Advances in horse morphometric measurements using LiDAR

1 Ad ances in ho se mo phome ic measu emen s using 1 LiDAR 2 3 M. Pé ez-Ruiz1,2; D. Ta a -Ma ín1; M. J. Sánchez-Gue e o3; M. Vale a3 4 5 1. Uni e sidad de Se illa. Á ea de Ingenie ía Ag o o es al 6 Dp o. de Ingenie ía Ae oespacial y Mecánica de Fluidos, Spain. 7 2. Co esponding Au ho ’s con ac in o ma ion: 8 ETSIA. C a. Se illa-U e a km 1 9 Se illa, Spain 41013 10 Phone: 954 481 389 11 Fax: 954 486 436 12 E-mail: [email p o ec ed] 13 3. Uni e sidad de Se illa. Dp o. de Ciencias Ag o o es ales, Spain 14 15 16 Abs ac . 17 Zoome ic measu emen s ha e a po en ial alue in di e en ialing be ween indi iduals and wi hin 18 popula ions. The measu emen o body size in ho ses and li es ock plays a signi ican ole in 19 unc ional longe i y, p oduc ion, and ep oduc i e pe o mance and heal h. In his con ex , he 20 measu emen s ob ained wi hou con ac by de ec ion sys ems and isualized by compu e s 21 could ep esen a g ea ad ance o e con en ional measu emen s ha a e edious, ime 22 consuming and s ess ul o he animals. This s udy p esen s a new app oach o aking 23 zoome ic measu emen s o an animal's body based on digi al h ee-dimensional modelling. The 24 cap u e o he da a se ies was ca ied ou by a LiDAR senso . The 16 lase beams o he senso 25 we e able o ully scan a ho se, pe o ming a 3D econs uc ion o he ho se's side, h ough 26 which body measu emen s we e ob ained. Fi e Pu a Raza Española ho ses (PRE) (3 s allions 27 and 2 ma es) wi h ages anging be ween 5 and 18 yea s old we e scanned. The PRE is he 28 mos ecognized na i e Spanish ho se popula ion o i s census (na ional and in e na ional), 29 2 cul u al and socioeconomic impo ance. Fo each ho se, 17 zoome ic measu emen s (linea 30 and angula ) we e aken bo h manually and using he LiDAR-based sys em o check he 31 use ullness o his non-in asi e echnology in ob aining quick li es ock measu emen s while 32 causing minimal s ess o he animals. O he 17 zoome ic measu emen s ob ained manually 33 and wi h he senso , 10 (58.82%) had a mean ela i e e o ha anged be ween > 0 and <10; 5 34 (29.41%) had an e o ha anged ≥10 and <20; and wo (As and AC ) had an e o ≥20 35 (11.76%). A o al o 82.5% o he ai s s udied had an accu acy ( 2) lowe han 5%. The e o e, 36 al hough his app oach could s ill be imp o ed, i e i ies he iabili y o noncon ac 37 measu emen s o la ge li es ock. 38 39 Keywo ds. Zoome ic measu emen s, digi al h ee-dimensional modelling, LiDAR senso , 40 Velodyne VLP-16 41 1. In oduc ion 42 The use o emo e sensing in ag icul u e is no new; i da es back decades. Howe e , 43 ecen echnological ad ances in senso s, ins umen a ion, and wi eless ne wo ks ha e b ough 44 inno a i e le els o moni o ing in o aising li es ock (Guo e al., 2017, 2019; Le Cozle e al., 45 2019). Senso s measu ing animal wel a e, heal h, g ow h a e, and physiological ea u es, 46 including shape, size, and weigh , ei he alone o in combina ions, ha e been applied in animal 47 b eeding. Kawasue e al. (2013) used h ee Kinec senso s o in es iga e he weigh and size o 48 ca le, as well as hei pos u es and shapes, wi h accu acies o up o 93% ela i e o manual 49 measu emen s. O he au ho s ha e also e alua ed he body shape and empe a u e o black 50 ca le using a he mal came a and a Kinec senso (Kawasue e al., 2017), he co ela ion 51 be ween manual measu emen s and s e eo ision in ho ses (Pallo ino e al., 2015), and he 52 3 combina ion o a binocula s e eo ision sys em wi h he LabView de elopmen pla o m o 53 es ima e pig body size and weigh in indoo a m condi ions (Shi e al., 2016). 54 The ho se indus y plays a pa in na ional, s a e, and local economies in Eu ope, No h 55 Ame ica, and Canada (C oss, 2019). The equine sec o is di e se, in ol ing ag icul u e, 56 business, ec ea ion, acing, compe i ions, and he gene a ion o specialized skills and gene al 57 employmen ac oss he boa d. B eeding, compe i ion, and leisu e ac i i ies in ol ing ho ses a e 58 impo an business conce ns, and while ho ses a e no longe used o p ima y anspo , hey 59 emain impo an asse s. Al hough no cu en ly popula in mos coun ies, he consump ion o 60 ho se mea is inc easing in se e al Wes e n Eu opean coun ies due o i s a ailabili y and 61 ecognized nu i ional alue and mainly as an al e na i e o he adi ional consump ion o ed 62 mea (Belaunza an e al., 2015). 