基线网平差（条件平差，间接平差，附有参数的条件平差，附有限制条件的间接平差）python3.7 自动生成矩阵并迭代平差
总述
先挖个坑 持续更新，目前已经实现了前三个
GayHub 有正常运行的案例
案例详情
一个由24条基线组成的GNSS网，共12个点，4个已知，data.csv如下
1,G01,G02,02912608_G01.zsd-02942607_G02.zsd,-366.6087,-252.4815,72.1291,8.1,11,13.7
2,G01,g03,02912608_G01.zsd-03662608_g03.zsd,-509.2608,22.3781,-426.4979,7.8,10.6,13.3
3,G01,g11,02942630_G01.zsd-03662630_g11.zsd,714.8437,446.3863,-70.7624,4.1,7.7,6.7
4,G02,g03,02942607_G02.zsd-03662608_g03.zsd,-142.6498,274.8577,-498.627,5.4,7.7,9.9
5,G04,g03,02942608_G04.zsd-03662609_g03.zsd,97.926,-192.0434,346.4421,5.3,6.9,5.8
6,G04,g05,02942625_G04.zsd-03662626_g05.zsd,312.0537,204.7549,-46.2603,13.4,38.8,18.6
7,g06,G04,02912631_g06.zsd-02942625_G04.zsd,-701.8827,-464.6907,110.2367,5.7,10.7,6.5
8,g06,g05,02912631_g06.zsd-03662626_g05.zsd,-389.8306,-259.9329,63.9769,5.6,5.9,4.1
9,g06,G07,02912612_g06.zsd-02942612_G07.zsd,348.1171,196.0223,-5.7602,3.3,4.8,3.9
10,g06,g09,02912612_g06.zsd-03662611_g09.zsd,167.4629,-202.2958,413.9846,2.9,4.3,3.3
11,G07,g09,02942612_G07.zsd-03662611_g09.zsd,-180.6533,-398.3194,419.7436,3.6,5.2,4.2
12,G07,g11,02942620_G07.zsd-03662620_g11.zsd,272.0238,-428.7511,818.176,3.7,9.2,4.6
13,G08,g03,02912609_G08.zsd-03662609_g03.zsd,-533.794,-312.1199,27.573,3.4,4.9,3.4
14,G08,G04,02912609_G08.zsd-02942608_G04.zsd,-631.7205,-120.0778,-318.8694,5.1,6.6,5.6
15,G08,g09,02942622_G08.zsd-03662622_g09.zsd,237.6263,142.3337,-15.1263,4.8,7,4.1
16,G10,G07,02912620_G10.zsd-02942620_G07.zsd,-242.2288,144.2423,-388.9749,3.3,6.6,4.1
17,G10,g11,02912620_G10.zsd-03662620_g11.zsd,29.795,-284.5064,429.2001,3.6,9,4.6
18,G12,G01,02912632_G12.zsd-02942630_G01.zsd,-390.41,-259.3411,63.9442,4.1,8.1,6.8
19,G12,G08,02912623_G12.zsd-02942622_G08.zsd,-365.8394,75.1172,-390.1562,6.1,8.8,4.9
20,G12,g09,02912623_G12.zsd-03662622_g09.zsd,-128.2126,217.4485,-405.2823,5.5,8.7,4.3
21,G12,g11,02912632_G12.zsd-03662630_g11.zsd,324.4349,187.0474,-6.82,3.7,8.6,5.7
间接平差
近似坐标手算的，也可以机器算，用的到在github里面的那个求路径的函数，我估计不会去花时间完善了
GNSS控制网类，主要作用是初始化P、X0，以及方便整个网内基线与点的查询
class station(object):
'''undirected unweighted graph'''
def __init__(self):
self.num = int(0)
self.cat = "未知点"
self.name = ""
self.X = 0.0
self.Y = 0.0
self.Z = 0.0
self.N = 0.0
self.E = 0.0
self.H = 0.0
#用来标记是第几个未知点，用来列B矩阵
self.BeiZhu = int(0)
def init_station(self,line):
line = line.rstrip('\n')
line = line.split(',')
self.num = int(line[0])
self.cat = line[1]
self.name = line[2]
if self.cat == "已知点":
self.X = float(line[3])
self.Y = float(line[4])
self.Z = float(line[5])
self.N = float(line[6])
self.E = float(line[7])
self.H = float(line[8])
#已经算过近似坐标了
def init_station2(self,line):
line = line.rstrip('\n')
line = line.split(',')
self.num = int(line[0])
self.cat = line[1]
self.name = line[2]
#
self.X = float(line[3])
self.Y = float(line[4])
self.Z = float(line[5])
self.N = float(line[6])
self.E = float(line[7])
self.H = float(line[8])
def station_inf(self):
return [ self.num ,self.name, self.X, self.Y , self.Z , self.cat , self.BeiZhu]
def station_match(self, point_name):
if self.name == point_name:
return True
else :
return False
def cat_match(self):
if self.cat == "已知点":
return True
elif self.cat == "未知点":
return False
else:
return None
def beizhu(self,i):
