PHP tclip extension calls opencv to read and parse image data from memory

Two days ago, I said there was a PHP extension, tclip Image cropping program.
Tclip is committed to face recognition and recognition of prominent areas of images. The biggest difference between Tclip and other image cropping programs is that the face and other prominent areas of the cropped image are retained, however, the image is still cropped.
Tclip depends on the Open Source opencv Library. Encyclopedia: OpenCV Full name: Open Source Computer Vision Library. OpenCV is a (open-source)-based cross-platform computer vision library that runs on Linux, Windows, and Mac OS. It is lightweight and efficient-it consists of a series of C functions and a small number of C ++ classes, and provides interfaces for Python, Ruby, MATLAB, and other languages, it implements many common algorithms for image processing and computer vision.
Back to the topic, tclip only uses a small part of opencv, that is, recognition of the significance area of the image.
The specific cropping algorithm does not do much research.
The problem is that you want to read the binary data of the image from the memory using PHP (this part of the data is prepared first) and pass it to opencv. You can view the API documentation of opencv, you can see the imdecode method to load images from the memory. This problem can be solved.
About installing tclip extension: https://github.com/exinnet/tclip
But it should be noted that the opencv-2.4.9 should have bugs, compilation is not successful, and the ubuntu-14.0.4 after the compilation and installation call tclip function error, I tested the version is:

CentOS-6.4
PHP-5.4.31
Opencv-2.4.4

Example

Installation steps:

The code is as follows:

Copy code

Download source code: Download opencv Source: http://www.bo56.com/download/opencv2.tar.bz2 Download Tclip source code: Http://code.taobao.org/p/tclip/src/ Https://github.com/exinnet/tclip Install opencv2: This extension depends on opencv2.0. Therefore, install opencv before installation. Follow these steps to install opencv: If you have any questions during installation, please contact qq 179815944. 1. Install the following dependency package: gtk +-devel pkgconfig libpng zlib libjpeg libtiff cmake For centos, run the following command to install the dependency package: Yum install gtk +-devel pkgconfig libpng zlib libjpeg libtiff cmake 2. Install opencv2 as follows: Decompress the installation package Cd to enter the installation package folder. Cmake CMakeLists.txt Make & make install Vim/etc/profile Note: When compiling opencv, you must use the gcc version 4.4 or later. Otherwise, an error is reported. 3. Set related environment variables Add before unset I Export PKG_CONFIG_PATH =/usr/lib/pkgconfig/:/usr/local/lib/pkgconfig: $ PKG_CONFIG_PATH After you exit, run the following command: Source/etc/profile Echo "/usr/local/lib/">/etc/ld. so. conf. d/opencv. conf Ldconfig Install tclip extension for php image cropping Cd to the php_ext folder in the source code directory Phpize ./Configure Make Cp modules/tclip. so to the extension directory Modify php. ini. Add extension = tclip. so Restart fpm Install command line If you want to use the command line method, you can install Cd to enter the soft folder of the installation package Chmod + x./tclip. sh ./Tclip. sh Usage instructions First: Image cropping format in php: Tclip (original file path, save path of the cropped image, width of the cropped image, and height of the cropped image) Example: $ Source_file = "/tmp/a.jpg "; $ Dest_file = "/www/a_dest.jpg "; $ Width = 400; $ Height = 200; $ Ret = tclip ($ source_file, $ dest_file, $ width, $ height ); Execution successful $ ret is true; otherwise, it is false. Type 2: command line Parameter description: -S source image path -D save path of the cropped image -W: width of the cropped image -H: specifies the image height after cropping. ./Tclip-s a.jpg-d a_dest.jpg-w 400-h 200

Finally, although PHP can directly transmit the binary data of the image, it is found that the binary data in the opencv source code is saved to a temporary file and then read. And temporary files won't be deleted in opencv-2.4.2 versions, which is a bit pitfall. However, PHP logic can still be reduced, so it is necessary to support binary image data.