C++ 中多维数组之间的快速转换(以1维2维数组的相互转换为例说明)
在图像处理中,大规格的数组之间的转换是家常便饭。
c++中多维数组在本质上就是一个地址问题,普通情况下,可以通过类似如下方法进行暴力转换
int arr[50];
int main()
{
unsigned i, j;
for (i = 0; i < 50; i++) {
arr[i] = 2 * i;
}
int bruteArray[7][7];
for (i = 0; i < 7; i++)
for (j = 0; j < 7; j++)
bruteArray[i][j] = arr[i * 7 + j];
cout << "[0][0]=" << bruteArray[0][0] << endl;
cout << "[1][0]=" << bruteArray[1][0] << endl;
return 0;
}
不过这样的话,开销比较大,效率就成了一个问题,尤其是在处理大规模矩阵式多维数组时,所以一般采取直接进行地址转换的方式,不必逐个从内存拷贝,就能得到正确的转换。
#include <iostream>
#include <stdio.h>
using namespace std;
int arr[50];
int main()
{
unsigned i, j;
for (i = 0; i < 50; i++) {
arr[i] = 2 * i;
}
//method1 -- convert 1d array to 2d
cout << "------------------------------------------------------" << endl;
int *pa[7];
for (j = 0; j < 7; j++) {
pa[j] = &arr[j*7];
for (i = 0; i < 7; i++) {
cout << pa[j][i] << endl;
}
}
//method2 --
cout << "------------------------------------------------------" << endl;
int array1[] = { 1,2,3,4,5,6,7,8,9,10 };
int array2[] = { 9,8,7,6,5,4,3,2,1,0 };
int *arr2d[2];
arr2d[0] = array1;
arr2d[1] = array2;
cout << "[0][0]=" << arr2d[0][0] << endl;
cout << "[1][0]=" << arr2d[1][0] << endl;
//method3 --
cout << "------------------------------------------------------" << endl;
int *arr2dv[] = { array1, array2 };
cout << "[0][0]=" << arr2dv[0][0] << endl;
cout << "[1][0]=" << arr2dv[1][0] << endl;
//method test0 -- convert 2d array to 1d
cout << "------------------------------------------------------" << endl;
int test2d[2][10] = { { 1,2,3,4,5,6,7,8,9,10 },{ 9,8,7,6,5,4,3,2,1,0 } };
int *test1 = test2d[0];
int *test2 = test2d[1];
cout << "[0][0]=" << test1[0] << endl;
cout << "[1][0]=" << test2[1] << endl;
return 0;
}
对于动态创建数组的情况,
int step = n; // columns
int rows = m;
int tot = m*n; // assume m rows, n columns
double* data = new double[tot];
double ** matrix = new double* [m];
for (int i = 0; i < m; i++) {
matrix[i] = &data[(int)(step * i)];
}
// do something here for matrix manipulation
// 这里干点什么。。。。例如,
matrix[i][j] = 1234;
//....
delete[] matrix;
delete[] data;
对于更高维数的数组,原理也是一样的。
下面演示一个三维数组的情况,该数组可以直接采用pmem[page][y][x]的形式快速访问。
const int COLOR_NUM = 10;
unsigned char* maskMem = (unsigned char*)malloc(totSize);
memset(maskMem, 0, totSize);
unsigned char *** pmem = new unsigned char** [COLOR_NUM];
int x = 0, y = 0, page = 0;
for (int clr = 0; clr < COLOR_NUM; clr++) {
pmem[clr] = new unsigned char* [ht]; // similar layer border marker (-10 for ext, 10 for inner) matrix
page = clr * mskSize;
for (y = 0; y < ht; y++) {
x = page + y * wd;
pmem[clr][y] = &maskMem[x];
}
}
另外,网上有人用memcpy的方式来实现内存的直接拷贝,这种方法并不能一直正确,因为多维数组取决于你创建的方式,内存未必是连续的。