-
Notifications
You must be signed in to change notification settings - Fork 0
/
studya.Rmd
188 lines (138 loc) · 6.25 KB
/
studya.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
---
title: "Pilot study A: Application of the Parrot Sequoia multispectral camera for surface water quality assesment"
subtitle: "Deriving turbidity and suspended matter from surface water"
output: html_document
---
```{r, out.width = "100%", echo=FALSE}
knitr::include_graphics("drone.PNG")
```
Parrot Sequoia multispectral camera mounted under a drone.
***
###Approach
{width="100%"}
*10 validation points (w1~w10)
**Study area Lollebeek (lat: 51.489167,long: 6.055111)**
{width="75%"}
Field campaign were carried out on April 10, 2018.
***
### Reflectance maps
{width="74%"}
{width="70%"}
{width="70%"}
Calculated band ratio and band index reflectance maps (scale is based on specific percentile range )
***
### Data analysis
**Descriptive statistics**
```{r, echo=FALSE, message=FALSE, warning=FALSE}
library(ggplot2)
library(reshape)
library(gridExtra)
library(dplyr)
library(scales)
library(magrittr)
data <- read.csv("meetpunten.csv")
easy_bp <- function(df,var,x_as,y_as,kleur){
subset <- df[c(var)]
subset.m <- melt(subset)
is_outlier <- function(x){
return(x < quantile(x, 0.25) - 1.5 * IQR(x) | x > quantile(x, 0.75) + 1.5 * IQR(x))
}
steps <- (max(subset.m$value)-min(subset.m$value))/5
means <- round(mean(subset.m$value),3)
means_mu <- paste(c("μ : ",means),collapse=" ")
medians <- round(median(subset.m$value),2)
medians_m <- paste(c("M: ",medians),collapse=" ")
subset.m %>%
group_by(variable)%>%
mutate(outlier = ifelse(is_outlier(value), value, as.numeric(NA))) %>%
ggplot(., aes(x = "", y = value))+
scale_y_continuous(labels = round(seq(min(subset.m$value),max(subset.m$value),steps),3),
breaks=seq(min(subset.m$value),max(subset.m$value),steps),
limits=c(min(subset.m$value),max(subset.m$value)))+
geom_boxplot(fill=kleur,colour="#636e72",fatten=NULL)+
geom_point(color="#636e72",size=1.5)+
geom_point(aes(y=outlier),color="red",size=1.5)+
xlab (x_as)+
ylab (y_as)+
theme_gray()+
geom_hline(yintercept=medians, linetype="solid",size=0.3,color = "#636e72")+
geom_text(aes(0,medians,label=medians_m,vjust=-0.5,hjust=0),size=2.5,fontface="italic")
}
red <- easy_bp(data,"red_mean","Red","Value","#ffeaa7")
rededge <- easy_bp(data,"rededge_me","Red Edge","Value","#ffeaa7")
nir <- easy_bp(data,"nir_mean","NIR","Value","#ffeaa7")
re_nir <- easy_bp(data,"re_nir_mea","Red Edge/ NIR","Value","#74b9ff")
r_nir <- easy_bp(data,"r_nir_mean","Red/ NIR","Value","#74b9ff")
g_nir <- easy_bp(data,"g_nir_mean","Green/ NIR","Value","#74b9ff")
r_re <- easy_bp(data,"r_re_mean","Red/ Red Edge","Value","#74b9ff")
ndvi <- easy_bp(data,"ndvi_mean","NDVI","Value","#55efc4")
ndwi <- easy_bp(data,"ndwi_mean","NDWI","Value","#55efc4")
ndre <- easy_bp(data,"ndre_mean","NDRE","Value","#55efc4")
turb <- easy_bp(data,"ntu_lab","Turbidity (lab)","NTU","#ff7675")
tsm <- easy_bp(data,"tsm_lab","TSM (lab)","mg/l","#ff7675")
grid.arrange(red,rededge,nir,re_nir,r_nir,g_nir,r_re,
ndvi,ndwi,ndre,turb,tsm,ncol=4)
```
Boxplots from 10 validation points.
M = median
**Best prediction model**
No relationship can be found between turbidity and suspended matter in our case. However, there is a relationship between turbidity and the Red/ NIR ratio. Three validation points (w2, w4, w7) were excluded to create the linear regression model below.
```{r, echo=FALSE, message=FALSE, warning=FALSE}
library(plotly)
library(ggplot2)
library(plyr)
easy_lm <- function(df,var_x,var_y,x_as,y_as,kleur,titel) {
## steps and model
steps_x <- (max(var_x)-min(var_x))/5
steps_y <- (max(var_y)-min(var_y))/5
model <- lm(var_y~var_x)
dat <- predict(model,interval="confidence")
data$inside <- ifelse(data$ntu_lab < dat[,"upr"] & data$ntu_lab > dat[,"lwr"], "", as.character(data$label))
subset <- df[c(var_x)]
subset.m <- melt(subset)
##equation stamp
lm_eqn = function(m) {
l <- list(a = format(coef(m)[1], digits = 2),
b = format(abs(coef(m)[2]), digits = 2),
r2 = format(summary(m)$r.squared, digits = 3));
if ( as.numeric(l[3]) <= 0.001 ){
l[3] <- as.character(round(as.numeric(l[3])),3)
}
if (coef(m)[2] >= 0) {
eq <- substitute(italic(y) == a + b %.% italic(x)*","~~italic(R)^2~"="~r2,l)
} else {
eq <- substitute(italic(y) == a - b %.% italic(x)*","~~italic(R)^2~"="~r2,l)
}
as.character(as.expression(eq));
}
##plotting
ggplot(df,aes(x=var_x,y=var_y)) +
stat_smooth(method=lm,color=kleur,fullrange=TRUE)+
scale_x_continuous(labels = round(seq(min(var_x),max(var_x),steps_x),3),
breaks=seq(min(var_x),max(var_x),steps_x))+
scale_y_continuous(labels = round(seq(min(var_y),max(var_y),steps_y),3),
breaks=seq(min(var_y),max(var_y),steps_y))+
geom_point(color="#34495e",alpha=0.8) + geom_line(color="#34495e",alpha=1,linetype="dotted")+
xlab (x_as)+
ylab (y_as)+
theme_grey()+
annotate("text",x=-Inf,y=Inf,label=lm_eqn(model),hjust=-0.75,vjust=1.0,parse=TRUE,size=4)+
geom_text(aes(label=data$inside,hjust=0.5,vjust=-0),size=4,color="#e74c3c",fontface="bold")+
theme(axis.text=element_text(size=12))+
labs(title=titel)+
scale_size_area()
}
data <- read.csv("meetpunten.csv")
data <- data[-c(7,4,2),]
r_nir <- easy_lm(data,data$r_nir_mean,data$ntu_lab,"Value Red/ NIR","Turbidity (NTU)(lab)", "#3498db","Model with 7 validation points")
r_nir
```
**Turbidity equation**
Based on the previous regression model the following turbidity equation can be derived:
$Turbidity (NTU) = 5.6 +13.4(\frac{Red}{NIR}) \pm 0.7$
***
### Turbidity map
Applying the turbidity equation to the Red/ NIR ratio reflectance map resulted into the turbidity coverage map below. The turbidity range scales from 14 to 19 NTU. Notice the patterns of similarity between the stream bed and turbidity layer.
**Interactive turbidity map**
<iframe width="100%" height="650px" src="turbidity.html" frameborder="0" allowfullscreen></iframe>
[Click here for a full screen preview ](https://williamtjiong.github.io/remotesensing/turbidity.html)