lock.cpp

// Grayson Pike, 2018

/*
	Facial Recognition Lock Software (Main program)
	Runs continuously looking for specific faces, then triggers a servo
	when an authenticated face is recognized.

	To prepare a model for this program, follow the steps in README.md
*/


#include <iostream>
#include <vector>
#include <string>
#include <stdlib.h>
#include <unistd.h>

#include <opencv2/core.hpp>
#include <opencv2/videoio.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/face.hpp>
#include <opencv2/imgcodecs.hpp>

#include "face_detect.hpp"
#include "hardware.hpp"
#include "config.hpp"


/*
    Returns model name from command line arguments or exits on failure  
*/
std::string get_model_name(int argc, char *argv[]) {

    std::string model_name;
    if (argc != 2) {
        std::cout << "Error: Correct usage: ./train <model_name>" << std::endl;
        exit(1);
    } else {
        return argv[1];
    }

}


int main(int argc, char *argv[]) {

	//GPIO input_pin("17", "in");
	//std::cout << input_pin.read_value() << std::endl;

	std::string model_name = get_model_name(argc, argv);

	// Load model
	std::cout << "Loading model..." << std::flush;
	cv::Ptr<cv::face::FaceRecognizer> model = cv::face::LBPHFaceRecognizer::create();
	model->read(std::string(MODEL_DIR) + model_name + ".xml");
	std::cout << "[DONE]" << std::endl;

	cv::VideoCapture camera = get_camera();

	// Loop to detect faces
	bool loop = true;
	while (loop) {

		cv::Mat image;

		// Capture an image
		flush_capture_buffer(camera);
		camera >> image;

		// Convert image to greyscale
		cv::cvtColor(image, image, cv::COLOR_RGB2GRAY);

		// Detect coordinates of a face, if any
		std::vector<cv::Rect> face_regions = detect_faces(image);
		std::cout << "Deteced " << face_regions.size() << " faces." << std::endl;

		// TODO: Make this work with more than one face in the picture
		if (face_regions.size() == 1) {

			// Crop image to face, then resize
			image = image(face_regions[0]);
			cv::resize(image, image, cv::Size(FACE_WIDTH, FACE_HEIGHT), 0, 0, cv::INTER_LANCZOS4);

			// Determine if face is allowed
			int label;
			double confidence;
			model->predict(image, label, confidence);

			std::cout << "Results: " << label << ", " << confidence << std::endl;

			// If a face is recognized and authorized
			if (label != 0 && confidence <= POSITIVE_THRESHOLD) {
				std::cout << "******   Detected allowed face with label: " << label << std::endl;
				// Move the lock to the unlocked position
				system("sudo python servo.py 250");
				// Sleep for 1 second (or 1000000 microseconds) to allow the servo to move
				usleep(1500000u);
				// Move back to the resting position
				system("sudo python servo.py 175");
				// Sleep for WAIT_TIME seconds before relocking the door
				usleep(1500000u * WAIT_TIME);
				// Move the lock to the locked position
				system("sudo python servo.py 100");
				// Sleep for 1 second (or 1000000 microseconds) to allow the servo to move
				usleep(1000000u);
				// Move back to the resting position
				system("sudo python servo.py 175");
			}

		}

		//std::string input;
		//std::cout << "Enter 'q' to quit, or any other key to take another picture: " << std::endl;
		//std::cin >> input;
		//if (input == "q" || input == "Q") {
		//	loop = false;
		//}

	}

}