##this is the example of my research,NDVI is the undependent variable while Biomass is the dependent variable
triandata<-read.table("unave_biomass$vis_622.csv",head=T,sep=",")## import your data into R
library(ggplot2) ##loading the package
y<-log(Biomass)
x<-log(NDVI)
##suppose you use the linear model
modelNDVI_3<-lm(y~x,data=traindata)
##set the range of fitting line 
xmin<-min(log(traindata$NDVI))
xmax<-max(log(traindata$NDVI))

NDVIpredicted<-data.frame(x=seq(xmin,xmax,length.out = 100))
##get the fitted value between the range of fitting line
NDVIpredicted$Biomass<-predict(modelNDVI_3,NDVIpredicted)
NDVIpre<-data.frame(NDVI=NDVIpredicted[,1],Biomass=NDVIpredicted[,2])
#using the ggplot2 to draw the scatter plot
p<-ggplot(traindata,aes(x=NDVI,y=Biomass))+geom_point(size=3)+scale_fill_brewer(palette = "Pastel2")+theme_bw()+guides(shape=guide_legend(title=NULL))
q<-p+guides(fill=FALSE)+theme(legend.position=c(0.2,0.8))
u<-q+scale_y_continuous(limits=c(0,1),breaks = c(0,0.2,0.4,0.6,0.8,1.0))+
  labs(y=expression(paste("Biomass",(kg/m^2))))+
  #labs(y=expression(paste("Aboveground biomass",(g/m^2))))+
  # ylab("Estimated Biomass (g/m^2)")+xlab("Measured Biomass (gm^2)")+
  scale_x_continuous(limits=c(0,0.8),breaks = c(0,0.2,0.4,0.6,0.8))
r<-u+theme(axis.title.x=element_text(size = 22),axis.title.y=element_text(size = 22))
s<-r+theme(axis.text.x=element_text(size=rel(2)),axis.text.y=element_text(size=rel(2)))
#added the fitting line according to the dataframe of "NDVIpre"
t<-s+geom_line(data=NDVIpre,size=1)+annotate("text",x=0.40,y=0.99,label="(a)",size=9)
t