aardio的opencv库

Last update: Dec 31, 2022

Related tags

Overview

opencv_aardio

dll库下载地址：https://github.com/xuncv/opencv-plugin/releases

import cv2
img = cv2.imread("./images/Lena.jpg",1)
img = cv2.medianBlur(img,5)
cv2.imshow( "窗口标题",img )
cv2.imwrite("result.jpg",img)
cv2.waitKey(0)

opencv具有500多个跨平台的图像处理函数，是目前应用最广的数字图像库。opencv历史发展过程中由C接口慢慢转向C++接口，使其他语言（python除外）调用opencv的难度增大。最近找到一个针对.NET的封装库项目OpenCvSharp ，对opencv的接口进行了重新封装，对opencv版本的跟进也很及时。

虽然这个库面向.NET，但其封装的接口对aardio异常友好。使用aardio调用dll时，甚至比.NET更简单方便。

比如Mat类作为参数时，.NET的实现是添加一个CvPtr成员，用来存储Mat指针，然后每次需取一次CvPtr作为dll接口的指针，一次api调用需要封装3次。aardio的实现就非常优雅，在mat类代码中只需修改下属性表,就可以让dll直接使用aardio的mat对象

_metaProperty = ..util.metaProperty(
		_topointer = function(){
			return owner.handle; 
		}
		...
)

甚至不需要声明API接口，就可以直接调用API。

var imread = dll["imgcodecs_imread"];
imread(srcMat,1);

基于以上几点，我开始尝试使用aardio对opencv进行封装，接口实现上尽量接近opencv-python风格，降低学习成本。

aardio原生支持opencv，我希望能解决以下问题：

充分融合aardio的胶水特性，增强aardio图像处理能力。
结合aardio桌面优势，如制作上位机软件时可提高工程进度。
使opencv项目轻量化。（我的conda文件夹已经10+G了）。
提高opencv启动速度，在算法测试中提高效率。
实时窗体显示,便于调参。

2021年4月18日

Comments

下载后提未错误putTextZH为NULL

aardio：运行时错误

错误行号:#4 文件:[string "import cv2..."]: 不支持此操作:call 定义类型:method(table) 名字:'putTextZH' 类型:null

调用栈: [string "import cv2..."]:4: in main chunk

确定

opened by kwan3277 2
提示：Cannot load library

已下载DLL放置在\lib\cv2.res，运行时提示：

aardio：运行时错误

文件:[string "import cv2..."] 错误行号:#1 错误:import cv2 failed :

文件:...esktop\opencv_aardio-main\lib\cv2_.aardio 错误行号:#2 错误:Cannot load library '~/lib/cv2/.res/OpenCvSharpExtern.dll'.

调用栈: [kernel]: in function 'import' [string "import cv2..."]:1: in main chunk

确定

是不是因为DLL文件下载的不对？系统win10x64

opened by kwan3277 2
CV2默认库里的bitwise_or，bitwise_xor函数是否有个小BUG？

`bitwise_or = function(src1,src2,mask){ var dst = ..cv2.mat(); mask = mask ? (mask.cvPtr) var err = dll.core_bitwise_or(src1.cvPtr,src2.cvPtr.dst.cvPtr,null); //这里是否应该为：src2.cvPtr,dst.cvPtr,null return dst; }

bitwise_xor = function(src1,src2,mask){ var dst = ..cv2.mat(); mask = mask ? (mask.cvPtr) var err = dll.core_bitwise_xor(src.cvPtr,src2.cvPtr,dst.cvPtr,null) //这里是否应该为src1.cvPtr,src2.cvPtr,dst.cvPtr,null return dst; }`

改后调用正常，没改前总报错

opened by zhhl99 1

反馈几个问题

1、cv2.Mat 和 cv2.MatExpr 库的库文件名，首字母未改为大写（mat.aardio matExpr.aardio），好像会造成传错参数后，无法弹出错误提示，直接退出的问题。 2、从 bitmap 创建 cv2.Mat 对象时，有些像素格式的图片显示不正常，我使用以下代码显示常见像素格式图片：

    //bitmap传入
    elseif(tArg == type.table && arg[["pBitmap"]]){
        var data = arg.lockMemory(,0x26200A);
        err,ret = dll.core_Mat_new8(arg.height,arg.width,0x18/*_CV_8UC4*/,data.Scan0,data.Stride,{ptr value});
        this.handle = ret.value;
        arg.unlockMemory(data);
    }