63 Equine p oduc ion is one o he main ag icul u al ac i i ies in Spain, esponsible o 0.5% o 64 he g oss na ional p oduc (Sánchez-Gue e o e al., 2018), wi h he Pu a Raza Española ho se 65 (PRE) being he mos impo an b eed in e ms o his o ical census (328,706 ho ses in 65 66 coun ies) and economic impac on in e na ional ade (Sánchez-Gue e o e al., 2016). Typical 67 gene ic selec ion p og ammes a e based on unc ionali y (d essage) and mo phological 68 cha ac e is ics. The e o e, ob aining accu a e zoome ic measu emen s, cha ac e ized by high 69 epea abili y and low a iance among obse e s, can be conside ed ideal condi ions in he 70 con ex o an animal b eeding p og amme (Duensign e al., 2014). In he case o ho ses, he 71 close connec ion be ween biome ic pa ame e s, locomo ion cha ac e is ics, and spo s 72 pe o mance is o en he only ool o gene ic imp o emen (Pallo ino e al., 2015) and as an 73 indica o o unc ional longe i y. In ecen decades, indi ec con o ma ion assessmen sys ems 74 based on linea mo phological ai s ha e been implemen ed o many equine b eeds (Kuhnke 75 e al., 2019; Folla e al., 2019). This sys em has a se ies o ad an ages, as he sco ing scale o 76 each egion co e s i s biological ange and a wide ange o nume ical classes can be used, 77 which allows subsequen s a is ical ea men o he da a in a con inuous scale. Howe e , highly 78 4 ained quali ie s a e equi ed o apply he linea mo phological sys em, and he ela ionship 79 be ween zoome ic measu emen s and he linea scale used has no been es ablished. The 80 main goal o he PRE b eeding p og amme is o imp o e no only animal unc ionali y bu also 81 i s con o ma ion o spo pe o mance (Sánchez-Gue e o e al., 2016). The e o e, in he PRE, 82 he impo ance o con o ma ion has been shown in s udies ca ied ou o demons a e i s 83 ela ionship wi h d essage (Sánchez-Gue e o e al., 2017; Solé e al., 2013), unc ional 84 longe i y (Solé e al., 2017), and gene ic imp o emen o mo pho unc ional ai s (Sánchez-85 Gue e o e al., 2017). 86 In he las decade, se e al compu e ized echniques ha e been p oposed o he collec ion 87 o biome ic da a h ough images (Wu e al., 2004; Viazzi e al., 2014). No el h ee-dimensional 88 sys ems can sol e he p oblems posed by con en ional wo-dimensional ision sys ems, 89 including s e eo pho og amme ic echniques. Howe e , hese pho og amme ic sys ems a e 90 di icul o implemen . 91 In ecen yea s, new echniques ha p o ide poin clouds in a as , non-in asi e and 92 inexpensi e way, such as he Kinec 2® (Mic oso Co p., Redmond, Washing on) o X ion P o 93 (USUSTeK COMPUTER Inc., Taipei, Taiwan), as well as in a ed (IR) ligh (Salau e al., 2017; 94 Kawasue e al., 2017; Viazzi e al., 2014; Mo ensen e al., 2016; Guo e al., 2017; Pezzuolo e 95 al., 2018; Song e al., 2018; Guo e al., 2019), a e being used o de elop new me hod o 96 au oma e he collec ion o mo phological in o ma ion om li es ock. In ac , Pezzuolo e al. 97 (2019) pe o med a me ological analysis o he S uc u e om Mo ion (S M) app oach, low-cos 98 LiDAR scanning, and he Mic oso Kinec 1 dep h came a o h ee-dimensional measu emen 99 o an animal's body, wi h speci ic e e ence o pigs. The esul s ob ained demons a ed he high 100 po en ial o he 3D Kinec and a highe oo mean squa e (RMS) alue o LiDAR wi h espec o 101 Kinec and S M, p obably due o he collec ion app oach based on indi idual p o iles ins ead o 102 su aces. 103 5 Howe e , despi e p og ess in he h ee-dimensional econs uc ion o he bodies o animals, 104 no obus desc ip o o he au oma ic es ima ion o con o ma ion has cu en ly been ound in 105 equine species. The e o e, he main objec i e o his s udy is o explo e a 3D econs uc ion 106 app oach o body measu emen s in ho ses using poin cloud analysis. Ou speci ic objec i es 107 we e (1) o de elop a po able scanning sys em o domes ica ed animals; (2) o ob ain aluable 108 zoo echnical in o ma ion h ough 3D poin cloud analysis ( ha would a oiding was ing ime and 109 money); and (3) o compa e he accu acy be ween he digi al and con en ional manual 110 measu emen s on i e PRE ho ses. 111 112 2. Ma e ials and me hods 113 2.1. Animals and mo phological a iables unde s udy 114 Fi e PRE ho ses, belonging o he b eeding p og amme o he “Yeguada Ca ujana del 115 Hie o del Bocado” (Je ez de la F on e a, Cádiz, Spain), si ua ed a la i ude 36°41′ no h and 116 longi ude 06°09′ wes , we e analysed (Table 1). 