Trapezoidal Decomposition梯形分解算法(TCD)

news2024/9/22 5:05:03

系列文章目录

提示:这里可以添加系列文章的所有文章的目录,目录需要自己手动添加
TODO:写完再整理

文章目录

  • 系列文章目录
  • 前言
  • Trapezoidal Decomposition梯形分解算法(TCD)
    • 原理
      • (1)第一种原理
      • (2)第二种原理
      • (3)第三种原理
    • 优缺点


前言

认知有限,望大家多多包涵,有什么问题也希望能够与大家多交流,共同成长!

本文先对**Trapezoidal Decomposition梯形分解算法(TCD)**做个简单的介绍,具体内容后续再更,其他模块可以参考去我其他文章


提示:以下是本篇文章正文内容

Trapezoidal Decomposition梯形分解算法(TCD)

原理

Trapezoidal Decomposition是最简单的精确细胞分解方法。该方法先使机器人沿着空间的边缘行走一圈,构建整个区域地图,然后用一条竖直的切割线从左至右扫描过整个区域,当切割线与多边形障碍物的顶点相遇时,就会分解出一个子区域,这样整个自由空间就被分解为诸多子区域,每个子区域都是梯形。梯形分解法的时间复杂度通常为O(n log n)。机器人在每个子区域中进行往复运动对子区域进行覆盖。

(1)第一种原理

在这里插入图片描述

在这里插入图片描述

在这里插入图片描述

(2)第二种原理

梯形分解方法假设一条垂直线段(称为slice)从左到右扫过由多边形障碍物填充的有界环境。单元格cell是通过一系列打开和关闭操作形成的

事件类型
当slice遇到一个事件时,有三种类型的事件:IN、OUT和MIDDLE。
1、在IN事件中,当前单元格关闭(从而完成其构造),两个新单元格打开(从而启动其构造)(图2)。

2、OUT事件正好相反:两个单元格关闭,一个新单元格打开(图3)。
IN事件可以看作是一个单元格分解成两个单元格,而OUT事件是两个单元格合并成一个单元格。

3、在MIDDLE事件中,当前单元格关闭,一个新单元格形成。这些操作的结果是自由空间被分解成梯形单元格。
在这里插入图片描述
在这里插入图片描述

(3)第三种原理

在这里插入图片描述
在这里插入图片描述
在这里插入图片描述
通过cell的面积最大化来确定cell的真正边界
在这里插入图片描述

代码实现

# Find a path avoiding obstacles using Vertical Cell Decomposition
# Author -- Shikhar Dev Gupta

import sys
from helpers.graph import *
from helpers.geometry import *;
import matplotlib.pyplot as plt

# Check for empty lines
file_handler = open("input_file","r");
raw_data = file_handler.read();
raw_data = raw_data.split("\n");

if(len(raw_data) <2):
	print("Incorrect format of the input file");
	exit;

def parse_input_line(line):
	temp2 = [];
	line = [i.strip() for i in line.split(",")];
	vertex = [];
	for index,i in enumerate(line):
		if(i[0] == "("):
			i = i[1:];
		if(i[len(i)-1] == ")"):
			i= i[:-1];
		vertex.append(int(i));
		if(index%2 != 0):
			temp2.append(vertex);
			vertex = [];
	return temp2;	

# Draw the obstacles and point the source and the destination----------------------------------------------
def draw_problem():
	bnd_x = [i.x for i in boundary];
	bnd_x.append(boundary[0].x);
	bnd_y = [i.y for i in boundary];
	bnd_y.append(boundary[0].y);
	poly_x = [];
	poly_y = []

	# Draw the boundary
	plt.plot(bnd_x, bnd_y);
	
	for index, i in enumerate(obstacles):
		poly_x.append([p[0] for p in i]);
		poly_y.append([p[1] for p in i]);
	
		plt.fill( poly_x[index], poly_y[index], color="#512DA8");
	
	plt.plot(source.x, source.y, marker="o");
	plt.plot(dest.x, dest.y, marker="o");
	plt.annotate('Source', xy=(source.x, source.y), xytext=(source.x+5, source.y-6) );
	plt.annotate('Destination', xy=(dest.x, dest.y), xytext=(dest.x-4, dest.y-10) );


# Extract vertices----------------------------------------------
temp = parse_input_line(raw_data[0]);
boundary = [point(i[0], i[1]) for i in temp];

