Research On Detecting And Resuming Incomplete Spatial Data

Incomplete Spatial Data (for short ISD) includes missed data and error during the collecting, inputting, disposing and employing of spatial data. However, the ISD are emerging widely in many application domain, such as geology, mine, remote sensing, mapping, city planning, house property, which touch severely the preciseness and reliability of analysing result so that the spatia data might not been utilized exactly. So, studying on detectiing and resuming ISD has very tremendous theoretic-and-realism significance.Firstly, after summarizing all related works of the traditional detectiing and resuming the incomplete data, this paper expounds the theory of detecting the incomplete data based on statistics theory, and then discusses the shortage of the detecting ISD method based on statistics theory. In succession, two detecting methods are proposed, which are basing on k nearest neighbors(for short k-NN) and basing on constrained nearest neighbors. Whereafter, this paper researches the resuming methods of ISD, which are basing on spatial interpolation, constrained interpolation and Geo-statistics theory.For the k-NN based detecting ISD method, k-NN of every spatial entity should been generated. Then, in every k-NN, the classical rule of "treble standard deviation" has been utilized to check whether the spatial entity is a ISD. However, the terrain factors, such as elevation and slope, have effect the non-spatial attribute value, so the k-NN based detecting ISD is defective because of ignoring the terrain factors. Thus, a constrained nearest neighbors based detectiing ISD method is proposed. The method divides the research region into some sub-regions according to terrain factors. Every sub-region's terrain factors are similitude. And then, in every sub-region, the constrained nearest neighbors are constructed. Lastly, in every constrained nearest neighbors, "the treble standard deviation" rule is employed to detect ISD.On the resuming method of incomplete spatial data, the primary researches are as follows:(1)constrained ascertainable interpolation based resuming ISD's value method, which divides the region into some sub-regions according the terrain factors, and then in every sub-region, the method is used to resume the ISD's value; (2) Geostatistics based resuming ISD's value method, which utilizes the geostatistics theory to estimate value of the ISD.At last, the test data, procession and result are introduced, and the design and the implement of the detecting and resuming ISD system are presented in detail.