 
-32.9473726529522, -31.9461975648206, -62.1495295587672, 55.18197515857, 
-48.1330783249247, -49.0512087567357, 64.1925509517545, -52.1377786774511, 
26.8971652392721, -49.1342534130563, 48.1737495416487, 44.0871255021365, 
-27.8548148859583, -59.1460042943724, 24.98071682091, 66.3509640507701, 
45.1702242772539, 49.1749246297804, 22.9783666446468, -82.253751548842, 
120.686872286734, 12.8807140054296, 48.0918258546629, -21.8477643571686, 
10.8783638291664, 14.8830641816928, 131.758025665537, 53.1796249823068, 
-60.147179382504, 60.1878505992281, 8.87601365290313, -13.8383636521157, 
23.9795417327784, 47.1725744535171, 57.1843253348332, -72.3200714342242, 
-53.1389537655828, 91.9951732050117, 40.1643488365959, -27.9414972122942, 
-22.8489394453002, 21.9771915565152, 50.2555352073846, 237.136779368164, 
55.3380380813223, 70.2790369700168, 21.8912897986141, -67.1554049994253, 
20.9760164683836, 45.0883005902681, 52.1784498941752, -119.654988035104, 
62.2696362649639, 50.176099717912, -57.1436541181092, -178.724318234869, 
-18.8442390927738, -33.8618654147479, -0.73510658654489, -110.579957573771, 
-90.3412230205932, 99.6621954359703, -72.1612804400833, 36.9948178784894, 
-29.8571650622215, -65.1530548231621, -51.0535589329989, -167.060502489566, 
48.2531850311214, 25.8959901511405, -31.8595152384847, 19.974841380252, 
-58.1448292062408, 183.819130248381, 62.8740557236012, -91.3423981087248, 
-89.3400479324616, -143.897766031004, -59.226724521815, -169.881611464456, 
29.900690503667, -42.9591235342683, -150.80781195583, 111.67629649355, 
-42.126027796135, -244.156619676724, -68.1565800875569, 82.2931380275962, 
72.8858066049173, 45.6767233434207, 169.021874270401, -81.2525764607104, 
20.8901147104825, -98.3506237256461, -61.2290746980781, -39.9555982698734, 
54.260235559911, -112.44218190202, -119.590533366955, -10.746857467861, 
-50.2161487286305, 132.700973344313, 237.190546304581, 121.688047374866, 
232.184670863923, -52.2184989048937, -53.6413626642005, 15.8842392698244, 
20.3784955094908, 211.865903318108, 154.664384192041, -63.1507046468988, 
-28.8559899740899, 236.184390219668, 239.160977781544, 46.1713993653855, 
17.9724912039888, -70.3928282004921, 53.2590604717795, 11.879538917298, 
-127.531648913209, 0.266068501586671, -3.82661277079962, 102.393267223508, 
-228.170369542773, -123.595233719481, -104.572907044981, -25.852464709695, 
1.35616883499019, -60.3810773191761, -248.019975196692, -210.085066262338, 
-240.190868827689, 237.158627605281, -218.036119344605, -200.964746054506, 
-122.59405863135, -56.0594343736569, -20.846589269037, -93.344748284988, 
3.78276259922932, 162.80832399966, -54.2208490811569, 175.751502133973, 
-180.941244291874, -92.3435731968564, 9.27664429477117, 170.023049358533, 
84.9869475880904, 18.8877645342192, -2.22428792956033, -63.3094956410397, 
127.900471265421, 127.856219152482, 32.154948131543), score = c(381.781574988754, 
376.867929304018, 364.626105647432, 299.691569724508, 294.231324378097, 
293.952772362833, 283.460261504884, 282.056364048014, 267.206929374373, 
261.168680064558, 252.414542012009, 237.042185845451, 234.078052148331, 
229.606725420143, 216.324860493748, 215.558533330939, 210.842537314501, 
208.368243554419, 205.181406815045, 204.261096051397, 197.324237979165, 
189.521926303097, 188.969964944608, 186.638575913254, 185.140684108428, 
183.377726563851, 182.124986813598, 182.045260423294, 181.846212597077, 
180.37944359477, 178.789891727661, 176.192227059953, 175.964328485497, 
174.846290171029, 171.988837570636, 169.019640895807, 167.9119495256, 
167.813177641545, 163.691114744162, 163.450749580731, 162.451649238718, 
158.403998247919, 156.155648844112, 153.590137368734, 152.769361891751, 
152.520713855705, 151.655777652606, 149.409331820614, 146.472014630364, 
146.115633442028, 146.052498649575, 135.864783446716, 134.768563651941, 
132.404742423211, 131.51576293036, 130.82284429468, 130.788174510703, 
128.424175293775, 126.128019200478, 124.336645293274, 123.597399463198, 
121.646401630518, 120.301723278087, 114.561484716175, 114.48553813731, 
113.216562649051, 112.226621711484, 111.461463120992, 110.013083445497, 
109.576273588556, 109.137012441314, 106.688602963489, 106.400626093183, 
106.133536519578, 105.777492905745, 105.688442177713, 104.456641874613, 
104.433669269611, 103.314623077763, 102.259746393435, 102.17938530452, 
101.670539083899, 100.452688050998, 99.02520484887, 98.4335113782726, 
98.0429505018094, 97.4402598820602, 95.7587658236876, 95.2950696240885, 
94.9586058341693, 94.3866751251166, 94.3158182244223, 93.7565070140147, 
92.127262977829, 91.8548747919413, 91.4386321895478, 91.3768416587042, 
91.3010019223121, 91.1345447367969, 90.3128424524546, 89.0661383970599, 
89.0055316281083, 88.6971624966804, 87.7983886555544, 87.0306396564274, 
85.9204774141763, 85.7965196497019, 85.3142922578242, 84.8425562837965, 
84.3735638666951, 83.9731763120239, 82.7466784247688, 82.6092763898421, 
82.0587829926224, 81.6777059131652, 81.4293731744454, 80.8438294356826, 
80.8211826121344, 80.6709609724947, 80.6465532824821, 80.5178349938752, 
80.1999503837411, 80.1719116395444, 79.8942969725428, 79.7436418638969, 
79.7277805289744, 79.70684435464, 79.6847128002474, 79.6806763649338, 
79.1777583656981, 78.6352312634342, 78.6274361426217, 78.5428920502696, 
77.7337176668642, 76.7946629291839, 76.4466244176081, 76.1485811758828, 
75.7612036684239, 75.2150823543861, 75.0345313303339, 74.8673067291731, 
74.8517038879151, 74.7508858198005, 74.5527292285188, 74.5267512278319, 
74.5227005364768, 74.4094234954551, 74.3674893962718, 74.3219691212919, 
73.8157794991736, 73.2286453984909, 73.222367796426, 72.8313773620846, 
72.7470649807689, 72.5137753468749, 72.5132078673106, 72.5025730252246, 
72.4385676060556, 72.1537622540259, 71.9425613728977, 71.6928128941157, 
71.5832397214361, 71.5322922095169, 71.2972159994947, 71.1810801604769, 
71.1642327631934, 71.076532336903, 71.0515945984565, 70.8404496937105, 
70.6690362061286, 70.3221433523102, 70.3120255269573, 70.3017098533045, 
70.265603928538, 70.2477297855539, 69.8757560784155, 69.8418754120254, 
69.7502537524278, 69.6096619396855, 69.0774383151024, 69.0200227396948, 
69.0172508598699, 68.9528342390049, 68.9023657817845, 68.5773542159423, 
68.305995695382, 68.1327451195585, 67.9244802616279, 67.6804525353541, 
67.5989028542724, 67.5963786570388, 67.4025447304631, 67.3888299703067, 
67.0351641892631, 66.9494920024158, 66.2982455895387, 66.2033914475987, 
66.0129552398745, 65.9579581030256, 65.6753707562488)), row.names = c(NA, 
200L), class = "data.frame")