# Extract source and dest
temp = parse_input_line(raw_data[len(raw_data)-1]);
source = point(temp[0][0], temp[0][1]);
dest = point(temp[1][0], temp[1][1]);

# Extract obstacles
obstacles = [];
for i in raw_data[1:len(raw_data)-1]:
	obstacles.append(parse_input_line(i) );

#sort by x-values
sorted_vertices = [];
for index,i in enumerate(obstacles):
	for j in i:
		j.append(index);
		sorted_vertices.append(j);
sorted_vertices.sort(key=lambda x: x[0]);

# Draw the problem
draw_problem();

new_sorted_vertices = [];

for i in sorted_vertices:
	temp = point(i[0], i[1], i[2]);
	new_sorted_vertices.append(temp);

new_obstacles = [];
for index, i in enumerate(obstacles):
	temp_obs = [];
	for j in i:
		temp = point(j[0], j[1], index);
		temp_obs.append(temp);
	new_obstacles.append(temp_obs);	


#-----------------------------------------------------------
# Find vertical lines
open_line_segments = [];

y_limit_lower = boundary[0].y;
y_limit_upper = boundary[2].y;

for pt in new_sorted_vertices:
	curr_line_segment = [ point(pt.x, y_limit_lower), point(pt.x, y_limit_upper) ]; 
	lower_obs_pt = curr_line_segment[0];
	upper_obs_pt = curr_line_segment[1];
	upper_gone = False;
	lower_gone = False;
	break_now = False;

	# Find intersection points with the vertical proposed lines. the intersection function returns false if segments are same, so no need to worry about same segment checking
	for index,obs in enumerate(new_obstacles):
		# Add the first point again for the last line segment of a polygon.
		
		obs.append( obs[0] );
		for vertex_index in range(len(obs)-1 ):
			res = segment_intersection( curr_line_segment[0], curr_line_segment[1], obs[vertex_index],  obs[vertex_index+1]);
			if (res!=-1):
				if ( index == pt.obstacle ):
					if pt.equals( res ) == False:
						if ( res.y > pt.y ):
							upper_gone = True;
						elif ( res.y < pt.y ):
							lower_gone = True;	
				else:	
					if pt.equals( res ) == False:
						if ( upper_gone is False ):
							if ( (res.y > pt.y) and res.y < (upper_obs_pt.y) ):
								upper_obs_pt = res;
						if ( lower_gone is False ):
							if ( (res.y < pt.y) and (res.y > lower_obs_pt.y) ):
								lower_obs_pt = res;
			if( upper_gone is True and lower_gone is True ):
				break_now = True;

		#No need to check for current point anymore...completely blocked
		if(break_now is True):
			break;		

	# Draw the vertical cell lines
	if(lower_gone is False):
		plt.plot( [lower_obs_pt.x, pt.x],  [lower_obs_pt.y, pt.y] );
		
	if(upper_gone is False):
		plt.plot( [pt.x, upper_obs_pt.x],  [pt.y, upper_obs_pt.y] );

	# Add to the global segment list
	if (lower_gone and upper_gone):
		open_line_segments.append([None, None]);
	elif (lower_gone):
		open_line_segments.append([None, upper_obs_pt]);
	elif (upper_gone):
		open_line_segments.append([lower_obs_pt, None]);
	else:
		open_line_segments.append([lower_obs_pt, upper_obs_pt]);


#------------------------------------------------------
# Find Polygon cells naiively. Will improve next. 
cells = [];

for index1 in range(len(open_line_segments) ):
	curr_segment = open_line_segments[index1];
	curr_vertex = new_sorted_vertices[index1];
	break_now = False;
	done = [False, False, True];
	if( curr_segment[0] is None ):
		done[0] = True; 
	if( curr_segment[1] is None ):
		done[1] = True;	
	if( curr_segment[1] is None and open_line_segments[index1][0] is None):
		done[2] = False;	

	for index2 in range(index1+1,  len(open_line_segments) ):
		next_segment = open_line_segments[index2];
		next_vertex = new_sorted_vertices[index2];			
		
		double_index1 = -2;
		double_index2 = -2;
		lines_to_check = [];
		trapezoids = [];
		double_check = False;

		if ( next_segment[0] is not None and next_segment[1] is not None ):
			double_check = True;

