Case introduction

 

Sampling survey to the total area of the forest of Heshun county in Shanxi province in year 2000

 

Description:

We need to take some samples from 2,697,460 image units to investigate (the minimum investigated unit is A TM image ) the area of forest coverage of Heshun county in Shanxi province in year 2000, require that the mean variance of the investigated forest area isn't greater than 65m2 (namely, User expectation accuracy is 65m2). According to the experience, standard deviation and investigation cost of each layer unit can be estimated.
 
Sampling method:
Spatial stratified sampling
 

Objective:

To estimate the forest area of Heshun county in year 2000 by sampling

 

Condition:

  1. Expectation variance isn't greater than 65m2
  2. According to the previous investigated results, the standard deviation ( Standard deviation is the root of the mean of sum of squares of deviations, which reflects the dispersion degree of this layer) and the cost of each layer is estimated as follows:
2
 

Request:

  1. Calculate the optimal sample size n according to the known information
  2. Population (2,697,460 pixels) is equally divided into n groups, namely, the whole zone is divided into n grids with the same size, and number the samples for each group; In the first group, the number L of the starting individual is confirmed by simple random sampling, then extract the Lth sample of the other groups.
  3. Estimate the total forest area and sampling accuracy according to the sampling pixels and investigated data.