MATLAB 定时截取桌面固定区域并OCR

0. 前言

使用MATLAB实现定时截取桌面固定区域并OCR。

对于数据实时变动比较大、肉眼难易分辨的数据均可采用该方法，使用本地定时器完成。
同时，本文也可应对大批量的数据处理。

本例程对应用场景进行了简化处理，为：
固定时间间隔——60秒，截取桌面右下角时间的分钟10次。

1. screencapture下载

对于截屏，MATLAB有现成的函数，直接调用即可：
ScreenCapture - screenshot of component, figure or screen
https://www.mathworks.com/matlabcentral/fileexchange/24323-screencapture-screenshot-of-component-figure-or-screen?s_tid=srchtitle_screencapture_2

本文例程中只需要知道当前屏幕分辨率及以下使用方法即可：

% 当前屏幕分辨率1440*900
% handle为0表示截取桌面
% position: [x,y,width,height]，以左下角屏幕为起点
% imageData = screencapture(handle, position, target, 'PropName',PropValue, ...)

如果你有更多使用需要，可以查看screencapture.m头部注释。
函数不同情况下的具体使用方法screencapture.m中介绍的都比较清楚：

% screencapture - get a screen-capture of a figure frame, component handle, or screen area rectangle
%
% ScreenCapture gets a screen-capture of any Matlab GUI handle (including desktop, 
% figure, axes, image or uicontrol), or a specified area rectangle located relative to
% the specified handle. Screen area capture is possible by specifying the root (desktop)
% handle (=0). The output can be either to an image file or to a Matlab matrix (useful
% for displaying via imshow() or for further processing) or to the system clipboard.
% This utility also enables adding a toolbar button for easy interactive screen-capture.
%
% Syntax:
%    imageData = screencapture(handle, position, target, 'PropName',PropValue, ...)
%
% Input Parameters:
%    handle   - optional handle to be used for screen-capture origin.
%                 If empty/unsupplied then current figure (gcf) will be used.
%    position - optional position array in pixels: [x,y,width,height].
%                 If empty/unsupplied then the handle's position vector will be used.
%                 If both handle and position are empty/unsupplied then the position
%                   will be retrieved via interactive mouse-selection.
%                 If handle is an image, then position is in data (not pixel) units, so the
%                   captured region remains the same after figure/axes resize (like imcrop)
%    target   - optional filename for storing the screen-capture, or the
%               'clipboard'/'printer' strings.
%                 If empty/unsupplied then no output to file will be done.
%                 The file format will be determined from the extension (JPG/PNG/...).
%                 Supported formats are those supported by the imwrite function.
%    'PropName',PropValue - 
%               optional list of property pairs (e.g., 'target','myImage.png','pos',[10,20,30,40],'handle',gca)
%               PropNames may be abbreviated and are case-insensitive.
%               PropNames may also be given in whichever order.
%               Supported PropNames are:
%                 - 'handle'    (default: gcf handle)
%                 - 'position'  (default: gcf position array)
%                 - 'target'    (default: '')
%                 - 'toolbar'   (figure handle; default: gcf)
%                      this adds a screen-capture button to the figure's toolbar
%                      If this parameter is specified, then no screen-capture
%                        will take place and the returned imageData will be [].
%
% Output parameters:
%    imageData - image data in a format acceptable by the imshow function
%                  If neither target nor imageData were specified, the user will be
%                    asked to interactively specify the output file.
%
% Examples:
%    imageData = screencapture;  % interactively select screen-capture rectangle
%    imageData = screencapture(hListbox);  % capture image of a uicontrol
%    imageData = screencapture(0);         % capture image of entire screen
%    imageData = screencapture(0,  [20,30,40,50]);  % capture a small desktop region
%    imageData = screencapture(gcf,[20,30,40,50]);  % capture a small figure region
%    imageData = screencapture(gca,[10,20,30,40]);  % capture a small axes region
%      imshow(imageData);  % display the captured image in a matlab figure
%      imwrite(imageData,'myImage.png');  % save the captured image to file
%    img = imread('cameraman.tif');
%      hImg = imshow(img);
%      screencapture(hImg,[60,35,140,80]);  % capture a region of an image
%    screencapture(gcf,[],'myFigure.jpg');  % capture the entire figure into file
%    screencapture(gcf,[],'clipboard');     % capture the entire figure into clipboard
%    screencapture(gcf,[],'printer');       % print the entire figure
%    screencapture('handle',gcf,'target','myFigure.jpg'); % same as previous, save to file
%    screencapture('handle',gcf,'target','clipboard');    % same as previous, copy to clipboard
%    screencapture('handle',gcf,'target','printer');      % same as previous, send to printer
%    screencapture('toolbar',gcf);  % adds a screen-capture button to gcf's toolbar
%    screencapture('toolbar',[],'target','sc.bmp'); % same with default output filename
%
% Technical description:
%    http://UndocumentedMatlab.com/blog/screencapture-utility/

2. 实现流程

1. 设置截屏存储的文件路径
   为防止文件夹重名截屏结果被覆盖，代码采用系统当前时间为文件名，这样每次截屏的结果存放在不同的文件夹里，不会起冲突
   如识别结果不理想，可以随时查看本地图片进行校验。在代码OCR调试成功后，便可将存储文件的代码屏蔽掉，只保留最后的识别结果。
2. 设置定时器周期及执行次数
3. 截屏，并保存
4. OCR识别，并判断
   ocr函数使用见MATLAB帮助即可
   如果识别结果错误，则将数据置零

3. MATLAB代码

% Author: Shaw
% Description: 定时多次截取桌面指定区域并OCR保存为DATA
% Date: 2021/11/26


close all
clear all
clc

% 全局变量
global i DATA count path file_name

% 以当前时间创建文件夹，用以存储截屏后的图片
file_clock = clock;
file_name = ['data-', num2str(file_clock(1)), '-', num2str(file_clock(2)), '-', num2str(file_clock(3)), '-', ...
            num2str(file_clock(4)), '-', num2str(file_clock(5)), '-', num2str(round(file_clock(6)))];
path = 'E:\ScreenCapture\';

mkdir(path,file_name)


% 执行次数计数
i = 0;
% OCR异常计数
count = 0;

% 周期：单位秒
Period = 60;
% 执行次数：单位次
TasksToExecute = 10;


% 定时器设置
t = timer('Period', Period, 'TasksToExecute', TasksToExecute, 'TimerFcn', @tTimerFcn, 'ExecutionMode', 'fixedRate');
t.start


function tTimerFcn(~, ~)
    global i DATA count path file_name
    i = i + 1
     
    % 确实识别区域
    imageData = screencapture(0,[1400,0,15,30]);

    % 将图片放大8倍
    imagePro = imresize(imageData, 8);

	% 保存图片
    cd([path,file_name])
    imwrite(imagePro,['Image' , num2str(i) , '.png']);  
    cd(path)

    % OCR识别
    ocrResults = ocr(imagePro);

    % 将识别结果转化为字符串
    recognizedText = string(ocrResults.Text);
    % 剔除因图像放大而导致的字符串识别结果中的空格
    recognizedTextModify = strrep(recognizedText,' ','');
    
    % 判断识别结果
    [recognizedNum,tf] = str2num(recognizedTextModify);
     if tf
         DATA(i) = recognizedNum;
     else
         count = count + 1;
         DATA(i) = 0;
     end
     
end

4. 实现效果

存储的截屏图片为：

绘制数据DATA如下：

结语

第51篇

这部分工作虽然是一个月前做的，但中间停滞到现在并没有做完，后续如有补充则再对该文进行修改。

个人水平有限，有问题欢迎各位大神批评指正！