		if( done[0] is False ):
			if( double_check ):
				double_index1 = len(lines_to_check);
				lines_to_check.append( [centroid([curr_segment[0], curr_vertex]), centroid([next_segment[0], next_vertex]), 0]);
				lines_to_check.append( [centroid([curr_segment[0], curr_vertex]), centroid([next_segment[1], next_vertex]), 0]);
				trapezoids.append([ curr_segment[0], next_segment[0], next_vertex, curr_vertex ]);
				trapezoids.append([ curr_segment[0], next_vertex, next_segment[1], curr_vertex ]);
			elif ( next_segment[0] is not None ):
				lines_to_check.append( [centroid([curr_segment[0], curr_vertex]), centroid([next_segment[0], next_vertex]), 0]);
				trapezoids.append([ curr_segment[0], next_segment[0], next_vertex, curr_vertex ]);
			elif( next_segment[1] is not None ):
				lines_to_check.append( [centroid([curr_segment[0], curr_vertex]), centroid([next_segment[1], next_vertex]), 0]);
				trapezoids.append([ curr_segment[0], next_vertex, next_segment[1], curr_vertex ]);
			else:
				lines_to_check.append( [centroid([curr_segment[0], curr_vertex]), next_vertex, 0]);
				trapezoids.append([ curr_segment[0], next_vertex, curr_vertex ]);

		if( done[1] is False ):
			if( double_check ):
				double_index2 = len(lines_to_check);
				lines_to_check.append( [centroid([curr_segment[1], curr_vertex]), centroid([next_segment[0], next_vertex]), 1]);
				lines_to_check.append( [centroid([curr_segment[1], curr_vertex]), centroid([next_segment[1], next_vertex]), 1]);
				trapezoids.append([ curr_vertex, next_segment[0], next_vertex , point(curr_segment[1].x, curr_segment[1].y,curr_segment[1].obstacle, 34)]);
				trapezoids.append([ curr_vertex, next_vertex, next_segment[1], curr_segment[1] ]);
			elif ( next_segment[1] is not None ):
				lines_to_check.append( [centroid([curr_segment[1], curr_vertex]), centroid([next_segment[1], next_vertex]), 1]);
				trapezoids.append([ curr_vertex, next_vertex, next_segment[1], curr_segment[1] ]);
			elif( next_segment[0] is not None ):
				lines_to_check.append( [centroid([curr_segment[1], curr_vertex]), centroid([next_segment[0], next_vertex]), 1]);
				trapezoids.append([ curr_vertex, next_segment[0], next_vertex , curr_segment[1] ]);
			else:
				lines_to_check.append( [centroid([curr_segment[1], curr_vertex]), next_vertex, 1]);
				trapezoids.append([ curr_vertex, next_vertex, curr_segment[1] ]);
		
		if( done[2] is False ):
			if(double_check):
				double_index = len(lines_to_check);
				lines_to_check.append( [curr_vertex, centroid([next_segment[0], next_vertex]), 2]);
				trapezoids.append([ curr_vertex,next_segment[0], next_vertex ]);
				lines_to_check.append( [curr_vertex, centroid([next_segment[1], next_vertex]), 2]);
				trapezoids.append([ curr_vertex, next_vertex, next_segment[1] ]);
			elif ( next_segment[0] is not None ):
				lines_to_check.append( [curr_vertex, centroid([next_segment[0], next_vertex]), 2]);
				trapezoids.append([ curr_vertex,next_segment[0], next_vertex ]);
			elif( next_segment[1] is not None ):
				lines_to_check.append( [curr_vertex, centroid([next_segment[1], next_vertex]), 2]);
				trapezoids.append([ curr_vertex, next_vertex, next_segment[1] ]);
			# Will this ever occur though??
			else:
				lines_to_check.append( [curr_vertex, next_vertex, 2]);
				trapezoids.append([curr_vertex, next_vertex]);

		temp_to_remove = [];
		for index5,q in enumerate(lines_to_check): 
			ok = [True, True, True];
			for index3,obs in enumerate(new_obstacles):
				# Add the last line to make closed polygon
				obs.append( obs[0] );
				for index4 in range(len(obs)-1):
					if (segment_intersection( q[0], q[1],  obs[index4],  obs[index4+1]) != -1):
						ok[q[2]] = False;
						if(index5 not in temp_to_remove):
							temp_to_remove.append(index5);
						

			if (  ok[q[2]] is True ):
				done[q[2]] = True;

		for i in range(len(lines_to_check)):
			if i not in temp_to_remove:
				cells.append(trapezoids[i]);
		
		if( done[0] == True and done[1] == True and done[2] == True ):
			break;

to_draw =[];
for i in cells:
	i.append(i[0]);
	to_draw.append(i);


