PLY文件格式及其MATLAB读写操作

     PLY是一种电脑档案格式，全名为多边形档案（Polygon File Format）或 斯坦福三角形档案（Stanford Triangle Format）。 史丹佛大学的 The Digital Michelangelo Project计划采用PLY格式储存极高分辨率之米开朗基罗的作品"大卫"雕塑。该格式主要用以储存立体扫描结果的三维数值，透过多边形片面的集合描述三维物体，与其他格式相较之下这是较为简单的方法。它可以储存的资讯包含颜色、透明度、表面法向量、材质座标与资料可信度，并能对多边形的正反两面设定不同的属性。在档案内容的储存上PLY有两种版本，分别是纯文字（ASCII）版本与二元码（binary）版本，其差异在储存时是否以ASCII编码表示元素资讯。

档案格式

（本文并未提供完整的格式描述，以下仅介绍PLY的基本概念与格式）

每个PLY档都包含档头（header），用以设定网格模型的“元素”与“属性”，以及在档头下方接着一连串的元素“数值资料”。一般而言，网格模型的“元素”就是顶点（vertices）、面（faces），另外还可能包含有边（edges）、深度图样本（samples of range maps）与三角带（triangle strips）等元素。无论是纯文字与二元码的PLY档，档头资讯都是以ASCII编码编写，接续其后的数值资料才有编码之分。PLY档案以此行：

ply

开头作为PLY格式的识别。接着第二行是版本资讯，目前有三种写法：

format ascii 1.0

format binary_little_endian 1.0

format binary_big_endian 1.0

其中ascii, binary_little_endian, binary_big_endian是档案储存的编码方式，而1.0是遵循的标准版本（现阶段仅有PLY 1.0版）。在档头中可使用’comment’作为一行的开头以编写注解，例如：

comment Thisisa comment!

描述元素及属性，必须使用’element’及’property’的关键字，一般的格式为element下方接着属性列表，例如:

element

property

‘property’不仅定义了资料的型态，其出现顺序亦定义了资料的顺序。内定的资料形态有两种写法：一种是char uchar short ushort int uint float double,另外一种是具有位元长度的int8 uint8 int16 uint16 int32 uint32 float32 float64。例如，描述一个包含12个顶点的物体，每个顶点使用3个单精度浮点数 (x,y,z）代表点的座标，使用3个unsigned char代表顶点颜色，颜色顺序为 (B, G, R),则档头的写法为：

element vertex 12

propertyfloatx

propertyfloaty

propertyfloatz

property uchar blue

property uchar green

property uchar red

其中vertex是内定的元素类型，接续的6行property描述构成vertex元素的数值字段顺序代表的意义，及其资料形态。

另一个常使用的元素是面。由于一个面是由3个以上的顶点所组成，因此使用一个“顶点列表”即可描述一个面, PLY格式使用一个特殊关键字’property list’定义之。例如，一个具有10个面的物体，其PLY档头可能包含：

element face 10

property list ucharintvertex_indices

‘property list’表示该元素face的特性是由一行的顶点列表来描述。列表开头以uchar型态的数值表示列表的项目数，后面接着资料型态为int的顶点索引值（vertex_indices），顶点索引值从0开始。

最后，标头必须以此行结尾：

end_header

档头后接着的是元素资料（端点座标、拓朴连结等）。在ASCII格式中各个端点与面的资讯都是以独立的一行描述，而二元编码格式则连续储存这些资料，加载时须以’element’定义的元素数目以及’property’中设定的资料形态计算各笔字段的长度。

范例

一个典型的PLY档案结构分成三部分：

档头 (从ply开始到end_header）

顶点元素列表

面元素列表

其中的顶点元素列表一般以x y z方式排列，形态如档头所定义；而面元素列表是以下列格式表示。

<組成面的端點數N> <端點#1的索引> <端點#2的索引> … <端點#N的索引>

例如画出一个有4个顶点，4个面的四面体，档案内容为:

ply

format ascii 1.0

comment這是一個正四面體

element vertex 4

propertyfloatx

propertyfloaty

propertyfloatz

element face 4

property list ucharintvertex_index

end_header

0 3 0

2.449 -1.0 -1.414

0 -1 2.828

-2.449 -1.0 -1.414

3 0 1 3

3 0 2 1

3 0 3 2

3 1 2 3

其中1~10行是档头, 11~14行是顶点元素列表, 15~18行则是面元素列表。

其中： 0 3 0是顶点

历史

PLY格式发展于90年代中期，在史丹佛大学图学实验室的Marc Levoy教授指导下，由Greg Turk及其他成员开发出来。PLY格式受Wavefront .obj格式的启发，但改进了Obj格式所缺少的对任意属性及群组的扩充性。因此PLY格式发明了”property”及”element”这两个关键词，来概括“顶点、面、相关资讯、群组”的概念。

