当我尝试使用用于RPi的工具链构建应用程序时遇到问题。工具链随git仓库中的教程一起安装。还有OpenCV 3.4.1。我将把文件cpp和cmakelists文件中的代码放在这里。
myClient.cpp：
entityManagerFactoryCMakeLists.txt：
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
optimizer = tf.contrib.opt.MultitaskOptimizerWrapper(params['optimization_algorithm'](params['training_rate']))
train_op = optimizer.minimize(loss)
global_step = tf.train.get_global_step()
update_global_step = tf.assign(global_step, global_step + 1, name = 'update_global_step')
return tf.estimator.EstimatorSpec(mode, loss = loss, train_op = tf.group(train_op, update_global_step))
和问题的文字在这里：
#include <sys/socket.h>
#include "opencv2/opencv.hpp"
#include <arpa/inet.h>
#include <net/if.h>
#include <unistd.h>
#include <string.h>
#include <cstdlib>
#include <pthread.h>
void run();
void * capture(void *);
int main()
{
run();
}
void run()
{
int sockSystemCall, acceptSystemCall,
bindSystemCall, portNumber;
char buffer[256];
pthread_t thread_id;
struct sockaddr_in serverAddress, clientAddress;
int clientAddSize;
portNumber = 3305;
sockSystemCall = socket(AF_INET, SOCK_DGRAM, 0);
if(sockSystemCall < 0)
exit(0);
// sin_family contains code for address family
serverAddress.sin_family = AF_INET;
// sin_addr.s_addr contains ip address
serverAddress.sin_addr.s_addr = INADDR_ANY;
// sin_port contains port number
serverAddress.sin_port = htons(portNumber);
// bind the server and check if it runs
bindSystemCall = bind(sockSystemCall, (struct sockaddr *) &serverAddress,
sizeof(serverAddress) );
if(bindSystemCall < 0)
exit(0);
// listen for connections
listen(sockSystemCall, 5);
// accept first incoming and exit if there is an issue
clientAddSize = sizeof(clientAddress);
acceptSystemCall = accept(sockSystemCall, (struct sockaddr *) &clientAddress,
(socklen_t *)&clientAddSize);
if(acceptSystemCall < 0)
exit(0);
pthread_create( &thread_id, nullptr, capture, &acceptSystemCall);
}
void * capture(void * pointer)
{
int serverFor = *(int *)pointer;
int rows, columns;
int numberOfDevice = 0;
cv::VideoCapture captureDevice(numberOfDevice);
// now it will gain data from camera and send it to server
cv::Mat image, grayImage;
rows = 640;
columns = 480;
// CV_8UC1 because i will use utf-8
image = cv::Mat::zeros(rows, columns, CV_8UC1);
if( !image.isContinuous() )
image = image.clone();
size_t imgSize = image.total() * image.elemSize();
ssize_t bytes;
while( true )
{
captureDevice >> image;
cvtColor(image, grayImage, CV_BGR2GRAY);
bytes = send(serverFor, grayImage.data, imgSize, 0);
if( bytes < 0 )
break;
}
}
对于这个项目，我正在使用CLion。现在真的不怎么处理它。一切对我来说都是新的，所以请忽略错误的代码。我将不胜感激。