二维平面上的矩阵运算
ps:这种线段树是不用lazy的, 因为不用push_down, 为啥不用push_down, 因为没有查询操作
同线段树 O l o g ( N ) Olog(N) Olog(N)
HDU 1542 [POJ 1151] Atlantis (矩形面积并)
HDU 1255 覆盖的面积 (矩形面积交)
HDU 1828 [POJ 1177] Picture(矩形周长并)
//矩阵面积并
/*
求N<=100个矩形的面积并
*/
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
using namespace std;
const int N = 205, INF = 0x3f3f3f3f, MOD = 1e9 + 7;
int n;
struct Seg {
double l, r, h; int d;
Seg() {}
Seg(double l, double r, double h, int d): l(l), r(r), h(h), d(d) {}
bool operator< (const Seg& rhs) const {return h < rhs.h;}
} a[N];
int cnt[N << 2]; //根节点维护的是[l, r+1]的区间
double sum[N << 2], all[N];//all 是离散化数组,sum是线段树维护的数组
#define lson l, m, rt << 1
#define rson m + 1, r, rt << 1 | 1
void push_up(int l, int r, int rt) { //因为是面积交
if(cnt[rt]) sum[rt] = all[r + 1] - all[l];
else if(l == r) sum[rt] = 0; //leaves have no sons
else sum[rt] = sum[rt << 1] + sum[rt << 1 | 1];
}
void update(int L, int R, int v, int l, int r, int rt) {
if(L <= l && r <= R) {
cnt[rt] += v;
push_up(l, r, rt);
return;
}
int m = l + r >> 1;
if(L <= m) update(L, R, v, lson);
if(R > m) update(L, R, v, rson);
push_up(l, r, rt);
}
int main() {
ios_base::sync_with_stdio(0);//这里存的矩阵是从下往上扫的。
int kase = 0;
while(scanf("%d", &n) == 1 && n) {
for(int i = 1; i <= n; ++i) {
double x1, y1, x2, y2;
scanf("%lf%lf%lf%lf", &x1, &y1, &x2, &y2);
a[i] = Seg(x1, x2, y1, 1);
a[i + n] = Seg(x1, x2, y2, -1);
all[i] = x1; all[i + n] = x2;
}
n <<= 1;
sort(a + 1, a + 1 + n);
sort(all + 1, all + 1 + n);
int m = unique(all + 1, all + 1 + n) - all - 1;
memset(cnt, 0, sizeof cnt);
memset(sum, 0, sizeof sum);
double ans = 0;
for(int i = 1; i < n; ++i) {
int l = lower_bound(all + 1, all + 1 + m, a[i].l) - all; //得到离散结果
int r = lower_bound(all + 1, all + 1 + m, a[i].r) - all;
if(l < r) update(l, r - 1, a[i].d, 1, m, 1);
ans += sum[1] * (a[i + 1].h - a[i].h);
}
printf("Test case #%d\nTotal explored area: %.2f\n\n", ++kase, ans);
}
return 0;
}
//HDU 1255 覆盖的面积 (矩形面积交)
/*
求N<=1000个矩形覆盖至少两次区域的面积,也就是矩形面积交
前面的与矩形面积并类似, 不同的是push_up的时候要考虑至少覆盖一次one和至少覆盖两次two的更新
尤其是当前被覆盖了一次的时候, 由于没有push_down操作, 父亲节点的信息是没有同步到儿子节点的, 这样的话push_up就要考虑了.
父亲被记录覆盖了一次, 但是如果儿子被覆盖过, 这些操作都是在这个父亲这个大区间上的, 就相当于父亲区间被覆盖了至少两次, 所以two和one都要更新
*/
#include
using namespace std;
#define pr(x) cout << #x << " = " << x << " "
#define prln(x) cout << #x << " = " << x << endl
const int N = 2e3 + 10, INF = 0x3f3f3f3f, MOD = 1e9 + 7;
int n;
struct Seg {
double l, r, h; int d;
Seg() {}
Seg(double l, double r, double h, double d): l(l), r(r), h(h), d(d) {}
bool operator< (const Seg& rhs) const {
return h < rhs.h;
}
} a[N];
int cnt[N << 2];
double one[N << 2], two[N << 2], all[N]; //one 为只相交一次的,two为相交两次的
#define lson l, m, rt << 1
#define rson m + 1, r, rt << 1 | 1
void push_up(int l, int r, int rt) {
if(cnt[rt] >= 2) two[rt] = one[rt] = all[r + 1] - all[l];
else if(cnt[rt] == 1) {
one[rt] = all[r + 1] - all[l];
if(l == r) two[rt] = 0;
else two[rt] = one[rt << 1] + one[rt << 1 | 1];
} else {
if(l == r) one[rt] = two[rt] = 0;
else {
one[rt] = one[rt << 1] + one[rt << 1 | 1];
two[rt] = two[rt << 1] + two[rt << 1 | 1];
}
}
}
void update(int L, int R, int v, int l, int r, int rt) {
if(L <= l && r <= R) {
cnt[rt] += v;