117 Table 1. Ages and coa colou s o he i e Pu a Raza Española ho ses used in his s udy. 118 S allion 1 S allion 2 S allion 3 Ma e 1 Ma e 2 Age 5 yea s old 6 yea s old 8 yea s old 12 yea s old 18 yea s old Coa Colou G ey Bay G ey G ey G ey 119 The animals had been used o b eeding and d essage exhibi ions. To educe possible 120 a ia ions, he zoome ic measu emen s aken by he manual me hod we e always ca ied ou 121 by he same e e ina ian. Fo each ho se, 17 zoome ic measu emen s (15 linea and 2 122 angula ) ela ing o spo pe o mance (Sánchez-Gue e o e al., 2016, 2017) we e collec ed 123 (Fig. 1). 124 6 Figu e. 1. G aphical ep esen a ion o he mo phological measu emen s aken in he Pu a Raza Espanola s allion 1 o he Yeguada Ca ujana del Hie o del Bocado. Heigh a he wi he s (HW); Heigh o he lowes poin a he wi he s (HWl); Heigh a he c oup (HC ); Leng h o he body (LB); Leng h o he head (LH); Wid h o he head (WH); Leng h o he neck (LN); Leng h o he shoulde (LS); Leng h o he o ea m (LFA); Do so-s e num diame e (DSD); Leng h o he c oup (LC ); Leng h o he emu (LF); Dep h o he c oup (DC ); Leng h o he gaskin (LG); Leng h o he bu ock (LBu); Angle o he shoulde (AS); Angle o he c oup (AC ). 125 The mo phological measu emen s aken we e as ollows: 126 1. Heigh a he wi he s (HW): measu ed om he g ound o he highes poin o he 127 wi he s. 128 2. Heigh o he lowes poin a he wi he s (HWl): measu ed om he g ound o he lowes 129 poin o he wi he s. 130 3. Heigh a he c oup (HC ): measu ed om he g ound o he highes poin o he ube 131 coxae. 132 4. Leng h o he body (LB): measu ed be ween he ube cle o he hume us and he ischial 133 ube osi y. 134 7 5. Leng h o he head (LH): dis ance om he nape o he c anial bo de o he snou . 135 6. Wid h o he head (WH): dis ance be ween he mos p o uding edge o he zygoma ic 136 a ches. 137 7. Leng h o he neck (LN): dis ance be ween he base o he ea and he middle poin o 138 he spine o he scapula. 139 8. Leng h o he scapula (LS): dis ance be ween he wi he and shoulde . 140 9. Leng h o he o ea m (LFA): dis ance be ween he pa allel s aigh lines d awn down 141 om he elbow and ca pal join midpoin . 142 10. Do so-s e num diame e (DSD): dis ance measu ed om he lowes poin in he wi he 143 decline o he s e nal a ea. 144 11. Leng h o he c oup (LC ): dis ance be ween he coxal ube osi y a i s midpoin and he 145 ischial ube osi y. 146 12. Leng h o he emu (LF): dis ance be ween he bu ock and s i le 147 13. Dep h o he c oup (DC ): dis ance be ween he hip and s i le. 148 14. Leng h o he gaskin (LG): dis ance be ween he s i le and he hock 149 15. Leng h o he bu ock (LBu): dis ance be ween he coxal ube osi y o he ilium and he 150 ischial ube osi y. 151 16. Angle o he shoulde (AS): angle o med by he line om he wi he s o he shoulde 152 wi h he ho izon al. 153 17. Angle o he c oup (AC ): angle o med by he line om he ischial ube osi y o he 154 ube coxae wi h he ho izon al. 155 156 2.2. Da a acquisi ion and da a p ocessing 157 2.2.1 Con en ional measu emen sys em 158 8 Fo each ho se, zoome ic measu emen s we e sys ema ically collec ed using s anda d 159 measu ing s icks, non-elas ic measu ing ape and zoome ic compasses (Sánchez-Gue e o e 160 al., 2016). All measu emen s we e aken om he le side o he ho se while i was s anding on 161 a ha d su ace and la g ound, assuming a na u al posi ion. The ho ses we e posi ioned o 162 measu emen wi h he on legs and hind ee pa allel and as nea o pe pendicula as possible; 163 he oes we e in line. No seda i es we e used. 164 165 2.2.2. Measu emen s based on he poin cloud sys em 166 The zoome ic measu emen s we e acqui ed h ough an op ical emo e sensing echnique 167 using a LiDAR senso . The senso uses lase ligh o ob ain a dense sampling o he a ge , 168 p oducing accu a e measu emen s in h ee dimensions. The e lec ion o lase ligh o he a ge 169 is de ec ed and analysed by he ecei e s in he LiDAR senso . These ecei e s, consis ing o a 170 ecei e elescope, il e s, and an op ical de ec o , eco d he p ecise ime be ween when he 171 lase pulse le he sys em and when i e u ned o calcula e he limi dis ance be ween he 172 senso and he a ge . These spa ially o ganized pos -p ocessed LiDAR da a a e known as poin 173 cloud da a, which is a la ge da a se consis ing o 3D poin da a. Each LiDAR poin can ha e an 174 assigned classi ica ion de ining he ype o objec ha e lec ed he lase pulse. 