Matlab: 深度学习目标检测xml标注信息批量统计

"""

https://blog.csdn.net/gusui7202/article/details/83239142

qhy。

"""

这个是matlab用于统计目标数量的，其他更多的统计写在另一个python脚本，有如下功能：

1. #计算每一个目标在原图中的占比

2. #计算目标的平均长度、

3. #计算平均宽度，

4. #计算平均面积、

5. #计算目标平均占比

6. ......

见：https://blog.csdn.net/gusui7202/article/details/85136268

本程序如下：

程序1：xml_read()   #xml读取  

程序2：mian() #xml内容统计

使用：将两个程序放入同一个文件夹下，修改路径即可。要统计不同的信息，修改main函数就行。

我主要统计的是数据集中每一类标签的框数量。

程序1: 数据读取模块

function [tree, RootName, DOMnode] = xml_read(xmlfile, Pref)
%XML_READ reads xml files and converts them into Matlab's struct tree.
%
% DESCRIPTION
% tree = xml_read(xmlfile) reads 'xmlfile' into data structure 'tree'
%
% tree = xml_read(xmlfile, Pref) reads 'xmlfile' into data structure 'tree'
% according to your preferences
%
% [tree, RootName, DOMnode] = xml_read(xmlfile) get additional information
% about XML file
%
% INPUT:
%  xmlfile    URL or filename of xml file to read
%  Pref     Preferences:
%    Pref.ItemName - default 'item' - name of a special tag used to itemize
%                    cell arrays
%    Pref.ReadAttr - default true - allow reading attributes
%    Pref.ReadSpec - default true - allow reading special nodes
%    Pref.Str2Num  - default 'smart' - convert strings that look like numbers
%                   to numbers. Options: "always", "never", and "smart"
%    Pref.KeepNS   - default true - keep or strip namespace info
%    Pref.NoCells  - default true - force output to have no cell arrays
%    Pref.Debug    - default false - show mode specific error messages
%    Pref.NumLevels- default infinity - how many recursive levels are
%      allowed. Can be used to speed up the function by prunning the tree.
%    Pref.RootOnly - default true - output variable 'tree' corresponds to
%      xml file root element, otherwise it correspond to the whole file.
%    Pref.CellItem - default 'true' - leave 'item' nodes in cell notation.
% OUTPUT:
%  tree         tree of structs and/or cell arrays corresponding to xml file
%  RootName     XML tag name used for root (top level) node.
%               Optionally it can be a string cell array storing: Name of
%               root node, document "Processing Instructions" data and
%               document "comment" string
%  DOMnode      output of xmlread
%
% DETAILS:
% Function xml_read first calls MATLAB's xmlread function and than
% converts its output ('Document Object Model' tree of Java objects)
% to tree of MATLAB struct's. The output is in format of nested structs
% and cells. In the output data structure field names are based on
% XML tags, except in cases when tags produce illegal variable names.
%
% Several special xml node types result in special tags for fields of
% 'tree' nodes:
%  - node.CONTENT - stores data section of the node if other fields are
%    present. Usually data section is stored directly in 'node'.
%  - node.ATTRIBUTE.name - stores node's attribute called 'name'.
%  - node.COMMENT - stores node's comment section (string). For global
%    comments see "RootName" output variable.
%  - node.CDATA_SECTION - stores node's CDATA section (string).
%  - node.PROCESSING_INSTRUCTIONS - stores "processing instruction" child
%    node. For global "processing instructions" see "RootName" output variable.
%  - other special node types like: document fragment nodes, document type
%   nodes, entity nodes, notation nodes and processing instruction nodes
%   will be treated like regular nodes
%
% EXAMPLES:
%   MyTree=[];
%   MyTree.MyNumber = 13;
%   MyTree.MyString = 'Hello World';
%   xml_write('test.xml', MyTree);
%   [tree treeName] = xml_read ('test.xml');
%   disp(treeName)
%   gen_object_display()
%   % See also xml_examples.m
%
% See also:
%   xml_write, xmlread, xmlwrite
%
% Written by Jarek Tuszynski, SAIC, jaroslaw.w.tuszynski_at_saic.com
% References:
%  - Function inspired by Example 3 found in xmlread function.
%  - Output data structures inspired by xml_toolbox structures.