注意

ply文件不支持中文格式的文件名字,所以在使用过程中避免使用中文来命名。

使用MATLAB对PLY文件进行读操作

function [ Elements, varargout ] = PLY_READ ( Path, Str )%*****************************************************************************80
%
%% PLY_READ reads a PLY 3D data file.
%
%   [DATA,COMMENTS] = PLY_READ(FILENAME) reads a version 1.0 PLY file
%   FILENAME and returns a structure DATA.  The fields in this structure
%   are defined by the PLY header; each element type is a field and each
%   element property is a subfield.  If the file contains any comments,
%   they are returned in a cell string array COMMENTS.
%
%   [TRI,PTS] = PLY_READ(FILENAME,'tri') or
%   [TRI,PTS,DATA,COMMENTS] = PLY_READ(FILENAME,'tri') converts vertex
%   and face data into triangular connectivity and vertex arrays.  The
%   mesh can then be displayed using the TRISURF command.
%
%   Note: This function is slow for large mesh files (+50K faces),
%   especially when reading data with list type properties.
%
%   Example:
%   [Tri,Pts] = PLY_READ('cow.ply','tri');
%   [Tri,Pts] = PLY_READ('bunny.ply','tri');
%   trisurf(Tri,Pts(:,1),Pts(:,2),Pts(:,3));
%   colormap(gray); axis equal;
%
%  Discussion:
%
%    The original version of this program had a mistake that meant it
%    did not properly triangulate files whose faces were not already triangular.
%    This has been corrected (JVB, 25 February 2007).
%
%    Glenn Ramsey pointed out and corrected a problem that occurred
%    with an uninitialized value of Type2, 27 August 2012.
%
%  Licensing:
%
%    This code is distributed under the GNU LGPL license.
%
%  Modified:
%
%    27 August 2012
%
%  Author:
%
%    Pascal Getreuer 2004
%
%  Parameters:
%
%  Local Parameters:
%
%    COMMENTS, any comments from the file.
%
%    ELEMENTCOUNT, the number of each type of element in file.
%
%    ELEMENTS, the element data.
%
%    PROPERTYTYPES, the element property types.
%
%    SIZEOF, size in bytes of each type.
%%
%  Open the input file in "read text" mode.
%[ fid, Msg ] = fopen ( Path, 'rt' );if ( fid == -1 )error ( Msg );endBuf = fscanf ( fid, '%s', 1 );if ( ~strcmp ( Buf, 'ply' ) )fclose ( fid );error('Not a PLY file.');end
%
%  Read the header.
%Position = ftell(fid);Format = '';NumComments = 0;Comments = {};NumElements = 0;NumProperties = 0;Elements = [];ElementCount = [];PropertyTypes = [];ElementNames = {};  % list of element names in the order they are stored in the filePropertyNames = [];  % structure of lists of property nameswhile ( 1 )
%
%  Read a line from the file.
%Buf = fgetl ( fid );BufRem = Buf;Token = {};Count = 0;
%
%  Split the line into tokens.
%while ( ~isempty(BufRem) )[ tmp, BufRem ] = strtok(BufRem);
%
%  Count the tokens.
%if ( ~isempty ( tmp ) )Count = Count + 1;Token{Count} = tmp;endend
%
%  Parse the line.
%if ( Count )switch lower ( Token{1} )
%
%  Read the data format.
%case 'format'if ( 2 <= Count )Format = lower ( Token{2} );if ( Count == 3 & ~strcmp ( Token{3}, '1.0' ) )fclose ( fid );error('Only PLY format version 1.0 supported.');endend
%
%  Read a comment.
%case 'comment'NumComments = NumComments + 1;Comments{NumComments} = '';for i = 2 : CountComments{NumComments} = [Comments{NumComments},Token{i},' '];end
%
%  Read an element name.
%case 'element'if ( 3 <= Count )if ( isfield(Elements,Token{2}) )fclose ( fid );error(['Duplicate element name, ''',Token{2},'''.']);endNumElements = NumElements + 1;NumProperties = 0;Elements = setfield(Elements,Token{2},[]);PropertyTypes = setfield(PropertyTypes,Token{2},[]);ElementNames{NumElements} = Token{2};PropertyNames = setfield(PropertyNames,Token{2},{});CurElement = Token{2};ElementCount(NumElements) = str2double(Token{3});if ( isnan(ElementCount(NumElements)) )fclose ( fid );error(['Bad element definition: ',Buf]);endelseerror(['Bad element definition: ',Buf]);end
%
%  Read an element property.
%case 'property'if ( ~isempty(CurElement) & Count >= 3 )NumProperties = NumProperties + 1;eval(['tmp=isfield(Elements.',CurElement,',Token{Count});'],...'fclose(fid);error([''Error reading property: '',Buf])');if ( tmp )error(['Duplicate property name, ''',CurElement,'.',Token{2},'''.']);end
%
%  Add property subfield to Elements.
%eval(['Elements.',CurElement,'.',Token{Count},'=[];'], ...'fclose(fid);error([''Error reading property: '',Buf])');