175 The LiDAR used was a Velodyne VLP-16 (Velodyne Lida , Inc., Cali o nia, EEUU), which is 176 a high-p ecision 3D lase senso LiDAR wi h a o a ing head con aining a ce ain numbe o 177 semiconduc o lase s o lase diodes. Each o he lase s has i s own de ec o (Fig. 2). 178 9 Figu e 2. Rep esen a ion o he lase s o he Velodyne VLP-16 179 The VLP-16 senso measu es he e lec i i y o an objec wi h 256-bi esolu ion 180 independen o lase powe and dis ance wi hin an in e al o 1-100 m. Comme cially a ailable 181 e lec i i y s anda ds we e used o absolu e e lec i i y calib a ion, which a e s o ed in a 182 calib a ion able wi hin he ield-p og ammable ga e a ay (FPGA) o he VLP-16. 183 The VLP-16 scanne has 16 indi idual lase s/de ec o s a anged in a 30° FOV, which yields 184 a e ical esolu ion o 2.0°. This senso has an FOV symme ical wi h he espec o he 185 ho izon al plane, and poin s can be ob ained up o 100 m away a a a e o app oxima ely 186 300,000 pe second in single e u n and 600,000 o dual e u n. The ho izon al FOV is 360°, 187 wi h an adjus able o a ion equency be ween 5 and 20 Hz. 188 Thanks o he di e gence o he lase beam, a single sho can hi mul iple objec s, and 189 di e en e u ns will occu . The VLP-16 has he abili y o analsze he mul iple e u ns and epo 190 he s onges e u n, he las e u n, o bo h (dual e u n o dual mode). Mul iple e u ns occu 191 when a lase pulse s ikes he ho se in a loca ion ha does no comple ely block he pa h o he 192 pulse, allowing he emaining po ion o he pulse o con inue o he nex seen objec . 193 Communica ion be ween he senso and he compu e pe o ming he da a analysis was 194 pe o med was ca ied ou h ough an in e ace box wi h an E he ne cables and use -assigned 195 IP add esses. 196 16 Table 2. Zoome ic measu emen s, ela i e e o , 2 e o and Pea son co ela ions in i e Pu a Raza Española ho ses om Es epe Ca ujana 286 ob ained using he con en ional p ocedu e (manual) and he LiDAR sys em (senso ) 287 M S allions 1 S allions 2 S allions 3 Ma e 1 Ma e 2 Global 2 m s .e. m s .e. m s .e. m s .e. m s .e. HW 156 160 -2.56 159 149 6.29 164 156 4.88 160 149 6.88 163 150 7.98 1.14 -0.32 HWl 151 141 6.62 150 148 1.33 153 155 -1.31 149 146 2.01 153 150 1.96 3.59 0.60 HC 158 145 8.23 159 155 2.52 159 140 11.95 160 154 3.75 162 161 0.62 23.17 0.77 LB 156 150 3.85 155 155 0.00 159 153 3.77 159 158 0.63 170 156 8.24 0.86 0.39 LH 60 63 -5.00 58 57 1.72 59 62 -5.08 61 57 6.56 62 60 3.23 0.52 0.06 WH 23 29 -26.09 23 22 4.35 24 23 4.17 21 17 19.05 23 25 -8.70 3.92 0.64 LN 74 78 -5.41 73 70 4.11 79 73 7.59 78 78 0.00 78 75 3.85 0.64 0.25 LS 62 53 14.52 62 54 12.9 62 67 -8,06 66 73 10.61 69 70 -1.45 0.98 0.71 LFA 54 54 0.00 47 56 19.15 49 40 18.37 44 37 15,91 47 51 8.51 1.07 0.48 DSD 72 54 25.00 71 54 23.94 74 62 16.22 74 66 10.81 77 67 12.99 3.58 0.90 LC 56 50 10.71 54 54 0.00 52 51 1.92 52 54 3.85 53 55 3.77 1.86 -0.45 LF 53 44 16.98 50 42 16.00 51 48 5.88 48 41 14.58 45 42 6.67 0.86 0.53 DCR 52 48 7.69 49 39 20.41 52 53 1.92 52 35 32.69 53 49 7.55 17.70 0.47 LG 51 64 25.49 48 49 2.08 50 50 0 57 62 8.77 54 55 1.85 2.32 0.62 LBu 47 45 4.26 45 41 8.89 48 47 2.08 42 39 7.14 45 47 4.44 0.25 0.79 AS 57 66 15.79 57 86 50.88 56 62 10.71 58 83 43.1 57 65 14.04 230.60 0.66 AC 14 25 78.57 18 10 44.44 18 23 27.78 17 21 23.53 18 20 11.11 4.17 -0.55 Measu emen s (M); con en ional p ocedu e (m) and he LiDAR sys em (s); ela i e e o ( .e.); Pea son co ela ions coe icien ( ); MHeigh a wi he s1 (HW); 288 Heigh a c oup1 (HC ); Leng h o head1 (LH); Wid h o head1 (WH); Leng h o neck1 (LN); Leng h o shoulde 1 (LS); Leng h o o ea m1 (LFA); Leng h o body1 289 (LB); Leng h o c oup1 (LC ); Leng h o emu 1 (LF); Leng h o gaskin1 (LG); Dep h o c oup1 (DC ); Leng h o bu ock1 (LBu); Do so-s e num diame e 1 (DSD); 290 Angle o shoulde 2 (AS); Angle o c oup2 (AC ). 