چكيده به لاتين
Nowadays, awareness about the types of land-cover and human activities as baseline data for various plans in different parts are very important. Remote sensing and satellite image processing can be used to obtain such information without the need for physical presence in the region and confront many restrictions.
Identifying the areas where poppy plants are grown is one application of this approach.
In this thesis, we aim to identify the areas where poppy is cultivated by using satellite imagery. To this end, extracting suitable features of poppy farms from images can help classifying the regions well. In this project we try to increase recognition accuracy by using several features such as NDVI, OSAVI, IPVI and MSAVI2. Also in addition to using red and near-infrared bands which are usually used in identifying different plants, the reflectivity property of plants in three other bands of electromagnetic spectrum is also used. To increase the classification accuracy, Multi temporal images in the plant's growing cycle are utilized.
In this thesis, landsat7 and landsat8 images with 30 meters resolution are used. To solve the problem of low resolution images, bicubic interpolation was employed. After performing required pre-processing and appropriate feature extraction phases, classification was performed using SVM classifier and a maximum accuracy of 97.4% was achieved. The results of this research can be used to detect pieces of poppy cultivated lands.