%
%  Add property subfield to PropertyTypes and save type.
%eval(['PropertyTypes.',CurElement,'.',Token{Count},'={Token{2:Count-1}};'], ...'fclose(fid);error([''Error reading property: '',Buf])');
%
%  Record property name order.
%eval(['PropertyNames.',CurElement,'{NumProperties}=Token{Count};'], ...'fclose(fid);error([''Error reading property: '',Buf])');elsefclose ( fid );if ( isempty(CurElement) )error(['Property definition without element definition: ',Buf]);elseerror(['Bad property definition: ',Buf]);endend
%
%  End of header.
%case 'end_header'break;endendend
%
%  Set reading for specified data format.
%if ( isempty ( Format ) )warning('Data format unspecified, assuming ASCII.');Format = 'ascii';endswitch Formatcase 'ascii'Format = 0;case 'binary_little_endian'Format = 1;case 'binary_big_endian'Format = 2;otherwisefclose ( fid );error(['Data format ''',Format,''' not supported.']);end
%
%  Read the rest of the file as ASCII data...
%if ( ~Format )Buf = fscanf ( fid, '%f' );BufOff = 1;else
%
%  ...or, close the file, and reopen in "read binary" mode.
%fclose ( fid );
%
%  Reopen the binary file as LITTLE_ENDIAN or BIG_ENDIAN.
%if ( Format == 1 )fid = fopen ( Path, 'r', 'ieee-le.l64' );elsefid = fopen ( Path, 'r', 'ieee-be.l64' );end
%
%  Find the end of the header again.
%  Using ftell on the old handle doesn't give the correct position.
%BufSize = 8192;Buf = [ blanks(10), char(fread(fid,BufSize,'uchar')') ];i = [];tmp = -11;while ( isempty(i) )i = findstr(Buf,['end_header',13,10]);   % look for end_header + CR/LFi = [i,findstr(Buf,['end_header',10])];  % look for end_header + LFif ( isempty(i) )tmp = tmp + BufSize;Buf = [Buf(BufSize+1:BufSize+10),char(fread(fid,BufSize,'uchar')')];endend
%
%  seek to just after the line feed
%fseek ( fid, i + tmp + 11 + (Buf(i + 10) == 13), -1 );end
%
%  Read element data.
%
%  PLY and MATLAB data types (for fread)
%PlyTypeNames = {'char','uchar','short','ushort','int','uint','float','double', ...'char8','uchar8','short16','ushort16','int32','uint32','float32','double64'};MatlabTypeNames = {'schar','uchar','int16','uint16','int32','uint32','single','double'};SizeOf = [1,1,2,2,4,4,4,8];for i = 1 : NumElements
%
%  get current element property information
%eval(['CurPropertyNames=PropertyNames.',ElementNames{i},';']);eval(['CurPropertyTypes=PropertyTypes.',ElementNames{i},';']);NumProperties = size(CurPropertyNames,2);%   fprintf('Reading %s...\n',ElementNames{i});
%
%  Read ASCII data.
%if ( ~Format )for j = 1 : NumPropertiesToken = getfield(CurPropertyTypes,CurPropertyNames{j});if ( strcmpi(Token{1},'list') )Type(j) = 1;elseType(j) = 0;end
%
%  Glenn Ramsey 20120827
%  Initialise Type2{} to prevent uninitialised value error.
%Type2{j} = '';end
%
%  Parse the buffer.
%if ( ~any(Type) )% no list typesData = reshape ( ...Buf(BufOff:BufOff+ElementCount(i)*NumProperties-1), ...NumProperties, ElementCount(i) )';BufOff = BufOff + ElementCount(i) * NumProperties;elseListData = cell(NumProperties,1);for k = 1 : NumPropertiesListData{k} = cell(ElementCount(i),1);end
%
% list type
%for j = 1 : ElementCount(i)for k = 1 : NumPropertiesif ( ~Type(k) )Data(j,k) = Buf(BufOff);BufOff = BufOff + 1;elsetmp = Buf(BufOff);ListData{k}{j} = Buf(BufOff+(1:tmp))';BufOff = BufOff + tmp + 1;endendendend
%
%  Read binary data.
%else
% translate PLY data type names to MATLAB data type namesListFlag = 0;  % = 1 if there is a list typeSameFlag = 1;     % = 1 if all types are the samefor j = 1 : NumPropertiesToken = getfield(CurPropertyTypes,CurPropertyNames{j});
%
%  Non-list type.
%if ( ~strcmp(Token{1},'list' ) )tmp = rem(strmatch(Token{1},PlyTypeNames,'exact')-1,8)+1;if ( ~isempty(tmp) )TypeSize(j) = SizeOf(tmp);Type{j} = MatlabTypeNames{tmp};TypeSize2(j) = 0;Type2{j} = '';SameFlag = SameFlag & strcmp(Type{1},Type{j});elsefclose(fid);error(['Unknown property data type, ''',Token{1},''', in ', ...ElementNames{i},'.',CurPropertyNames{j},'.']);endelse           % list typeif ( length(Token) == 3 )ListFlag = 1;SameFlag = 0;tmp = rem(strmatch(Token{2},PlyTypeNames,'exact')-1,8)+1;tmp2 = rem(strmatch(Token{3},PlyTypeNames,'exact')-1,8)+1;if ( ~isempty(tmp) & ~isempty(tmp2) )TypeSize(j) = SizeOf(tmp);Type{j} = MatlabTypeNames{tmp};TypeSize2(j) = SizeOf(tmp2);Type2{j} = MatlabTypeNames{tmp2};elsefclose(fid);error(['Unknown property data type, ''list ',Token{2},' ',Token{3},''', in ', ...ElementNames{i},'.',CurPropertyNames{j},'.']);endelsefclose(fid);error(['Invalid list syntax in ',ElementNames{i},'.',CurPropertyNames{j},'.']);endendend% read fileif ( ~ListFlag )