#-------------------------------------------------------
# Merge overlapping Polygons
quad_cells = [i for i in cells if len(i)>3];
tri_cells = [i for i in cells if len(i)==3];
others = [i for i in cells if len(i)<3];
quads_to_remove = [];
quads_to_add = [];

quads_to_remove = [];
quads_to_add = [];
for index_cell in range(len(quad_cells)):
	for index_cell2,cell in enumerate(quad_cells):
		if(index_cell != index_cell2):
			if(quad_cells[index_cell][0].x == cell[0].x and quad_cells[index_cell][1].x == cell[1].x):
					temp1 = list(quad_cells[index_cell]);
					temp1.append(temp1[0]);
					temp2 = list(cell);
					temp2.append(temp2[0]);
					area1 = polygon_area(temp1,4); area2 = polygon_area(temp2,4);
					new_quad=[];
					
					new_quad.append( point(temp1[0].x, min(temp1[0].y, temp2[0].y)) );
					new_quad.append( point(temp1[1].x, min(temp1[1].y, temp2[1].y)) );
					new_quad.append( point(temp1[1].x, max(temp1[2].y, temp2[2].y)) );
					new_quad.append( point(temp1[0].x, max(temp1[3].y, temp2[3].y)) );
					new_quad.append( point(temp1[0].x, min(temp1[0].y, temp2[0].y)) );
					area3 = polygon_area(new_quad, 4);
					if( area1 + area2 >= area3):
						#merge
						quads_to_remove.append(index_cell);
						quads_to_remove.append(index_cell2);
						
						quads_to_add.append(new_quad);

quads_to_remove = list(set(quads_to_remove));
for index in sorted(quads_to_remove, reverse=True):
    del quad_cells[index];

for i in quads_to_add:
	quad_cells.append(i);

# Remove duplicates
to_remove = [];
for index1 in range(len(quad_cells)):
	for index2 in range(index1+1, len(quad_cells)):
		duplicate = True;
		for k,m in zip(quad_cells[index1], quad_cells[index2]):
			if k.equals(m) is False:
				duplicate = False;
				break;
		if(duplicate is True):
			if index2 not in to_remove:
				to_remove.append(index2);		

for index in sorted(to_remove, reverse=True):
    del quad_cells[index];

# One more pass to remove extra quads generated because of cross - segments
quads_to_remove = [];
for index1 in range(len(quad_cells)):
	for index2 in range(len(quad_cells)):
		if(index1 != index2 and quad_cells[index1][0].x == quad_cells[index2][0].x and quad_cells[index1][1].x == quad_cells[index2][1].x):
			
			if( (quad_cells[index1][0].y<= quad_cells[index2][0].y) and  (quad_cells[index1][1].y<= quad_cells[index2][1].y)
				and (quad_cells[index1][2].y>= quad_cells[index2][2].y) and (quad_cells[index1][3].y >= quad_cells[index2][3].y)):			
				quads_to_remove.append(index2);


quads_to_remove = list(set(quads_to_remove) );
for index in sorted(quads_to_remove, reverse=True):
    del quad_cells[index];


#------------------------------------------------------
# Add boundary lines
if( boundary[0].x != new_sorted_vertices[0].x):
	quad_cells.append([boundary[0], point(new_sorted_vertices[0].x, y_limit_lower), point(new_sorted_vertices[0].x, y_limit_upper), boundary[3]]);
if( boundary[1].x != new_sorted_vertices[len(new_sorted_vertices)-1].x):
	quad_cells.append([point(new_sorted_vertices[len(new_sorted_vertices)-1].x ,y_limit_lower), boundary[1], boundary[2], point(new_sorted_vertices[len(new_sorted_vertices)-1].x, y_limit_upper) ]);

#-------------------------------------------------------
# Plot final cells
to_draw = quad_cells+tri_cells+others;
for i in to_draw:
	x = [j.x for j in i];
	y = [j.y for j in i];
	plt.plot(x, y);