1Cen im e s 2Deg ees 291 17 Size and body con o ma ion a e c i ically impo an ai s in nea ly all ho ses b eeds, which 292 a e p esumably subjec o a s ic p ocess o selec ion o e ime (B ooks e al., 2010). The 293 zoome ic measu emen s eco ded du ing his s udy we e simila o hose ob ained p e iously in 294 he same b eed (Molina e al., 1999; Gómez e al., 2009). Al hough he bay PRE ho ses a e he 295 alles in his s udy, he heigh o s allion 2 was be ween ha o s allions 1 and 3. 296 The measu emen s we e aken bo h in ma es and s allions o measu e he possible 297 di e ence in accu acy o he measu emen a he ho se's mane; in ma es, he mane is usually 298 cu , and in s allions i is usually le long (Fig. 1 and 4). In ou esul s, s allions had a highe 299 ela i e e o han ma es (10.83 and 8.80, espec i ely) and a highe a e age 2 alue. 300 Among he h ee s allions, he ela i e e o anged be ween 0 (HC, LFA, LC and LC) and 301 78.57 (Ac ). A o al o 7.84% o he zoome ic measu emen s s udied had a ela i e e o o 0; 302 o 52.94% had an e o > 0 and <10; o 21.57% had an e o ≥10 and <20; o 11.76% had an 303 e o ≥20 and <30; and o 5.88% had an e o ≥ 30. Fo he wo ma es s udied, he ela i e 304 e o anged be ween 0 (LN) and 43.10 (AS). A o al o 2.94% o he zoome ic measu emen s 305 s udied had a ela i e e o o 0; o 64.71% had an e o > 0 and <10; o 23.53% had an e o 306 ≥10 and <20; o 2.94% had an e o ≥20 and <30; and o 5.88% had an e o ≥ 30 (Table 2). 307 The mean ela i e e o was 10.83% and 8.00% wi hou angle measu emen s. A o al o 82.5% 308 o he s udied ea u es had an accu acy ( 2) o less han 5%; only one o he angle 309 meau emen s (AS) had a highe 2 alue. The Pea son co ela ion be ween he manual and 310 senso measu emen s anged om 0.90 (DSD) o 311 -0.55 (Ac ) o -0.45 (LC ) i he angles we e no aken in o accoun (Table 3). The angles a e, in 312 gene al, e y di icul o measu e also in he adi ional sys em and in he linea mo phological 313 assesmen s since hey a e in luenced by he ho se's pos u e (Sánchez e al., 2013). We 314 sugges ha pe haps his p oblem could be sol ed wi h he use o LiDAR o a simila echnique, 315 since he scan could be pe o med when he animal is pe ec ly poised, and in eal li e, he 316 ho se is always mo ing o some deg ee while zoome ic measu emen s a e being aken. The R² 317 18 and RMSE es ima ed by compa ing he zoome ic measu emen s ob ained using he 318 con en ional p ocedu e (manual) and he LiDAR sys em (senso ) we e 0.97 and 7.62, 319 espec i ely (Figu e 8). 320 y = 2.4251 + 0.94268 * x = 0.9861; R 2 =0.9723; RMSE=7.6172 020 40 60 80 100 120 140 160 180 Manual zoome ic measu emen 0 20 40 60 80 100 120 140 160 180 LIDAR zoome ic measu emen 0.95 Con idence In e al 321 Figu e 8. A sca e plo wi h eg ession analysis be ween he coe icien o de e mina ion (R²) and oo -322 mean-squa e e o (RMSE) compa ing zoome ic measu emen s ob ained o i e Pu a Raza Española 323 ho ses om Es epe Ca ujana pe o med using he con en ional p ocedu e (manual) and he LiDAR 324 sys em (senso ) 325 326 Thus, ou esul s di e om hose epo ed o di e en body measu es on Medi e anean 327 bu aloes (Neg e i e al., 2007), co ela ing adi ional measu es wi h p edic ed ones using 328 image analysis and ob aining coe icien s anging be ween 0.91 and 0.99 depending on he 329 measu ed ai . Mo eo e , Pallo ino e al. (2015) co ela ed adi ional measu emen s wi h hose 330 ob ained wi h a dual web-came a sys em ( = 0.998) in Lipizzan ho ses. In any case, I is 331 19 impo an o highligh ha he Pea son co ela ion coe icien s a e no he bes way o de ec he 332 di e ences be ween wo me hods o measu emen , since we a e no as in e es ed in whe he 333 hey inc ease o dec ease in he same di ec ion. The main in e es is o ensu e ha he 334 di e ence be ween he wo measu es is as small as possible. This could be be e es ima ed 335 wi h he ela i e e o . Fo he body size measu emen s o sheep, a low-cos dual web came a 336 was used, yielding a mean ela i e e o o 5% (Menesa i e al., 2014). In addi ion, 16 adul 337 dai y cows we e s udied wi h pai s o pic u es, aken wi h a po able ins umen inco po a ing 338 wo synch onized came as, wi h pho og amme ic measu emen s ha ing an accu acy wi hin 0.5 339 cm (Gaudioso e al., 2014). The use o a s uc u ed ligh dep h came a o h ee-dimensional 340 body measu emen s o dai y cows in ee-s all ba ns has also been e alua ed (Pezzuolo e al., 341 2018), ob aining coe icien s o de e mina ion o R2 > 0.84 and de ia ions lowe han 6% wi h 342 espec o manual measu emen s. As in ou