%
%  No list types, all the same type (fast)
%if ( SameFlag )Data = fread(fid,[NumProperties,ElementCount(i)],Type{1})';
%
%  No list types, mixed type.
%elseData = zeros(ElementCount(i),NumProperties);for j = 1 : ElementCount(i)for k = 1 : NumPropertiesData(j,k) = fread(fid,1,Type{k});endendendelseListData = cell(NumProperties,1);for k = 1 : NumPropertiesListData{k} = cell(ElementCount(i),1);endif ( NumProperties == 1 )BufSize = 512;SkipNum = 4;j = 0;
%
%  List type, one property (fast if lists are usually the same length)
%while ( j < ElementCount(i) )BufSize = min(ElementCount(i)-j,BufSize);Position = ftell(fid);
%
%  Read in BufSize count values, assuming all counts = SkipNum
%[Buf,BufSize] = fread(fid,BufSize,Type{1},SkipNum*TypeSize2(1));Miss = find(Buf ~= SkipNum);     % find first count that is not SkipNumfseek(fid,Position + TypeSize(1),-1);   % seek back to after first countif ( isempty(Miss) )
% all counts are SkipNumBuf = fread(fid,[SkipNum,BufSize],[int2str(SkipNum),'*',Type2{1}],TypeSize(1))';fseek(fid,-TypeSize(1),0);     % undo last skipfor k = 1:BufSizeListData{1}{j+k} = Buf(k,:);endj = j + BufSize;BufSize = floor(1.5*BufSize);else
%
%  Some counts are SkipNum.
%if ( 1 < Miss(1) )Buf2 = fread(fid,[SkipNum,Miss(1)-1],[int2str(SkipNum),'*',Type2{1}],TypeSize(1))';for k = 1:Miss(1)-1ListData{1}{j+k} = Buf2(k,:);endj = j + k;end
%
%  Read in the list with the missed count.
%SkipNum = Buf(Miss(1));j = j + 1;ListData{1}{j} = fread(fid,[1,SkipNum],Type2{1});BufSize = ceil(0.6*BufSize);endendelse
%
%  List type(s), multiple properties (slow)
%Data = zeros(ElementCount(i),NumProperties);for j = 1:ElementCount(i)for k = 1:NumPropertiesif ( isempty(Type2{k}) )Data(j,k) = fread(fid,1,Type{k});elsetmp = fread(fid,1,Type{k});ListData{k}{j} = fread(fid,[1,tmp],Type2{k});endendendendendend
%
%  Put data into Elements structure
%for k = 1 : NumPropertiesif ( ( ~Format & ~Type(k) ) || (Format & isempty(Type2{k})) )eval(['Elements.',ElementNames{i},'.',CurPropertyNames{k},'=Data(:,k);']);elseeval(['Elements.',ElementNames{i},'.',CurPropertyNames{k},'=ListData{k};']);endendendclear Dataclear ListData;fclose ( fid );
%
%  Output the data as a triangular mesh pair.
%if ( ( nargin > 1 & strcmpi(Str,'Tri') ) || nargout > 2 )
%
%  Find vertex element field
%Name = {'vertex','Vertex','point','Point','pts','Pts'};Names = [];for i = 1 : length(Name)if ( any ( strcmp ( ElementNames, Name{i} ) ) )Names = getfield(PropertyNames,Name{i});Name = Name{i};break;endendif ( any(strcmp(Names,'x')) & any(strcmp(Names,'y')) & any(strcmp(Names,'z')) )eval(['varargout{1}=[Elements.',Name,'.x,Elements.',Name,'.y,Elements.',Name,'.z];']);elsevarargout{1} = zeros(1,3);endvarargout{1} = varargout{1}';varargout{2} = Elements;varargout{3} = Comments;Elements = [];
%
% Find face element field
%Name = {'face','Face','poly','Poly','tri','Tri'};Names = [];for i = 1 : length(Name)if ( any(strcmp(ElementNames,Name{i})) )Names = getfield(PropertyNames,Name{i});Name = Name{i};break;endendif ( ~isempty(Names) )% find vertex indices property subfieldPropertyName = {'vertex_indices','vertex_indexes','vertex_index','indices','indexes'};for i = 1 : length(PropertyName)if ( any(strcmp(Names,PropertyName{i})) )PropertyName = PropertyName{i};break;endend
%
%  Convert face index list to triangular connectivity.
%if ( ~iscell(PropertyName) )eval(['FaceIndices=varargout{2}.',Name,'.',PropertyName,';']);N = length(FaceIndices);Elements = zeros(3,N*2);Extra = 0;for k = 1 : NElements(1:3,k) = FaceIndices{k}(1:3)';
%
%  The original code had an error in the following loop.
%for j = 4 : length(FaceIndices{k})Extra = Extra + 1;Elements(1,N + Extra) = FaceIndices{k}(1);Elements(2,N + Extra) = FaceIndices{k}(j-1);Elements(3,N + Extra) = FaceIndices{k}(j);endend
%
%  Add 1 to each vertex value; PLY vertices are zero based.
%Elements = Elements(:,1:N+Extra) + 1;endendelsevarargout{1} = Comments;endreturn
end**使用MATLAB对PLY文件进行写操作**
实现将矩阵的内容转换到封装的结构体中