%% 默认设置
DPref.TableName  = {'tr','td'}; % name of a special tags used to itemize 2D cell arrays
DPref.ItemName  = 'item'; % name of a special tag used to itemize 1D cell arrays
DPref.CellItem  = false;  % leave 'item' nodes in cell notation
DPref.ReadAttr  = true;   % allow reading attributes
DPref.ReadSpec  = true;   % allow reading special nodes: comments, CData, etc.
DPref.KeepNS    = true;   % Keep or strip namespace info
DPref.Str2Num   = 'smart';% convert strings that look like numbers to numbers
DPref.NoCells   = true;   % force output to have no cell arrays
DPref.NumLevels = 1e10;   % number of recurence levels
DPref.PreserveSpace = false; % Preserve or delete spaces at the beggining and the end of stings?
RootOnly        = true;   % return root node  with no top level special nodes
Debug           = false;  % show specific errors (true) or general (false)?
tree            = [];
RootName        = [];

%% 检查matlab版本
v = ver('MATLAB');
version = str2double(regexp(v.Version, '\d.\d','match','once'));
if (version<7.1)
  error('Your MATLAB version is too old. You need version 7.1 or newer.');
end

%%读取配置参数
if (nargin>1)
  if (isfield(Pref, 'TableName')), DPref.TableName = Pref.TableName; end
  if (isfield(Pref, 'ItemName' )), DPref.ItemName  = Pref.ItemName;  end
  if (isfield(Pref, 'CellItem' )), DPref.CellItem  = Pref.CellItem;  end
  if (isfield(Pref, 'Str2Num'  )), DPref.Str2Num   = Pref.Str2Num ;  end
  if (isfield(Pref, 'NoCells'  )), DPref.NoCells   = Pref.NoCells ;  end
  if (isfield(Pref, 'NumLevels')), DPref.NumLevels = Pref.NumLevels; end
  if (isfield(Pref, 'ReadAttr' )), DPref.ReadAttr  = Pref.ReadAttr;  end
  if (isfield(Pref, 'ReadSpec' )), DPref.ReadSpec  = Pref.ReadSpec;  end
  if (isfield(Pref, 'KeepNS'   )), DPref.KeepNS    = Pref.KeepNS;    end
  if (isfield(Pref, 'RootOnly' )), RootOnly        = Pref.RootOnly;  end
  if (isfield(Pref, 'Debug'    )), Debug           = Pref.Debug   ;  end
  if (isfield(Pref, 'PreserveSpace')), DPref.PreserveSpace = Pref.PreserveSpace; end
end
if ischar(DPref.Str2Num), % convert from character description to numbers
  DPref.Str2Num = find(strcmpi(DPref.Str2Num, {'never', 'smart', 'always'}))-1;
  if isempty(DPref.Str2Num), DPref.Str2Num=1; end % 1-smart by default
end

%% 读取xml
if isa(xmlfile, 'org.apache.xerces.dom.DeferredDocumentImpl');
  % if xmlfile is a DOMnode than skip the call to xmlread
  try
    try    # try-catch主要用来查找用户里面的错误，或者可以判断是否存在变量的使用
      DOMnode = xmlfile;
    catch ME
      error('Invalid DOM node: \n%s.', getReport(ME));
    end
  catch %#ok<CTCH> catch for mablab versions prior to 7.5
    error('Invalid DOM node. \n');
  end
else         % we assume xmlfile is a filename
  if (Debug) % in debuging mode crashes are allowed
    DOMnode = xmlread(xmlfile);
  else       % in normal mode crashes are not allowed
    try
      try
        DOMnode = xmlread(xmlfile);