3、cv2.Mat 对象的toBitmap() 好像会造成内存泄漏（即使转换后，手动释放 cv2.Mat对象和bitmap对象），我使用以下代码实现，暂时未发现内存大幅上涨的现象（转换后，手动释放 cv2.Mat对象和bitmap对象）

toBitmap = function(){
	if(this.empty() or this.depth() != 0/*_CV_8U*/) return null;
    var pixelFormats = {
    	[1] = 0x30803;
    	[3] = 0x21808;
    	[4] = 0x26200A;
    }
    
    var pixelFormat = pixelFormats[this.channels()];
    var stride = this.step1();
    if(stride % 4 ){
    	stride = stride - (stride % 4 ) + 4;
    }
	return ..gdip.bitmap( this.width,this.height,pixelFormat,this.data,stride );
}

opened by iycai 0

Releases(0.0.1)

0.0.1(Jun 22, 2021)

models store
Source code(tar.gz)
Source code(zip)
detect.caffemodel(942.80 KB)
detect.prototxt(41.65 KB)
face_mask_detection.caffemodel(3.86 MB)
face_mask_detection.prototxt(13.17 KB)
sr.caffemodel(23.36 KB)
sr.prototxt(5.84 KB)
yolo-fastest-xl.cfg(9.47 KB)
yolo-fastest-xl.weights(3.48 MB)

Owner

GitHub Repository

Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Vedaldi, Andrew Zisserman, CVPR 2016.

SynthText Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Ved

1.8k Dec 28, 2022

aardio的opencv库

Related tags

Overview

opencv_aardio

You might also like...

Comments

下载后提未错误putTextZH为NULL

aardio：运行时错误

错误行号:#4 文件:[string "import cv2..."]: 不支持此操作:call 定义类型:method(table) 名字:'putTextZH' 类型:null

调用栈: [string "import cv2..."]:4: in main chunk

确定

提示：Cannot load library

已下载DLL放置在\lib\cv2.res，运行时提示：

aardio：运行时错误

文件:...esktop\opencv_aardio-main\lib\cv2_.aardio 错误行号:#2 错误:Cannot load library '~/lib/cv2/.res/OpenCvSharpExtern.dll'.

调用栈: [kernel]: in function 'import' [string "import cv2..."]:1: in main chunk

确定

CV2默认库里的bitwise_or，bitwise_xor函数是否有个小BUG？

反馈几个问题

Releases(0.0.1)

0.0.1(Jun 22, 2021)

Owner

Code for generating synthetic text images as described in "Synthetic Data for Text Localisation in Natural Images", Ankush Gupta, Andrea Vedaldi, Andrew Zisserman, CVPR 2016.

Python tool that takes the OCR.space JSON output as input and draws a text overlay on top of the image.

A simple OCR API server, seriously easy to be deployed by Docker, on Heroku as well

OCR powered screen-capture tool to capture information instead of images

A semi-automatic open-source tool for Layout Analysis and Region EXtraction on early printed books.

天池2021"全球人工智能技术创新大赛"【赛道一】：医学影像报告异常检测 - 第三名解决方案

Pixel art search engine for opengameart

Textboxes_plusplus implementation with Tensorflow (python)

Python Computer Vision Aim Bot for Roblox's Phantom Forces

An expandable and scalable OCR pipeline

Demo processor to illustrate OCR-D Python API

Convolutional Recurrent Neural Networks(CRNN) for Scene Text Recognition

📷 This repository is focused on having various feature implementation of OpenCV in Python.

Roboflow makes managing, preprocessing, augmenting, and versioning datasets for computer vision seamless.

Thresholding-and-masking-using-OpenCV - Image Thresholding is used for image segmentation

Code for the head detector (HeadHunter) proposed in our CVPR 2021 paper Tracking Pedestrian Heads in Dense Crowd.

A tensorflow implementation of EAST text detector

Source code of RRPN ---- Arbitrary-Oriented Scene Text Detection via Rotation Proposals

Implement 'Single Shot Text Detector with Regional Attention, ICCV 2017 Spotlight'

A pure pytorch implemented ocr project including text detection and recognition