function [Data1,Data2]=change(A,B)
Data1.vertex.x = A(:,1);
Data1.vertex.y = A(:,2);
Data1.vertex.z = A(:,3);
Data2.vertex.x = B(:,1);
Data2.vertex.y = B(:,2);
Data2.vertex.z = B(:,3);

function ply_write ( Elements, Path, Format, Str )%*****************************************************************************80
%
%% PLY_WRITE writes 3D data as a PLY file.
%
%  Discussion:
%
%    PLY_WRITE(DATA,FILENAME) writes the structure DATA as a binary
%    PLY file.  Every field of DATA is interpreted as an element
%    and every subfield as an element property.  Each subfield of
%    property data must either be an array or a cell array of
%    arrays.  All property data in an element must have the same
%    length.
%
%    A common PLY data structure has the following fields:
%      DATA.vertex.x = x coordinates, [Nx1] real array
%      DATA.vertex.y = y coordinates, [Nx1] real array
%      DATA.vertex.z = z coordinates, [Nx1] real array
%
%      DATA.face.vertex_indices = vertex index lists,
%         an {Mx1} cell array where each cell holds a one-
%         dimesional array (of any length) of vertex indices.
%
%    Some other common data fields:
%
%      DATA.vertex.nx = x coordinate of normal, [Nx1] real array
%      DATA.vertex.ny = y coordinate of normal, [Nx1] real array
%      DATA.vertex.nz = z coordinate of normal, [Nx1] real array
%
%      DATA.edge.vertex1 = index to a vertex, [Px1] integer array
%      DATA.edge.vertex2 = second vertex index, [Px1] integer array
%
%    Many other fields and properties can be added.  The PLY format
%    is not limited to the naming in the examples above -- they are
%    only the conventional naming.
%
%    PLY_WRITE(DATA,FILENAME,FORMAT) write the PLY with a specified
%    data format, where FORMAT is
%      'ascii'                  ASCII text data
%      'binary_little_endian'   binary data, little endian
%      'binary_big_endian'      binary data, big endian (default)
%
%    PLY_WRITE(DATA,FILENAME,FORMAT,'double') or
%
%    PLY_WRITE(DATA,FILENAME,'double') write floating-point data as
%    double precision rather than in the default single precision.
%
%  Example:
%
%    % make a cube
%    clear Data;
%    Data.vertex.x = [0;0;0;0;1;1;1;1];
%    Data.vertex.y = [0;0;1;1;0;0;1;1];
%    Data.vertex.z = [0;1;1;0;0;1;1;0];
%    Data.face.vertex_indices = {[0,1,2,3],[7,6,5,4], ...
%         [0,4,5,1],[1,5,6,2],[2,6,7,3],[3,7,4,0]};
%    ply_write(Data,'cube.ply','ascii');
%  Licensing:
%
%    This code is distributed under the GNU LGPL license.
%
%  Modified:
%
%    01 July 2016    Seth Billings   bug fix: 'ushort' / 'uint16' data type was
%                                    not included in the min/max value arrays; now
%                                    excluding 'ushort' from the integer data types
%                                    (rather than fixing the min/max arrays), since
%                                    some programs such as MeshLab do not handle the
%                                    'ushort' data type properly anyway
%
%    01 July 2016    Seth Billings   bug fixes for max data value of 'short' / 'int16'
%                                    data type and for selection of output data type
%    01 March 2007
%
%  Author:
%
%    Pascal Getreuer 2004
%    Seth Billings 2016