case, lowe pe o mances we e ob ained o he 343 angle measu emen s (back slope: R2 = 0.12). In ou s udy, he AS measu emen had a e y 344 high 2 alue, while AC had a low 2 alue in bo h gende s. This is he i s s udy in which 17 345 zoome ic measu emen s we e analysed, wi h he majo i y o he p e ious s udies ha ing ewe 346 ai s analysed ( anging om 3 o 8 body measu emen s (Huang e al., 2018; Pezzuolo e al., 347 2018)). Taking accu a e manual measu emen da a as alida ion c i e ia, LiDAR me hodology 348 was used in ano he s udy o measu e h ee li e ca le, and he expe imen al esul s showed 349 ha he inal de ia ions we e close o 2 mm and wi hin app oxima ely 2% (Huang e al., 2018). 350 The ep oducibili y o hese ai s in he linea mo phological sys em anged om 0.86 o 0.99 in 351 PRE (Sánchez e al., 2013). In s udying ano he g oup o zoome ic measu emen s, Gaudioso 352 e al. (2014) ound a s anda d de ia ion o 2.0 cm. I shold be no ed ha we ha e no been able 353 o s udy he p ecision o hese 17 zoomome ic measu emen s since eplica ions we e no 354 ca ied ou . 355 Once he me hodology p oposed in his p ojec has been p esen ed and he esul s 356 ob ained om his p ocess ha e been analysed, he inal de elopmen o wo conclusions can 357 20 be p esen ed, one om a gene ic poin o iew and ano he om a speci ic poin o iew, in 358 addi ion o a oadmap o u u e de elopmen s. 359 4. Conclusions 360 A gene al me hodology has been c ea ed ha es ablishes a da a e alua ion p ocess, so ha in 361 he u u e, applica ions can be de eloped in which we can encompass he en i e p ocess. The 362 use o a LiDAR sys em has simpli ied he aking o measu emen s, gi ing ise o a me hod ha 363 can ob ain op imal esul s in moni o ing he g ow h o he li es ock. This me hodology speci ies 364 a clea , p ecise and logical p ocess. 365 In heo y, his me hod can be ex apola ed o any equine b eed, as well as o o he li es ock 366 species. Al hough i is a pilo s udy and i s accu acy mus be imp o ed, i could help in 367 p e en ing excess spending and s essing animals wi h he cu en manual zoome ic 368 measu emen . Mo e speci icically, he ollowing conclusions can be d awn based on he esul s 369 ob ained om implemen ing he me hodology. The app oach ca ied ou in his wo k e i ies he 370 iabili y o he noncon ac measu emen o la ge li es ock. Knowledge o he scalable g ow h 371 and animal wel a e a he loca ion whe e he animals a e housed will imp o e he quali y o he 372 animals and gene ic b eeding. Howe e , as can be seen in he esul s sec ion, due o he 373 cons an mo emen o he ho ses and he limi a ion o single- ame analysis, i is no possible o 374 ully analyse he h ee-dimensional image o he ho se. The geome ic accu acy o hese 375 senso s is closely ela ed o he quali y o hei ex e nal o ien a ion. This ex e nal o ien a ion is 376 wha we will ob ain om he in eg a ed na iga ion sys ems (GNSS and INS). The lase scanne 377 measu es only he ec o o ien ed om he opening o he lase sys em o an objec poin 378 no mally on he g ound, he ho se. Th ee-dimensional poin s can only be calcula ed i a any 379 momen he posi ion and o ien a ion o he LiDAR senso wi h espec o a coo dina e sys em is 380 known. A u u e s udy will y o imp o e hese d awbacks by in eg a ing wo synch onized poin 381 21 clouds h ough wo LiDARs posi ioned on each side o he animal and pe o ming epea ed 382 measu emen s wi h he con en ional p ocedu e (manual) and he LiDAR sys em (senso ). 383 Fo his p oblem, in he nea u u e his sys em will con inue o be op imized by ac i a ing he 384 LiDAR senso and an in eg a ed na iga ion sys em ha will allow he senso o mo e so ha he 385 ho se can be o a ed while i is being scanned. This will u he allow a comple e econs uc ion 386 o he ho se in 3D o g ea e accu acy in ob aining he zoome ic a iables. 387 388 Acknowledgemen s 389 The esea ch was suppo ed by p ojec AGL2016-78964-R unded by he Spanish Minis y o 390 Economics and Compe ence. Addi ionally, he au ho s wan o hank he Yeguada Ca ujana-391 Hie o del Bocado in Je ez (Cádiz) o p o iding us wi h hei acili ies and access o hei 392 animals. 393 Re e ences 394 Belaunza an, X., Bessa, R.J.B., La ín, P., Man ecón, A.R., K ame , J.K.G., Aldai, N., 2015. 395 Ho se-mea o human consump ion - Cu en esea ch and u u e oppo uni ies (Re iew). 396 Mea Science. 