OpenVINO计算机视觉模型加速-002

#include <opencv2/opencv.hpp>
#include <opencv2/dnn.hpp>
#include <iostream>

using namespace cv;
using namespace cv::dnn;
using namespace std;

int main(int argc, char** argv) {
    string xml = "D:/projects/models/vehicle-license-plate-detection-barrier-0106/FP32/vehicle-license-plate-detection-barrier-0106.xml";
    string bin = "D:/projects/models/vehicle-license-plate-detection-barrier-0106/FP32/vehicle-license-plate-detection-barrier-0106.bin";
    Mat src = imread("D:/images/vehicle_01.jpg");
    if (src.empty()) {
        printf("could not load image...\n");
        return -1;
    }
    namedWindow("input", WINDOW_NORMAL);
    namedWindow("ssd-model optimizer demo", WINDOW_NORMAL);
    imshow("input", src);

    // 加载模型
    Net net = readNetFromModelOptimizer(xml, bin);
    net.setPreferableBackend(DNN_BACKEND_INFERENCE_ENGINE);
    net.setPreferableTarget(DNN_TARGET_CPU);

    // 设置输入
    printf("ssd network model loaded...\n");
    Mat blob = blobFromImage(src, 1.0, Size(300, 300), Scalar(), true, false, 5);
    net.setInput(blob);

    // 推断预测
    float threshold = 0.5;
    Mat detection = net.forward();

    // 获取推断时间
    vector<double> layerTimings;
    double freq = getTickFrequency() / 1000;
    double time = net.getPerfProfile(layerTimings) / freq;
    ostringstream ss;
    ss << "infernece : " << time << " ms";
    putText(src, ss.str(), Point(50, 50), FONT_HERSHEY_SIMPLEX, 1, Scalar(255, 0, 0), 2, 8);

    // 解析输出结果
    Mat detectionMat(detection.size[2], detection.size[3], CV_32F, detection.ptr<float>());
    for (int i = 0; i < detectionMat.rows; i++) {
        float confidence = detectionMat.at<float>(i, 2);
        if (confidence > threshold) {
            size_t obj_index = (size_t)detectionMat.at<float>(i, 1);
            float tl_x = detectionMat.at<float>(i, 3) * src.cols;
            float tl_y = detectionMat.at<float>(i, 4) * src.rows;
            float br_x = detectionMat.at<float>(i, 5) * src.cols;
            float br_y = detectionMat.at<float>(i, 6) * src.rows;
            Rect object_box((int)tl_x, (int)tl_y, (int)(br_x - tl_x), (int(br_y - tl_y)));
            rectangle(src, object_box, Scalar(0, 0, 255), 2, 8, 0);
        }
    }
    imshow("ssd-model optimizer demo", src);
    waitKey(0);
    destroyAllWindows();
    return 0;
}