%if ( nargin < 4 )Str = '';if ( nargin < 3 )Format = 'binary_big_endian';elseif ( strcmpi ( Format, 'double' ) )Str = 'double';Format = 'binary_big_endian';endend[ fid, Msg ] = fopen ( Path, 'wt' );if ( fid == -1 )error(Msg);end
%
%  Bug Fix: Seth Billings
%  - correcting max integer value for 'short' / 'int16' from 2^16-1 to 2^15-1
%  - excluding 'ushort' from integer types since MeshLab does not
%    read 'ushort' data type properly.
%PlyTypeNames = {'char','uchar','short','int','uint','float','double', ...'char8','uchar8','short16','int32','uint32','float32','double64'};FWriteTypeNames = {'schar','uchar','int16','int32','uint32','single','double'};MatlabTypeNames = {'int8','uint8','int16','int32','uint32','single','double'};PrintfTypeChar = {'%d','%u','%d','%d','%u','%-.6f','%-.14e'};IntegerDataMin =  [-128,   0, -2^15,  -2^31,   0];IntegerDataMax =  [ 127, 255,  2^15-1, 2^31-1, 2^32-1];numTypes = length(MatlabTypeNames);%PlyTypeNames = {'char','uchar','short','ushort','int','uint','float','double', ...%  'char8','uchar8','short16','ushort16','int32','uint32','float32','double64'};%FWriteTypeNames = {'schar','uchar','int16','uint16','int32','uint32','single','double'};%MatlabTypeNames = {'int8','uint8','int16','uint16','int32','uint32','single','double'};%PrintfTypeChar = {'%d','%u','%d','%u','%d','%u','%-.6f','%-.14e'};  %IntegerDataMin = [-128, 0, -2^15, -2^31, 0];%IntegerDataMax = [127, 255, 2^16-1, 2^31-1, 2^32-1];
%
%  Write the PLY header.
%fprintf(fid,'ply\nformat %s 1.0\ncomment created by MATLAB ply_write\n',Format);ElementNames = fieldnames(Elements);NumElements = length(ElementNames);Data = cell(NumElements,1);for i = 1 : NumElementseval(['tmp=isa(Elements.',ElementNames{i},',''struct'');']);if ( tmp )eval(['PropertyNames{i}=fieldnames(Elements.',ElementNames{i},');']);elsePropertyNames{i} = [];endif ( ~isempty(PropertyNames{i}) )eval(['Data{i}{1}=Elements.',ElementNames{i},'.',PropertyNames{i}{1},';']);ElementCount(i) = prod(size(Data{i}{1}));Type{i} = zeros(length(PropertyNames{i}),1);elseElementCount(i) = 0;endfprintf(fid,'element %s %u\n',ElementNames{i},ElementCount(i));for j = 1 : length(PropertyNames{i})eval(['Data{i}{j}=Elements.',ElementNames{i},'.',PropertyNames{i}{j},';']);if ( ElementCount(i) ~= prod(size(Data{i}{j})) )fclose(fid);error('All property data in an element must have the same length.');endif ( iscell(Data{i}{j}) )Type{i}(j) = numTypes + 1;  % Seth Billings%Type{i}(j) = 9;Data{i}{j} = Data{i}{j}{1};endfor k = 1 : length(MatlabTypeNames)if ( isa(Data{i}{j},MatlabTypeNames{k}) )Type{i}(j) = Type{i}(j) + k;break;endendif ( ~rem(Type{i}(j),numTypes+1) ) % Seth Billings%if ( ~rem(Type{i}(j),9) )fclose(fid);error('Unsupported data structure.');end
%
%  Try to convert float data to integer data
%
%  Array data.
%if ( Type{i}(j) <= numTypes)  % Seth Billings%if ( Type{i}(j) <= 8 )if any(strcmp({'single','double'},MatlabTypeNames{Type{i}(j)}))if ~any(floor(Data{i}{j}) ~= Data{i}{j})  % data is integerMinValue = min(min(Data{i}{j}));MaxValue = max(max(Data{i}{j}));% choose smallest possible integer data format            % Bug Fix: Seth Billingstmp = min(find(MinValue >= IntegerDataMin & MaxValue <= IntegerDataMax));%tmp = max(min(find(MinValue >= IntegerDataMin)),min(find(MaxValue <= IntegerDataMax)));if ~isempty(tmp)Type{i}(j) = tmp;endendendelse        % cell array dataeval(['Data{i}{j}=Elements.',ElementNames{i},'.',PropertyNames{i}{j},';']);tmp = 1;for k = 1:prod(size(Data{i}{j}))tmp = tmp & all(floor(Data{i}{j}{k}) == Data{i}{j}{k});endif tmp  % data is integerMinValue = inf;MaxValue = -inf;for k = 1:prod(size(Data{i}{j}))MinValue = min(MinValue,min(Data{i}{j}{k}));MaxValue = max(MaxValue,max(Data{i}{j}{k}));end