108, pp. 74-81. h ps://doi.o g/10.1016/j.mea sci.2015.05.006 397 B ooks, S.A., Mak andi-Nejad, S., Chu, E., Allen, J.J., S ee e , C., Gu, E., McClee y, B., 398 Mu phy, B.A., Bellone, R., Su e , N.B., 2010. Mo phological a ia ion in he ho se: 399 de ining complex ai s o body size and shape. Animal Gene ics. 41(2), pp.159–65. 400 h ps://doi.o g/10.1111/j.1365-2052.2010.02127.x 401 C oss, P., 2019. Global Ho se s a is ics in e nal. Global Ho se s a is ics. HPAB (HiPoin Ag o 402 Bedding Co p.) Helping he Plane h ough Animal Biodi e si y. Global Ho se. 403 Duensing, J., S ock, K.F., K ie e , J., 2014. Implemen a ion and p ospec s o linea p o iling in 404 he wa mblood ho se. Jou nal o Equine Ve e ina y Science. 34, pp. 360–368. 405 h ps://doi.o g/10.1016/j.je s.2013.09.002 406 Folla, F., Sa o i, C., Guzzo, N., Pigozzi, G., Man o ani, R., 2019. Gene ics o linea ype ai s 407 sco ed on young oals belonging o he I alian Hea y D augh Ho se b eed. Li es ock 408 Science. 219, pp. 91–96. h ps://doi.o g/10.1016/j.li sci.2018.11.019 409 Gaudioso, V., Sanz-Ablanedo, E., Lomillos, J.M., Alonso, E., Ja a es-Mo illo, L., Rod íguez, P., 410 2014. “Pho ozoome e ”: A new pho og amme ic sys em o ob aining mo phome ic 411 22 measu emen s o elusi e animals. Li es ock Science.165, pp 147–156. 412 h ps://doi.o g/10.1016/j.li sci.2014.03.028 413 Gómez, M.D., Goyache, F., Molina, A., Vale a, M., 2009. Si e × s ud in e ac ion o body 414 measu emen ai s in Spanish Pu eb ed ho ses. Jou nal o Animal Science. 87, pp. 2502–415 2509. h ps://doi.o g/10.2527/jas.2008-0841 416 Guo, H., Ma, X., Ma, Q., Wang, K., Su, W., Zhu, D., 2017. LSSA_CAU: An in e ac i e 3d poin 417 clouds analysis so wa e o body measu emen o li es ock wi h simila o ms o cows o 418 pigs. Compu e s and Elec onics in Ag icul u e. 138, pp. 60–68. 419 h ps://doi.o g/10.1016/j.compag.2017.04.014 420 Guo, H., Bo Li, Z., Ma, Q., Zhu, D., Su, W., Wang, K., Ma inello, F., 2019. A bila e al symme y 421 based pose no maliza ion amewo k applied o li es ock body measu emen in poin 422 clouds. Compu e s and Elec onics in Ag icul u e. 160, pp. 59-70. h ps://doi.o g/ 423 10.1016/j.compag.2019.03.010 424 Huang, L., Li, S., Zhu, A., Fan, X., Zhang, C., Wang, H., 2018. Non-con ac body measu emen 425 o qinchuan ca le wi h LiDAR senso . Senso s. 18(9), pp. 3014. 426 h ps://doi.o g/10.3390/s18093014 427 Kawasue, K., Ikeda, T., Tokunaga, T., Ha ada, H., 2013. Th ee-dimensional shape 428 measu emen sys em o black ca le using KINECT senso . In e na ional Jou nal o 429 Ci cui s, Sys em and Signal P ocessing. 7, pp. 222–230. 430 Kawasue, K., Win, K.D., Yoshida, K., Tokunaga, T., 2017. Black ca le body shape and 431 empe a u e measu emen using he mog aphy and KINECT senso . A i icial Li e and 432 Robo ics. 22, pp. 464–470. h ps://doi.o g/10.1007/s10015-017-0373-2 433 Kuhnke, S., Bä , K., Bosch, P., Rensing, M., Bo s el, U.K.V., 2019. E alua ion o a No el 434 Sys em o Linea Con o ma ion, Gai , and Pe sonali y T ai Sco ing and Au oma ic 435 Ranking o Ho ses a B eed Shows: A Pilo S udy in Ame ican Qua e Ho ses. Jou nal o 436 Equine Ve e ina y Science. 78, pp. 53–59. h ps://doi.o g/10.1016/j.je s.2019.04.002 437 Le Cozle , Y., Allain, C., Caillo , A., Deloua d, J.M., Dela e, L., Luginbuhl, T., Fa e din, P., 438 2019. High-p ecision scanning sys em o comple e 3D cow body shape imaging and 439 analysis o mo phological ai s. Compu e s and Elec onics in Ag icul u e. 157, pp. 447–440 453. h ps://doi.o g/10.1016/j.compag.2019.01.019 441 Menesa i, P., Cos a, C., An onucci, F., S e i, R., Pallo ino, F., Ca illo, G., 2014. A low-cos 442 s e eo ision sys em o es ima e size and weigh o li e sheep. Compu e s and Elec onics 443 in Ag icul u e. 103, pp. 33–38. h ps://doi.o g/10.1016/j.compag.2014.01.018 444 Molina, A., Vale a, M., Dos San os, R., Rode o, A., 1999. Gene ic pa ame e s o 445 mo pho unc ional ai s in Andalusian ho se. Li es ock P oduc ion Science. 60, pp. 295–446 303. h ps://doi.o g/10.1016/S0301-6226(99)00101-3 447 Mo ensen, A.K., Lisouski, P., Ah end , P., 2016. Weigh p edic ion o b oile chickens using 3d 448 compu e ision. Compu e s and Elec onics in Ag iculu e. 123, pp. 319–326. 