opencv在3.0之后就支持调用深度学习模型。OpenCV dnn模块目前支持Caffe、TensorFlow、Torch、PyTorch等深度学习框架。另外，新版本中使用预训练深度学习模型的API同时兼容C++和Python。

参照官方教程进行一个分类模型的调用caffe模型进行分类，Load Caffe framework models。注意版本，我的版本是3.4.1，选择对应版本。

image.png

在opencv的解压文件下面有sample/dnn/所有的例子。这个例子来自sample/dnn//caffe_googlenet.cpp
下载两个文件bvlc_googlenet.prototxt and bvlc_googlenet.caffemodel

把上面三个文件放在与caffe_googlenet.cpp同级目录。
新建工程把caffe_googlenet.cpp添加进去。

caffe_googlenet.cpp 代码

#include <opencv2/dnn.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>
#include <opencv2/core/utils/trace.hpp>
using namespace cv;
using namespace cv::dnn;

#include <fstream>
#include <iostream>
#include <cstdlib>
using namespace std;

/* Find best class for the blob (i. e. class with maximal probability) */
static void getMaxClass(const Mat &probBlob, int *classId, double *classProb)
{
    Mat probMat = probBlob.reshape(1, 1); //reshape the blob to 1x1000 matrix
    Point classNumber;

    minMaxLoc(probMat, NULL, classProb, NULL, &classNumber);
    *classId = classNumber.x;
}

static std::vector<String> readClassNames(const char *filename)
{
    std::vector<String> classNames;

    std::ifstream fp(filename);
    if (!fp.is_open())
    {
        std::cerr << "File with classes labels not found: " << filename << std::endl;
        exit(-1);
    }

    std::string name;
    while (!fp.eof())
    {
        std::getline(fp, name);
        if (name.length())
            classNames.push_back(name.substr(name.find(' ') + 1));
    }

    fp.close();
    return classNames;
}

const char* params
= "{ help         | false | Sample app for loading googlenet model }"
"{ proto          | bvlc_googlenet.prototxt | model configuration }"
"{ model          | bvlc_googlenet.caffemodel | model weights }"
"{ label          | classification_classes_ILSVRC2012.txt | names of ILSVRC2012 classes }"
"{ image          | space_shuttle.jpg | path to image file }"
"{ opencl         | false | enable OpenCL }"
;

int main(int argc, char **argv)
{
    CV_TRACE_FUNCTION();

    CommandLineParser parser(argc, argv, params);

    if (parser.get<bool>("help"))
    {
        parser.printMessage();
        return 0;
    }

    String modelTxt = parser.get<string>("proto");
    String modelBin = parser.get<string>("model");
    String imageFile = parser.get<String>("image");
    String classNameFile = parser.get<String>("label");

    Net net;
    try {
        //! [Read and initialize network]
        net = dnn::readNetFromCaffe(modelTxt, modelBin);
        //! [Read and initialize network]
    }
    catch (const cv::Exception& e) {
        std::cerr << "Exception: " << e.what() << std::endl;
        //! [Check that network was read successfully]
        if (net.empty())
        {
            std::cerr << "Can't load network by using the following files: " << std::endl;
            std::cerr << "prototxt:   " << modelTxt << std::endl;
            std::cerr << "caffemodel: " << modelBin << std::endl;
            std::cerr << "bvlc_googlenet.caffemodel can be downloaded here:" << std::endl;
            std::cerr << "http://dl.caffe.berkeleyvision.org/bvlc_googlenet.caffemodel" << std::endl;
            exit(-1);
        }
        //! [Check that network was read successfully]
    }

    if (parser.get<bool>("opencl"))
    {
        net.setPreferableTarget(DNN_TARGET_OPENCL);
    }

    //! [Prepare blob]
    Mat img = imread(imageFile);
    if (img.empty())
    {
        std::cerr << "Can't read image from the file: " << imageFile << std::endl;
        exit(-1);
    }

    //GoogLeNet accepts only 224x224 BGR-images
    Mat inputBlob = blobFromImage(img, 1.0f, Size(224, 224),
        Scalar(104, 117, 123), false);   //Convert Mat to batch of images
    //! [Prepare blob]
    net.setInput(inputBlob, "data");        //set the network input
    Mat prob = net.forward("prob");         //compute output

    cv::TickMeter t;
    for (int i = 0; i < 10; i++)
    {
        CV_TRACE_REGION("forward");
        //! [Set input blob]
        net.setInput(inputBlob, "data");        //set the network input
        //! [Set input blob]
        t.start();
        //! [Make forward pass]
        prob = net.forward("prob");                          //compute output
        //! [Make forward pass]
        t.stop();
    }

    //! [Gather output]
    int classId;
    double classProb;
    getMaxClass(prob, &classId, &classProb);//find the best class
    //! [Gather output]

    //! [Print results]
    std::vector<String> classNames = readClassNames(classNameFile.c_str());
    std::cout << "Best class: #" << classId << " '" << classNames.at(classId) << "'" << std::endl;
    std::cout << "Probability: " << classProb * 100 << "%" << std::endl;
    //! [Print results]
    std::cout << "Time: " << (double)t.getTimeMilli() / t.getCounter() << " ms (average from " << t.getCounter() << " iterations)" << std::endl;
    // input samething
    int a;
    std::cin >> a;

    return 0;
} //main

按照对应的版本使用例子，一般不会出错。

这里有几个比较重要的参数，需改对应的参数就可进行不同图片的分类了。

const char* params
= "{ help         | false | Sample app for loading googlenet model }"
"{ proto          | bvlc_googlenet.prototxt | model configuration }"
"{ model          | bvlc_googlenet.caffemodel | model weights }"
"{ label          | synset_words.txt | names of ILSVRC2012 classes }"
"{ image          | space_shuttle.jpg | path to image file }"
"{ opencl         | false | enable OpenCL }"

proto: caffe的配置文件
model：caffe的参数文件
image：图片
opencl：是否使用opencl进行加速

result

Attempting to upgrade input file specified using deprecated V1LayerParameter: bvlc_googlenet.caffemodel
Successfully upgraded file specified using deprecated V1LayerParameter
[ INFO:0] Initialize OpenCL runtime...

Net Outputs(1):
prob
Best class: #812 'shuttle'
Probability: 99.993%
Time: 1492.05 ms (average from 10 iterations)

使用c++ opencv调用tensorflow训练好的卷积神经网络
 opencv官方教程caffe model调用
 OpenCV调用TensorFlow预训练模型
 基于opencv dnn模块的caffe模型的调用

C++ opencv-3.4.1 调用caffe训练好的模型

C++ opencv-3.4.1 调用caffe训练好的模型

按照对应的版本使用例子，一般不会出错。

result