#----------------------------------------------------------------------
# Get the graph
graph_vertices = [];
graph_edges = [];

for index1 in range(len(quad_cells)):
	same_boundary = [];
	for index2 in range(len(quad_cells)):
		if(index1 != index2):
			if( (quad_cells[index1][1].x == quad_cells[index2][0].x ) and ((quad_cells[index1][2].y in [quad_cells[index2][0].y, quad_cells[index2][3].y]) or (quad_cells[index1][1].y in [quad_cells[index2][0].y, quad_cells[index2][3].y]) ) ):
				same_boundary.append(index2);

	temp = quad_cells[index1][0:4];
	centroid_vertex = centroid(temp);
	place = centroid_vertex.find_point(graph_vertices)
	if( place == -1):
		graph_vertices.append(centroid_vertex);

	if(len(same_boundary)==1):
		temp_edge_middle = centroid([quad_cells[index1][1], quad_cells[index1][2]]);
		graph_vertices.append(temp_edge_middle);
		n = len(graph_vertices)-1;
		if(place != -1):
			graph_edges.append([place, n]);
		else:
			graph_edges.append([n-1, n]);
		temp = quad_cells[same_boundary[0]][0:4];
		curr_centroid_vertex = centroid(temp);
		place2 = curr_centroid_vertex.find_point(graph_vertices);
		if( place2 == -1 ):
			graph_vertices.append(curr_centroid_vertex);
			graph_edges.append([n, n+1]);
		else:
			graph_edges.append([n, place2]);

	elif(len(same_boundary)>1):
		n = len(graph_vertices)-1;
		if(place != -1):
			use = place;
		else:
			use = n;	
		for index, i in enumerate(same_boundary):
			temp = quad_cells[i][0:4];
			curr_centroid_vertex = centroid(temp);
			temp_edge_middle = centroid([quad_cells[i][0], quad_cells[i][3]]);
			graph_vertices.append(temp_edge_middle);
			pl1 =len(graph_vertices)-1;
			hmmm= curr_centroid_vertex.find_point(graph_vertices);
			if (hmmm == -1):
				graph_vertices.append(curr_centroid_vertex);
				pl2 =len(graph_vertices)-1;
			else:
				pl2 = hmmm;	
			graph_edges.append([use, pl1]);
			graph_edges.append([pl1, pl2]);		


# Add source and dest to graph
# Find the smallest distance vertex on graph and see if its clear to traverse
# Source------------------------------
min_ind = -1; min = 9999999;
for index, i in enumerate(graph_vertices):
	if( check_obstruction(new_obstacles, [source, i]) is True ):
		dist = find_dist(i, source);
		if( dist < min):
			min = dist;
			min_ind = index;

graph_vertices.append(source);
m = len(graph_vertices)-1;
graph_edges.append([min_ind, m]);	

# Destination------------------------------------
min_ind = -1; min = 9999999;
for index, i in enumerate(graph_vertices):
	if( check_obstruction(new_obstacles, [dest, i]) is True ):
		dist = find_dist(i, dest);
		if( dist < min):
			min = dist;
			min_ind = index;

graph_vertices.append(dest);
m = len(graph_vertices)-1;
graph_edges.append([min_ind, m]);

# Convert graph in adjacency list format
graph = [];
for j in range(len(graph_vertices)):
	graph.append([]);
	for i in graph_edges:
		if(i[0]==j):
			graph[j].append(i[1]);
		elif(i[1]==j):
			graph[j].append(i[0]);	

path = bfs(graph, len(graph_vertices)-2, len(graph_vertices)-1);

if(path is None):
	print "No path found. Sorry";
	sys.exit();
else:
	print "Path found."	;

# Draw everything--------------
for index,i in enumerate(graph_vertices):
	plt.annotate(str(index), xy=(i.x, i.y), xytext=(i.x+2, i.y-2) );
	# plt.plot(i.x,i.y, marker="x");

for i in graph_edges:
	temp_x = [graph_vertices[i[0]].x, graph_vertices[i[1]].x];
	temp_y = [graph_vertices[i[0]].y, graph_vertices[i[1]].y];
	plt.plot(temp_x,temp_y);


# draw path
temp_x = [graph_vertices[i].x for i in path];
temp_y = [graph_vertices[i].y for i in path];
plt.plot(temp_x,temp_y, color="#0F0F0F", linewidth=2);	