449 h ps://doi.o g/10.1016/j.compag.2016.03.011 450 Neg e i, P., Bianconi, G., Ba occi, S., Te amoccia, S., Ve na, M., 2008. De e mina ion o li e 451 weigh and body condi ion sco e in lac a ing Medi e anean bu alo by Visual Image 452 Analysis. Li es ock Science. 113, pp. 1–7. h ps://doi.o g/10.1016/j.li sci.2007.05.018 453 Pallo ino, F., S e i, R., Menesa i, P., An onucci, F., Cos a, C., Figo illi, S., Ca illo, G., 2015. 454 Compa ison be ween manual and s e eo ision body ai s measu emen s o Lipizzan 455 23 ho ses. Compu e s and Elec onics in Ag icul u e. 118, pp. 408–413. 456 h ps://doi.o g/10.1016/j.compag.2015.09.019 457 Pezzuolo, A., Gua ino, M., Sa o i, L., Ma inello, F., 2018. A Feasibili y s udy on he use o a 458 s uc u ed ligh dep h-came a o h ee-dimensional body measu emen s o dai y cows in 459 ee-s all ba ns. Senso s. 18 (2), pp. 673. h ps://doi.o g/10.3390/s18020673 460 P eisinge , R., Wilkens, J., Kalm, E., 1991. Es ima ion o gene ic pa ame e s and b eeding 461 alues o con o ma ion ai s o oals and ma es in he T akehne popula ion and hei 462 p ac ical implica ions. Li es ock P oduc ion Science. 29 (1), pp. 77-86. 463 h ps://doi.o g/10.1016/0301-6226(91)90121-6 464 Poly, J., Pou ous, M., Calomi i, S., Suzanne C., 1967. Le con ole lai ie mensual al e né (AT). 465 I._P écision is-a- is d’un con ole mensual ou bimes iel pou la p oduci on “de lai en 466 305 jou s”. Ann. Zoo ech. 16, pp. 183–190. 467 Sánchez, M.J., Gómez, M.D., Molina, A., Vale a, M., 2013. Gene ic analyses o linea 468 con o ma ion ai s in Pu a Raza Español ho ses. Li es ock Science. 157, pp. 57–64. 469 h ps://doi.o g/10.1016/j.li sci.2013.07.010 470 Sánchez-Gue e o, M.J., Molina, A., Gómez, M.D., Peña, F., Vale a, M., 2016. Rela ionship 471 be ween mo phology and pe o mance: Signa u e o mass-selec ion in Pu a Raza Español 472 ho se. Li es ock Science. 185, pp.148–155. h ps://doi.o g/10.1016/j.li sci.2016.01.003 473 Sánchez-Gue e o, M.J., Ce an es, I., Molina, A., Gu ié ez, J.P., Vale a, M., 2017. Designing 474 an ea ly selec ion mo phological linea ai s index o d essage in he Pu a Raza Español 475 ho se. Animal. 11(6), pp. 948-957. h ps://doi.o g/10.1017/S1751731116002214 476 Sánchez-Gue e o, M.J., Neg o-Rama, S., Daymas-Pey ás, S., Solé-Be ga, M., Azo -O iz, P.J., 477 Vale a-Có doba, M., 2018. Mo phological and gene ic di e si y o Pu a Raza Español 478 ho se ega ding o he coa colou . Animal Science Jou nal. 90, pp.14–22. 479 h ps://doi.o g/10.1111/asj.13102 480 Salau, J., Haas, J.H., Junge, W., Thalle , G., 2017. A mul i-kinec cow scanning sys em: 481 calcula ing linea ai s om manually ma ked eco dings o hols ein- iesian dai y cows. 482 Biosys em Enginee ing. 157, pp. 92–98. 483 h ps://doi.o g/10.1016/j.biosys emseng.2017.03.001 484 Shi, C., Teng, G., Li, Z., 2016. An app oach o pig weigh es ima ion using binocula s e eo 485 sys em based on LabVIEW. Compu e s and Elec onics in Ag icul u e. 129, pp. 37–43. 486 h ps://doi.o g/10.1016/j.compag.2016.08.012 487 Solé, M., San os, R., Gómez, M.D., Galis eo, A.M., Vale a, M., 2013. E alua ion o con o ma ion 488 agains ai s associa ed wi h d essage abili y in un idden Ibe ian ho ses a he o . 489 Resea ch in Ve e ina y Science. 95, pp. 660–666. 490 h ps://doi.o g/10.1016/j. sc.2013.06.017 491 Solé, M., Sánchez, M.J., Vale a, M., Molina, A., Azo , P.J., Sölkne , J., Mészá os, G., 2017. 492 Assessmen o spo i e longe i y in Pu a Raza Español d essage ho ses. Li es ock 493 Science. 203, pp. 69-75. h ps://doi.o g/10.1016/j.li sci.2017.07.007 494 Song, X., Bokke s, E.A.M., an de Tol, P.P.J., Koe kamp, P.W.G.G., an Mou ik, S., 2018. 495 Au oma ed body weigh p edic ion o dai y cows using 3-dimensional ision. Jou nal o 496 Dai y Science.101 (5), pp. 4448–4459. h ps://doi.o g/10.3168/jds.2017-13094 497 Viazzi, S., Ismayilo a, G., Oczak, M., Sonoda, L.T., Fels, M., Gua ino, M., V anken, E., Ha ung, 498 J., Bah , C., Be ckmans, D., 2014. Image ea u e ex ac ion o classi ica ion o agg essi e 499 24 in e ac ions among pigs. Compu e s and Elec onics in Ag icul u e. 104, pp. 57–62. 500 h ps://doi.o g/10.1016/j.compag.2014.03.010 501 Wu, J., Tille , R., McFa lane, N., Ju, X., Siebe , J.P., Scho ield, P., 2004. Ex ac ing he h ee-502 dimensional shape o li e pigs using s e eo pho og amme y. Compu e s and Elec onics 503 in Ag icul u e. 44, pp. 203–222. h ps://doi.o g/10.1016/j.compag.2004.05.003 504 505 506