self.BeiZhu = i
def unknow_num(self):
return self.BeiZhu
class Baseline(object):
'''undirected unweighted graph'''
def __init__(self):
self.num = int(0)
self.origin = ""
self.target = ""
self.name = ""
self.DX = 0.0
self.DY = 0.0
self.DZ = 0.0
self.sigema_DX = 0.0
self.sigema_DY = 0.0
self.sigema_DZ = 0.0
def init_baseline(self,line):
self.num = int(line[0])
self.origin = line[1]
self.target = line[2]
self.name = line[3]
self.DX = float(line[4])
self.DY = float(line[5])
self.DZ = float(line[6])
self.sigema_DX = float(line[7])
self.sigema_DY = float(line[8])
self.sigema_DZ = float(line[9])
def baseline_inf(self):
return [self.num,self.origin,self.target,self.name ,self.DX,self.DY,self.DZ,self.sigema_DX,self.sigema_DY,self.sigema_DZ,]
def baseline_match(self,baseline):
if baseline[0] == self.origin and baseline[1] == self.target:
return True,1,self.num
elif baseline[0] == self.target and baseline[1] == self.origin:
return True,-1,self.num
else:
return False,0,-1
class GNSSNet(object):
'''undirected unweighted graph'''
def __init__(self):
self.BaselineSet = []
self.StationSet = []
self.ControlPts = []
self.UnknowPts = []
def insert_Baseline(self,baseline):
self.BaselineSet.append(baseline)
#初始化站点信息
def insert_Station(self, station):
self.StationSet.append(station)
#已知点
def insert_ControlPoint(self, station):
self.StationSet.append(station)
#未知点
def insert_UnknowPoint(self, station):
self.StationSet.append(station)
def Net_baseline_match(self,baseline):
for Baseline in self.BaselineSet:
a,towards,number = Baseline.baseline_match(baseline)
if a:
return towards,self.BaselineSet[number-1].baseline_inf()
def station_match(self,station_name):
for station in self.StationSet:
if station.station_match(station_name):
return True,station.station_inf()
return False,None
def Station_Cat_Match(self,station_name):
for station in self.StationSet:
if station.station_match(station_name):
return True,station.cat_match(),station.station_inf()
return False,None,None
def init_P(self,P):
for Baseline in self.BaselineSet:
P[Baseline.num*3-3][Baseline.num*3-3] = 1.0 / (Baseline.sigema_DX * Baseline.sigema_DX)
P[Baseline.num*3-2][Baseline.num*3-2] = 1.0 / (Baseline.sigema_DY * Baseline.sigema_DY)
P[Baseline.num*3-1][Baseline.num*3-1] = 1.0 / (Baseline.sigema_DZ * Baseline.sigema_DZ)
return P
def init_L(self,L):
for Baseline in self.BaselineSet:
L[Baseline.num*3-3][0] = Baseline.DX
L[Baseline.num*3-2][0] = Baseline.DY
L[Baseline.num*3-1][0] = Baseline.DZ
return L
def init_X0(self,X0):
for station in self.StationSet:
if not station.cat_match():
num = int(station.unknow_num())
X0[num*3-3][0] = station.X
X0[num*3-2][0] = station.Y
X0[num*3-1][0] = station.Z
print(station.station_inf()[5] , station.station_inf()[6])
return X0
非常直白的平差过程
'''
间接平差
main.py
'''
import pandas as pd
import numpy as np
import sys
np.set_printoptions(threshold=sys.maxsize) #print显示完整array
from gnssnet import *
import math
def Save2Excel(mats,name):
data = pd.DataFrame(mats)
writer = pd.ExcelWriter("C:\\Users\\sheld\\Desktop\\result3\\"+ name + ".xlsx")
data.to_excel(writer, "page_1", float_format = '%.6f')#浮点数，精确到6位小数
writer.save()
writer.close()
def init_GNSSNet(G):
#读入所有的基线坐标信息