push_up(l, r, rt);
return;
}
int m = l + r >> 1;
if(L <= m) update(L, R, v, lson);
if(R > m) update(L, R, v, rson);
push_up(l, r, rt);
}
int main() {
ios_base::sync_with_stdio(0);
int t; scanf("%d", &t);
while(t--) {
scanf("%d", &n);
for(int i = 1; i <= n; ++i) {
double x1, y1, x2, y2;
scanf("%lf%lf%lf%lf", &x1, &y1, &x2, &y2);
a[i] = Seg(x1, x2, y1, 1);
a[i + n] = Seg(x1, x2, y2, -1);
all[i] = x1; all[i + n] = x2;
}
n <<= 1;
sort(a + 1, a + 1 + n);
sort(all + 1, all + 1 + n);
int m = unique(all + 1, all + 1 + n) - all - 1;
memset(cnt, 0, sizeof cnt);
memset(one, 0, sizeof one);
memset(two, 0, sizeof two);
double ans = 0;
for(int i = 1; i < n; ++i) {
int l = lower_bound(all + 1, all + 1 + m, a[i].l) - all;
int r = lower_bound(all + 1, all + 1 + m, a[i].r) - all;
if(l < r) update(l, r - 1, a[i].d, 1, m, 1);
ans += two[1] * (a[i + 1].h - a[i].h);
}
printf("%.2f\n", ans);
}
return 0;
}
//HDU 1828 [POJ 1177] Picture(矩形周长并)
/*
求N<=5000个矩形的轮廓长度,也就是矩形周长并
可以用类似矩形面积并的办法, 不过这次我们不乘高, 不算面积罢了.
需要注意的是, 由于周长的线会被重复覆盖, 我们每次需要和上一次的作差.
但是这样仅仅是x轴的, 不过我可以再y轴做一次加起来就可以了
*/
#include
using namespace std;
#define pr(x) cout << #x << " = " << x << " "
#define prln(x) cout << #x << " = " << x << endl
const int N = 1e4 + 10, INF = 0x3f3f3f3f, MOD = 1e9 + 7;
int n, m[2];
int sum[N << 2], cnt[N << 2], all[2][N];
struct Seg {
int l, r, h, d;
Seg() {}
Seg(int l, int r, int h, int d): l(l), r(r), h(h), d(d) {}
bool operator< (const Seg& rhs) const {return h < rhs.h;}
} a[2][N];
#define lson l, m, rt << 1
#define rson m + 1, r, rt << 1 | 1
void push_up(int p, int l, int r, int rt) {
if(cnt[rt]) sum[rt] = all[p][r + 1] - all[p][l];
else if(l == r) sum[rt] = 0;
else sum[rt] = sum[rt << 1] + sum[rt << 1 | 1];
}
void update(int p, int L, int R, int v, int l, int r, int rt) {
if(L <= l && r <= R) {
cnt[rt] += v;
push_up(p, l, r, rt);
return;
}
int m = l + r >> 1;
if(L <= m) update(p, L, R, v, lson);
if(R > m) update(p, L, R, v, rson);
push_up(p, l, r, rt);
}
int main() {
ios_base::sync_with_stdio(0);
while(scanf("%d", &n) == 1) {
for(int i = 1; i <= n; ++i) {
int x1, y1, x2, y2;
scanf("%d%d%d%d", &x1, &y1, &x2, &y2);
all[0][i] = x1, all[0][i + n] = x2;
all[1][i] = y1, all[1][i + n] = y2;
a[0][i] = Seg(x1, x2, y1, 1);
a[0][i + n] = Seg(x1, x2, y2, -1);
a[1][i] = Seg(y1, y2, x1, 1);
a[1][i + n] = Seg(y1, y2, x2, -1);
}
n <<= 1;
sort(all[0] + 1, all[0] + 1 + n);
m[0] = unique(all[0] + 1, all[0] + 1 + n) - all[0] - 1;
sort(all[1] + 1, all[1] + 1 + n);
m[1] = unique(all[1] + 1, all[1] + 1 + n) - all[1] - 1;
sort(a[0] + 1, a[0] + 1 + n);
sort(a[1] + 1, a[1] + 1 + n);
// for(int i = 0; i < 2; ++i){
// for(int j = 1; j <= m[i]; ++j) cout << all[i][j] <<' '; cout << endl;
// } cout << endl;
int ans = 0;
for(int i = 0; i < 2; ++i) { //上下做一遍,左右做一遍
int t = 0, last = 0;
memset(cnt, 0, sizeof cnt);
memset(sum, 0, sizeof sum);
for(int j = 1; j <= n; ++j) {
int l = lower_bound(all[i] + 1, all[i] + 1 + m[i], a[i][j].l) - all[i];
int r = lower_bound(all[i] + 1, all[i] + 1 + m[i], a[i][j].r) - all[i];
if(l < r) update(i, l, r - 1, a[i][j].d, 1, m[i], 1);
t += abs(sum[1] - last);
last = sum[1];
}
ans += t;
}
printf("%d\n", ans);
}
return 0;
}