%
%  choose smallest possible integer data format
%  Bug Fix: Seth Billings
%tmp = min(find(MinValue >= IntegerDataMin & MaxValue <= IntegerDataMax));%tmp = max(min(find(MinValue >= IntegerDataMin)),min(find(MaxValue <= IntegerDataMax)));if ( ~ isempty ( tmp ) )Type{i}(j) = tmp + numTypes + 1;%Type{i}(j) = tmp + 9;endendend% convert double to single if specifiedif rem(Type{i}(j),numTypes+1) == numTypes & ~strcmpi(Str,'double') % Seth Billings%if rem(Type{i}(j),9) == 8 & ~strcmpi(Str,'double')Type{i}(j) = Type{i}(j) - 1;endif Type{i}(j) <= numTypes  % Seth Billings%if Type{i}(j) <= 8fprintf(fid,'property %s %s\n',PlyTypeNames{Type{i}(j)},PropertyNames{i}{j});elsefprintf(fid,'property list uchar %s %s\n',PlyTypeNames{Type{i}(j)-(numTypes+1)},PropertyNames{i}{j}); % Seth Billings%fprintf(fid,'property list uchar %s %s\n',PlyTypeNames{Type{i}(j)-9},PropertyNames{i}{j});endendendfprintf(fid,'end_header\n');switch Formatcase 'ascii'Format = 0;case 'binary_little_endian'fclose(fid);fid = fopen(Path,'a','ieee-le');Format = 1;case 'binary_big_endian'fclose(fid);fid = fopen(Path,'a','ieee-be');Format = 2;endfor i = 1 : NumElementsif ~isempty(PropertyNames{i})if ~Format          % write ASCII datafor k = 1:ElementCount(i)for j = 1:length(PropertyNames{i})if Type{i}(j) <= numTypes  % Seth Billings%if Type{i}(j) <= 8fprintf(fid,[PrintfTypeChar{Type{i}(j)},' '],Data{i}{j}(k));elsefprintf(fid,'%u%s ',length(Data{i}{j}{k}),sprintf([' ',PrintfTypeChar{Type{i}(j)-(numTypes+1)}],Data{i}{j}{k})); % Seth Billings%fprintf(fid,'%u%s ',length(Data{i}{j}{k}),sprintf([' ',PrintfTypeChar{Type{i}(j)-9}],Data{i}{j}{k}));endendfprintf(fid,'\n');endelse            % write binary dataif all(Type{i} <= numTypes) & all(Type{i} == Type{i}(1))  % Seth Billings%if all(Type{i} <= 8) & all(Type{i} == Type{i}(1))% property data without list types (fast)tmp = zeros(length(PropertyNames{i}),ElementCount(i));for j = 1:length(PropertyNames{i})tmp(j,:) = Data{i}{j}(:)';endfwrite(fid,tmp,FWriteTypeNames{Type{i}(j)});elseif all(Type{i} > numTypes) % Seth Billings%elseif all(Type{i} > 8)% only list typesType{i} = Type{i} - (numTypes+1); % Seth Billings%Type{i} = Type{i} - 9;if length(PropertyNames{i}) == 1% only one list propertytmp = FWriteTypeNames{Type{i}(1)};for k = 1:ElementCount(i)fwrite(fid,length(Data{i}{1}{k}),'uchar');fwrite(fid,Data{i}{1}{k},tmp);endelse% multiple list propertiesfor k = 1:ElementCount(i)for j = 1:length(PropertyNames{i})fwrite(fid,length(Data{i}{j}{k}),'uchar');fwrite(fid,Data{i}{j}{k},FWriteTypeNames{Type{i}(j)});endendendelse% mixed typefor k = 1:ElementCount(i)for j = 1:length(PropertyNames{i})if Type{i}(j) <= numTypes % Seth Billings%if Type{i}(j) <= 8fwrite(fid,Data{i}{j}(k),FWriteTypeNames{Type{i}(j)});elsefwrite(fid,length(Data{i}{j}{k}),'uchar');fwrite(fid,Data{i}{j}{k},FWriteTypeNames{Type{i}(j)-(numTypes+1)}); % Seth Billings%fwrite(fid,Data{i}{j}{k},FWriteTypeNames{Type{i}(j)-9});endendendendendendendfclose ( fid );return
end

