Heyelanların izlenmesi araştırmalarının en önde gelen amacı, heyelanın önceden haber alınmasıdır. Bunu izleyen amaç ise önlem geliştirmektir. Bu amaçlar için, hareket edecek kitlenin büyüklüğü ile birlikte hareket yönünün saptanması gereklidir. Heyelan gerçekleşmeden, kontrol noktalarındaki kanıtlanmış küçük zemin hareketlerini gösteren deformasyon vektörleri bir ölçüde heyelanın yönünü göstermektedir. Hareket edecek zemin kitlesinin (heyelan bloklarının) belirlenmesi için, sezginin ötesinde matematik modellerin geliştirilmesi gereklidir. Bu bağlamda, mevcut verilerden seçilen girdi değişkenlerinden çıktı değişkenlerinin elde edilmesini sağlamak amacıyla bulanık küme ilkelerini kullanan Bulanık Çıkarım Sistemleri (BÇS) elverişli bir yöntem olarak öne çıkmaktadır. Heyelan bloklarının belirlenmesi 2D (yatay) + 1D (düşey) konum bileşenlerinin kombinasyonu ile (Quasi) 3D bir sistemde gerçekleştirilmiştir. Blok sınırlarının belirlenmesi amacıyla; kontrol noktalarının deformasyon vektörleri, afin koordinat dönüşümü ile irdelenmekte ve farklı bloklardaki kontrol noktalarının belirlenmesi iteratif bir çözümle yapılmakta ve çözüm algoritmasının bazı adımları Bulanık Çıkarım Sistemleri (BÇS) ile gerçekleştirilmektedir. Ölçme kampanyaları arasında gerçekleştirilen afin transformasyonundan elde edilen karesel ortalama hata (s0), gerilme (strain) parametresi bileşenleri (e1, e2), iqr değeri, deformasyon vektörlerinin yönü ve büyüklüğü ile kontrol noktalarına ait düşey konum değişimi gibi veriler BÇS’nin girdi değişkenleridir. Bu çalışmada, Marmara Denizi kıyısındaki Büyükçekmece-Gürpınar (İstanbul) heyelan bölgesinde gerçekleştirilen, GPS ölçmeleri ile heyelan izleme projesinin verileri kullanılarak bölgedeki heyelan blokları BÇS ile belirlenmiştir.

Landslides are serious geologic disasters that threat human life and property in every country. On account of their negative consequences, landslides loom large among natural hazards. They not only cause life and property losses where they occur but also harm economically important structures such as transportation lines (highways, railways) and agricultural fields or arable land. In addition, landslides are one of the most important natural phenomena, on which directly or indirectly effect countries’ economy. Turkey is also the country that is under the threat of landslides. Landslides frequently occur in all of the Black Sea region as well as in many parts of Marmara, East Anatolia, and Mediterranean regions. Since these landslides resulted in destruction, they are ranked as the second important natural phenomenon that comes after earthquake in Turkey. Therefore it is needed to monitor the landslides. Landslide areas can be divided into different blocks moving in different directions with different velocities. Determination of block boundaries provides important information that can be useful in implementing more effective landslide monitoring studies and in the studies aiming at reducing landslide effects. Information regarding the relative movements of the blocks is a very important indicator for future movement of the blocks. Coordinate transformation is one of the widely applied issues in geodesy. Coordinates in one coordinate system can be determined in another system through transformation. Transformation type in geodetic studies is decided upon the objective of transformation and the number of common points available. In order to determine the block boundaries, displacement vectors of observation points are analysed employing affine transformations. The determination of observation points on different blocks can be achieved in an iterative solution. Some steps in the solution algorithm can be accomplished by Fuzzy Inference Systems (FIS). In the fuzzy logic approach such parameters as the strain parameters obtained from transformations and the standard deviation of unit weight and iqr value and deformation vector and height changes are used as input parameters. Fuzzy Systems was developed by Lotfi Zadeh in the mid 1960’s as an alternative to conventional reasoning and probability theory. Most of the conventional methods employed for modelling and reasoning are straight-forward, numerical and yield exact solutions. However, real cases are somehow uncertain and fuzzy in many respects. As a result of the lack of information, future position of a system cannot be properly predicted. So there are two problems for the outside world applications. (1) Real cases are mostly rigorous, non-numerical and cannot be defined in a certain way. (2) For the definition of a real system much more information than instantaneous definition and understanding of a person is needed. In such cases, human decisions are based on uncertainties and insensitivities that can be expressed orally. Fuzzy logic is one of the methods used for modelling such decision making processes The data collected at Buyukcekmece Gurpinar Village, where a landslide monitoring project for the determination of ground movements was carried out between 1996 and 1998, was used for landslide block determination through Fuzzy Logic Method. In this research, an alternative methodology for 3D determination of landslide blocks was applied. When 3D affine transformation between measurement periods is applied, the number of unknown parameters is twelve. For the solution of unknown parameters at least four common points are needed. In the least squares estimation of transformation parameters, generally more common points than required are used. Affine transformation, therefore, requires at least five common points. This may be a problem in geodetic deformation monitoring, because generally there are not so many points in such studies. In order to overcome this drawback, this problem could be solved by the combination of 2D coordinate transformation and vertical components. In block determination algorithm, fuzzy inference systems can also be employed. The input values in FIS application are strain parameters, directions of displacement vectors, displacement values, root mean square error, and the mean value of height changes. The results show that fuzzy logic approach in the determination of landslide block boundaries could be employed as a powerful tool.