source: proiecte/pmake3d/make3d_original/Make3dSingleImageStanford_version0.1/image3dstiching/README @ 37

Last change on this file since 37 was 37, checked in by (none), 14 years ago

Added original make3d

File size: 5.3 KB
Line 
1Scaling:
21) Individual model scale: the depth scale generated from Mono-inference
3        - ImgInfo.Model.Depth....
4        - ImgInfo.Model.Constrain.Depth_modified % matched triangulated depth
5        - ImgInfo.Model.PlaneParaInfo.PlanePara
6        - ImgInfo.Model.PlaneParaInfo.FitDepth
7        - ImgInfo.Model.PlaneParaInfo.Position3DFited
82) PairMatchScale
9        - Pair.T
10        - Pair.lamda
113) GlobalModelScale
12        -
13
14Scale Transform:
151) Pair.DepthScale: Individual model scale*Pair.DepthScale = PairMatchScale
162) ModelStatus.Scale: Individual model scale*ModelStatus.Scale = GlobalModelScale
17
18Storaged Data Structure
19
200) First stage data from GPS, IMU, and Google earth modification
21Description:
22        ImgInfo in the camera Exif file
23        Geo info from GPS and Modified by Google Earth
24        Orientation info from IMU and Modified by Google Earth 
25
26Path:
27        ./data/
28
29FileName:
30        ImgInfo.mat
31
32Variables:
33        ImgInfo - Struture
34                .ExifInfo - Struture
35                        .name - String, the name of the image (including file extension)
36                        ---------------
37                        all other infomation depending on your camera default
38                        ex: .FocalLength
39                        ---------------
40                .Geo - Structure
41                        .Longitude - double
42                        .Latitude - double
43                        .altitude - double
44                .Rw - 3 x 1,  [ Pitch Yaw Roll]
45                .X_world - 3 by 1, global Cartesian coordinate, calculated from Geo infomation
46Comments:       
47       
48       
491) Individual image infomation
50
51Description:
52        Infomation of indiviual image infomation
53        Start form Mono-infomation
54        Then evlove with Matches infomation
55        and Occlusion Refinement infomation
56        (Notice always maintained in the Scale of beginning Mono-Model)
57
58Path:
59        ./data/"Img1Name"/
60FileName:
61        "WrlName"_"Img1Name"_"PostFix".mat
62        (ex: Multi_IMG_0606_Mono.mat)
63
64Variables:
65        model - structure
66                (From Mono)
67                .Depth - Structure
68                        .FitDepth (nxm) double, DepthMap after Mono PlaneParameter Model fiting depths
69                        .RawDepth (nxm) double, DepthMap from image learning   
70                .Ray - 3xnxm double, unit 3-d vector indiacte the track of ray a specific image point come from
71                        ( 3DPosition = Ray(:,:).*repmat(FitDepth(:)',3,1))
72                .Sup - nxm integer, index that group each point into a Superpixel
73                .maskG - nxm logical, 1 indicated Ground
74                .maskSky - nxm logical, 1 indicated sky
75                .SupNeighborTable - MaxIndexSup x MaxIndexSup, is a look up table of which Sups are neighboring each other
76                .MultiScaleSupTable - NumUniqueSup x 14(14 type of Bigger group of points)
77
78                (From Matches)
79                .Constrain - Structure
80                        .RayMatche - NumMatches x 3, the Ray that correspoding to the matches feature points
81                        .Depth_modified - 1 x NumMatches, the triangulated depth in Indivial Model Scale
82                        .SupMatched - NumMatches x 1, the sup index of the matches
83
84                (UptoDate Model after any Re-Inference)
85                .PlaneParaInfo - Structure
86                        .PlanePara - 3 by NumUniqueSup, the plane parameter after Re-inference in Indivial Model Scale
87                        .Sup2Para - 1 by MaxIndexofSupEpand, since not all the sup have a PlaneParameter,
88                                        Sup2Para is the hash table of Sup Index to PlaneParameter Index
89                        .SupEpand - nxm integer, Superpixel index after some Process to clean the Sky Boundary
90                        .maskSkyEroded - nxm logical, new sky mask after some process
91                        .Ray -  3xnxm double, unit 3-d vector indiacte the track of ray a specific image point come from
92                                ( 3DPosition = Ray(:,:).*repmat(FitDepth(:)',3,1))     
93                        .PositionTex - n x m x 2 double, used my VRMl rendering
94                                        (ex: PositionTex(:,:,1) are the X position of the FaceSet Points in the range from 0 to 1)
95                        .FitDepth - (nxm) double, DepthMap after Mono PlaneParameter Model fiting depths
96                        .Position3DFited - (n x m x 3) double, 3D position of each FaceSet Points
97                        .SupOri - nxm integer, index that group each point into a Superpixel
98
99                (From Refinement)
100                .ConstrainOccluMatch
101Comments:       
102
103 
1042) Pair image infomation
105% ============================================================================
106Description:
107        Infomation obtained after matching pair of images
108
109Path:
110        ./data/
111
112FileName:
113        "Img1Name"_"Img2Name"_PoseMatch.mat
114
115Variables:
116        Error - integer from -1 to 6, indicating different error type when matched
117        Matches - (2 by NumMatches) integer, indicating the index of the whole surf Features Set
118        Pair - Structure
119                .R - 3x3 double, the rotation matrix from Img1 Coordinate to Img2 Coordinate
120                .T - 3x1 double, the translation matrix from Img1 Coordinate to Img2 Coordinate
121                        (ex: X2 = R*X1+T)
122                .Xim - 4 by NumMatches integer,
123                        Xim(1:2,:) indicating the (X Y) position (in pixel) of the Matched points in Img1
124                        Xim(3:4,:) indicating the (X Y) position (in pixel) of the Matched points in Img2
125                .lamda - 2 by NumMatches double,
126                        lamda(1,:) is the triangulated depths in Img1 in Pair Match Scale
127                        lamda(2,:) is the triangulated depths in Img2 in Pair Match Scale
128                .DepthScale - 2x1 double, indicate the Scaling factoe when applying the single model depths the the triangulated depths
129                                (ex: DepthScale(1)*IndivialModelDepth = PairMatchedDepths in Img1)     
130                .t -empty
131Comments:       
132        Should merge Error and Matches into Pair.Error Pair.Matches,
133        and erase Pair.t
134% =============================================================================== //Min to do
135Description:
136Path:
137        ./data/
138
139FileName:
140        "Img1Name"_"Img2Name"_OccluDetect_Ray.mat
141
142Variables:
143Comments:       
144
145% ===============================================================================
1463) Whole model infomation
147
148Template -
149
150Description:
151Path:
152FileName:
153Variables:
154Comments:       
155
Note: See TracBrowser for help on using the repository browser.