with open("data.csv",'r') as f:
for line in f.readlines():
b= Baseline()
b.init_baseline(line.split(','))
G.insert_Baseline(b)
f.close()
#已经手算过近似坐标了
with open("zhandian.csv",'r',encoding='UTF-8-sig') as f:
i = 1
for line in f.readlines():
p = station()
p.init_station2(line)
if not p.cat_match():
p.beizhu(i)
i = i + 1
G.insert_Station(p)
f.close()
n = 63
t = 24
r = n - t
global G
G = GNSSNet()
init_GNSSNet(G)
P = np.zeros([n,n], dtype = float)
L = np.zeros([n,1], dtype = float)
l = np.zeros([n,1], dtype = float)
'''
L^ = BX^ + d
l = L - (BX0 + d)
V = Bx^ - l
'''
B = np.zeros([n,t], dtype = float)
d = np.zeros([n,1], dtype = float)
X0 = np.zeros([t,1], dtype = float)
x = np.zeros([t,1], dtype = float)
#写入B,d矩阵
for baseline in G.BaselineSet:
bl_inf = baseline.baseline_inf()
num = bl_inf[0]
origin = bl_inf[1]
target = bl_inf[2]
#起点
found,ifknow,stat_inf = G.Station_Cat_Match(origin)
if ifknow:
d[3 * num - 3][0] = d[3 * num - 3][0] - stat_inf[2]
d[3 * num - 2][0] = d[3 * num - 2][0] - stat_inf[3]
d[3 * num - 1][0] = d[3 * num - 1][0] - stat_inf[4]
## print(num,bl_inf,"起点是已知点",stat_inf)
elif not ifknow:
B[3 * num - 3][3 * stat_inf[6] - 3] = -1
B[3 * num - 2][3 * stat_inf[6] - 2] = -1
B[3 * num - 1][3 * stat_inf[6] - 1] = -1
## print(num,bl_inf,"起点是未知点",stat_inf)
else:
print("有错误的站点名")
#末尾
found,ifknow,stat_inf = G.Station_Cat_Match(target)
if ifknow:
d[3 * num - 3][0] = d[3 * num - 3][0] + stat_inf[2]
d[3 * num - 2][0] = d[3 * num - 2][0] + stat_inf[3]
d[3 * num - 1][0] = d[3 * num - 1][0] + stat_inf[4]
## print(num,bl_inf,"终点是已知点",stat_inf)
elif not ifknow:
B[3 * num - 3][3 * stat_inf[6] - 3] = 1
B[3 * num - 2][3 * stat_inf[6] - 2] = 1
B[3 * num - 1][3 * stat_inf[6] - 1] = 1
## print(num,bl_inf,"终点是未知点",stat_inf)
else:
print("有错误的站点名")
## print(B[3 * num - 3])
## print(B[3 * num - 2])
## print(B[3 * num - 1])
## print(d[3 * num - 3])
## print(d[3 * num - 2])
## print(d[3 * num - 1])
#定权
P = G.init_P(P)
L = G.init_L(L)
x0 = G.init_X0(X0)
P = np.matrix(P)
L = np.matrix(L)
X0 = np.matrix(X0)
d = np.matrix(d)
B = np.matrix(B)
Q = P.I
print("X0:\n",X0)
print("L:\n",L)
print("d:\n",d)
print("B:\n",B)
l = L - np.dot(B,X0) - d
print("l:\n",l)
#化成毫米单位
l = np.matrix(np.dot(l, 1000))
#Nbbx^ - W = 0
Nbb = np.dot(np.dot(B.T, P), B)
W = np.dot(np.dot(B.T, P), l)
#矩阵的秩
print(np.linalg.matrix_rank(B, tol=None, hermitian=False))
print(np.linalg.matrix_rank(l, tol=None, hermitian=False))
#x^
x = np.dot(Nbb.I,W)
V = np.dot(B,x) - l
print('x',x,'\n')
print('V',V,'\n')
time = 1
V_total = V
x_total = x
##
##while abs(V.max()) > 0.00001 and time < 10000:
## L = L + V/1000
## X0 = X0 + x/1000
##
## l = L - np.dot(B,X0) - d
##
## l = np.matrix(np.dot(l, 1000))
## x = np.dot(Nbb.I,W)
## V = np.dot(B,x) - l
##
## V_total = V_total + V
## x_total = x_total + x
## time = time + 1
##
##
L = L + V/1000
X0 = X0 + x/1000
print('L^',L)
print("time:",time,'\n')
print('V_total',V_total,'\n')
print('x_total',x_total,'\n')
sigema02 = np.dot(np.dot(V_total.T,P),V_total)/r
sigema0 = math.sqrt(sigema02)
print(sigema0)
Save2Excel(B,"B")
Save2Excel(d,"d")
Save2Excel(L,"L^")
Save2Excel(l,"l")
Save2Excel(X0,"X^")
Save2Excel(V_total,"V_total")
Save2Excel(x_total,"x_total")
Save2Excel(P,"P")