#----------------------------------------------------
# output into a file
file_output = open("vertical_cell_output", "w" );
str_to_write = "";
for index in range(len(graph_vertices)):

	str_to_write = str_to_write + ", "+str(index)+":"+"("+ str(int(graph_vertices[index].x) )+  ", "+ str(int(graph_vertices[index].y) ) + ")";

str_to_write = str_to_write[1:];

total_write = str_to_write+"\n";
str_to_write="";
for i in graph:
	if (i == []):
		continue;
	str_to_write = str_to_write + ",(";
	for j in i:
		str_to_write = str_to_write + str(j) + ",";
	str_to_write = str_to_write[:-1];
	str_to_write = str_to_write + ")";

str_to_write = str_to_write[1:];

total_write = total_write+ str_to_write + "\n";

str_to_write = "";
str_to_write =','.join(str(x) for x in path);

total_write = total_write + str_to_write;

file_output.write(total_write);
print "Output written to file.. Drawing the result";

plt.show();

优缺点

产生大量的子区域,造成了大量的路径冗余


本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处:http://www.coloradmin.cn/o/2154194.html

如若内容造成侵权/违法违规/事实不符,请联系多彩编程网进行投诉反馈,一经查实,立即删除!

相关文章

粒子向上持续瀑布动画效果(直接粘贴到记事本改html即可)

代码&#xff1a; 根据个人喜好修改即可 <!DOCTYPE html> <html lang"zh"> <head><meta charset"UTF-8"><meta name"viewport" content"widthdevice-width, initial-scale1.0"><title>宽粒子向上…

技术周总结 09.16~09.22 周日

文章目录 一、09.16 周一1.1&#xff09;问题01&#xff1a; 软件质量属性中"质量属性场景"、"质量属性环境分析"、"质量属性效用树"、"质量属性需求用例分析"分别是什么&#xff1f;1.2&#xff09;问题02&#xff1a; 软件质量属性中…

Sharding-Jdbc+Mybatis+SpringBoot的分库分表实现

Sharding-JdbcMybatisSpringBoot的分库分表实现 1、sharding jdbc简介 Sharding-JDBC定位为轻量级java框架&#xff0c;使用客户端直连数据库&#xff0c;以jar包形式提供服务&#xff0c;未使用中间层&#xff0c;无需额外 部署&#xff0c;无其他依赖&#xff0c;DBA也无需…

SQL - 进阶语法(一)

1. SELECT TOP SELECT TOP Number|Percentage column1 from ... 选择所需要的数据的前多少行&#xff0c;Number表示具体数值&#xff0c;Percentage表示百分比 2. LIKE 搜索列元素的指定模式&#xff0c;匹配对应的信息进行自定义搜索 通常需要搭配通配符进行使用 %替代…

FPGA与Matlab图像处理之直方图均衡化

文章目录 一、什么是直方图?二、什么是直方图均衡化&#xff1f;三、Matlab实现直方图均衡化的步骤第一步&#xff1a; 彩色图像转成灰度图像第二步&#xff1a;提取亮度通道的直方图第三步&#xff1a;累计亮度通道的像素值频率第四步&#xff1a; 映射到新的灰度值 四、Veri…

使用pe工具制作ubuntu备份系统和还原系统

使用pe工具制作ubuntu备份系统和还原系统 备份系统还原系统修复磁盘教程修复引导教程为什么使用pe工具 1,因为我个人觉得这个工具实现起来比systemback软件操作起来报错少些,而且装的快,其他系统同理 实验准备 1,一个电脑,一个pe启动U盘 备份系统 插入U盘,开机进入pe系…

数据结构|二叉搜索树

&#x1f36c; mooridy-CSDN博客 &#x1f36c;数据结构专栏&#xff08;更新中&#xff09; 目录 1. ⼆叉搜索树的概念 2. ⼆叉搜索树的性能分析 3.⼆叉搜索树key和key/value key搜索场景 key/value搜索场景 4. 二叉搜索树的代码实现 4.1 ⼆叉搜索树的插⼊ 4.2 ⼆叉搜索…

harbor私有镜像仓库,搭建及管理

私有镜像仓库 docker-distribution docker的镜像仓库&#xff0c;默认端口号5000 做个仓库&#xff0c;把镜像放里头&#xff0c;用什么服务&#xff0c;起什么容器 vmware公司在docker私有仓库的基础上做了一个web页面&#xff0c;是harbor docker可以把仓库的镜像下载到本地&…

docker入门总结(附错误处理,持续更新)

安装、启动、卸载 卸载掉旧版本的 Docker yum remove -y docker docker-client docker-client-latest docker-common docker-latest docker-latest-logrotate docker-logrotate docker-selinux docker-engine-selinux docker-engineDocker安装&#xff08;选其一&#xff09;…