PLY文件格式及其MATLAB读写操作相关推荐

hdf5 matlab,hdf5格式的matlab读写操作
最近要用caffe处理一个multi-label的回归问题,就是输出是一个向量,不是一个具体的数值,这个时候之前的leveldb格式就不凑效了,因为caffe源代码里面默认label是一个数值,网上搜 ...
hdf5格式的matlab读写操作
最近要用caffe处理一个multi-label的回归问题,就是输出是一个向量,不是一个具体的数值,这个时候之前的leveldb格式就不凑效了,因为caffe源代码里面默认label是一个数值,网上搜 ...
matlab编写文件格式,MATLAB程序设计教程(4)——MATLAB文件操作
MATLAB程序设计教程(4)--MATLAB文件操作第4章MATLAB文件操作 4.1 文件的打开与关闭 4.2 文件的读写操作 4.3 数据文件定位 4.1 文件的打开与关闭 4.1.1 文 ...
MATLAB对ply文件格式的读取和显示
转自:https://blog.csdn.net/lafengxiaoyu/article/details/60574150 在网上搜索这个题目可以找到一些类似的文章,其来源大致都是http://pe ...
matlab用excel的数据,使用MATLAB对excel文件数据的读写操作
最近建模,刚好用matlab对excel中的数据进行处理,由于自己是个菜鸟,基本都是边问度娘边动手操作.感觉以后还会用得到,在此做个简单的总结,为下一次省点力... excel读写操作:xlsread ...
matlab对nii文件处理,使用Nibabel库对nii格式图像的读写操作
因为后期主要的研究方向是医学图像处理,而现有手头的大部分数据都是nii格式或者是hdr,img格式的数据,所以首先第一步我们需要解决图像的读写问题. 其实使用OpenCV也可以方便的进行图像读取,但是 ...
matlab %4.3f,MATLAB程序设计教程(4)—MATLAB文件操作
第4章 MATLAB文件操作 4.1 文件的打开与关闭 4.2 文件的读写操作 4.3 数据文件定位 4.1 文件的打开与关闭 4.1.1 文件的打开 fopen函数的调用格式为: fid= fope ...
QT5 文件读写操作
QFile Class 1.read读文件加载文件对象 QFile file("文件地址"); 打开加载的文件file.open(打开方式); 操作文件关闭打开的文件file ...
OpenCV FileStorage类的数据读写操作
OpenCV FileStorage类的数据读写操作 OpenCV的许多应用都需要使用数据的存储于读取,例如经过3D校准后的相机,需要存储校准结果矩阵,以方便下次调用该数据:基于机器学习的应用,同样需 ...

PLY文件格式及其MATLAB读写操作

PLY文件格式及其MATLAB读写操作相关推荐

最新文章

热门文章