旁注:数据是通过对图像进行霍夫变换获得的。

编辑:示例数据:

> df
  theta rho score
1    10  10   200
2    15  15   150
3    16  16   145
4    25  25   100
5    50  50    50
6    60  60    40
7    70  70    20
8    75  75     5
9   110 110   100

所需的输出:

> df
  theta rho score
1    10  10   200
2    50  50    50
3   110 110   100

对于我的实际数据,theta周围的间隔为8,rho周围的间隔为10

2 个答案:

答案 0 :(得分:0)

整洁的方法

所需输出的样本数据

df <- read.table(text="id theta rho score
1    10  10   200
                 2    15  15   150
                 3    16  16   145
                 4    25  25   100
                 5    50  50    50
                 6    60  60    40
                 7    70  70    20
                 8    75  75     5
                 9   110 110   100", header = TRUE)

代码

library( tidyverse )

df %>%
  #first, arrange by theta
  arrange( theta ) %>%
  #mark rows with theta <=10 of next row
  mutate( theta_mark = if_else( abs( theta - lag(theta) ) <= 10, 1, 0 ) ) %>%
  #second, arrange by rho
  arrange( rho ) %>%
  #mark rows with rho <=15 of next row
  mutate( rho_mark = if_else( abs( rho - lag(rho) ) <= 15, 1, 0 ) ) %>%
  #replace NA's with 0
  replace_na( list( theta_mark = 0, rho_mark = 0 ) ) %>%
  #filter all unwanted rows
  filter( theta_mark == 0 & rho_mark == 0 ) %>%
  #drop unwanted columns
  select( -theta_mark, -rho_mark )

输出

#   id theta rho score
# 1  1    10  10   200
# 2  5    50  50    50
# 3  9   110 110   100

答案 1 :(得分:0)

参考我对重叠重复项的评论,您可以迭代地执行此操作,删除得分最高的重复项,然后重复其余得分的下一个最高得分等。下面的代码应该为您做到这一点,并且在您发布的数据集中,剩下27分。尽管效率很低,但循环是我想出快速答案的最简单方法。我确信没有循环就可以做到这一点,如果您拥有数千万个点,您将需要它,但是在这种情况下,确实没有必要。

df1 <- df
thetaInterval <- 8
rhoInterval <- 10
i <- 0
while(i < nrow(df1)){
  i <- i + 1
  df1 <- df1[order(df1$score, decreasing = TRUE), ]
  dups <- abs(df1$theta - df1$theta[i]) <= 8 &
    abs(df1$rho - df1$rho[i]) <= 10
  df1 <- df1[!dups, ]
}