CTC loss 博客转载

论文地址&#xff1a; https://www.cs.toronto.edu/~graves/icml_2006.pdf 为了对应这个图&#xff0c;我们假设一种符合的模型情况&#xff1a; 英文OCR&#xff0c;37个类别&#xff08;26个小写字母10个汉字空格&#xff09;&#xff0c;最大输出长度8个字符 模型预测结果…

C++:多态(协变,override,final,纯虚函数抽象类,原理)

目录 编译时多态 函数重载 模板 运行时多态 多态的实现 实现多态的条件 协变 析构函数的重写 override 关键字 final 关键字 重载、重写、隐藏对比 纯虚函数和抽象类 多态的原理 多态是什么&#xff1f; 多态就是有多种形态 多态有两种&#xff0c;分别是编译时…

【Linux-基础IO】C语言文件接口回顾 系统文件概念及接口

目录 一、C语言文件接口回顾 C语言基础知识 C中文件操作示例 二、系统文件概念及接口 重定向基本理解的回顾 文件的基本概念 系统调用接口 open read write close lseek 什么是当前路径 一、C语言文件接口回顾 引言&#xff1a;我们并不理解文件&#xff01;从语…

【JavaEE】——多线程(join阻塞,计算,引用,状态)

阿华代码&#xff0c;不是逆风&#xff0c;就是我疯&#xff0c;你们的点赞收藏是我前进最大的动力&#xff01;&#xff01;希望本文内容能够帮助到你&#xff01; 目录 一&#xff1a;join等待线程结束 1&#xff1a;知识回顾 2&#xff1a;join的功能就是“阻塞等待” …

MK-1000门控管理系统

MK-1000门控系统简介 1.1 前言 中大型仓库一般由多个仓库组成&#xff0c;每个仓库具有上百个仓库门&#xff0c;人工控制仓库门的开启、闭合等耗时耗力&#xff0c;随着电子信息自动化发展&#xff0c;集中控制仓库门成为现实。本系统可通过网络灵活控制某个仓库的某个门开启、…

共轭传热和浸没边界耦合相关的论文的阅读笔记

Diffuse-interface immersed-boundary framework for conjugate-heat-transfer problems https://doi.org/10.1103/PhysRevE.99.053304 三区 混合交错离散体积和非交错离散体积 看不懂&#xff0c;不会 FVM 计算固体体积分数的步骤 对于每一个点&#xff1a; 找固体表面上…

【多维动态规划】64. 最小路径和(面试真题+面试官调整后的题目)

64. 最小路径和 难度&#xff1a;中等 力扣地址&#xff1a;https://leetcode.cn/problems/minimum-path-sum/description/ 1. 原题以及解法 1.1 题目 给定一个包含非负整数的 m x n 网格 grid &#xff0c;请找出一条从左上角到右下角的路径&#xff0c;使得路径上的数字总和…

数据库提权【笔记总结】

文章目录 UDF提权以有webshell只有数据库权限条件复现msf工具sql语句提权 MOF提权前言条件复现msf工具php脚本提权 sqlserver提权前言条件xp_cmdshell提权复现 沙盒提权介绍复现 Oracle提权靶场搭建执行任意命令复现 通过注入存储过程提权&#xff08;低权限提升至DBA&#xff…

HTML-DOM模型

1.DOM模型 window对象下的document对象就是DOM模型。 DOM描绘了一个层次化的节点树&#xff0c;每一个节点就是一个html标签&#xff0c;而且每一个节点也是一个DOM对象。 2.操作DOM 2.1.获取DOM对象常用方法 获取DOM对象的常用方法有如下几种&#xff1a; getElementById(…

Linux入门学习:Git

文章目录 1. 创建仓库2. 仓库克隆3. 上传文件4. 相关问题4.1 git进程阻塞4.2 git log4.3 上传的三个步骤在做什么4.4 配置邮箱/用户名 本文介绍如何在Linux操作系统下简单使用git&#xff0c;对自己的代码进行云端保存。 1. 创建仓库 &#x1f539;这里演示gitee的仓库创建。…

Linux 基础IO 1

文件操作 基础IO 在程序中若需要读取文件数据或者将数据永久存入需要使用文件因为程序的产生的数据都是在内存中的&#xff0c;只要关闭程序就会释放掉不会&#xff0c;所以在一些程序中我们难免需要对文件进程操作。无论我们是存入数据还是提取数据前提是有这个数据并清楚它…