چکیده:
از آنجاکه زمین به عنوان یکی از نهاده های بخش تولید میباشد، نه تنها در اقتصاد کشاورزی و منابع طبیعی بلکه در اقتصاد کل کشور نقش به سزایی دارد و توجه به زمین و تغییرات به وجود آمده در آن، امری ضروری است. تحقیق حاضر به منظور بررسی روند تغییرات کاربری اراضی حوضه آبریز اوجان چای با استفاده از سنجش از دور و GIS انجام شده است. از تصاویر چند زمان سنجدة TMسال 1987و +ETM سال2002 و +ETM سال 2015 استفاده شد و نقشه های کاربری اراضی بر اساس پردازش رقومی حداکثر احتمال و ماشین بردار پشتیبان تهیه شد.نقشه های کاربری اراضی به همراه اطلاعات زمینی وارد محیط GIS شدند و میزان و نوع تغییرات کاربری اراضی در منطقه به دست آمد. با توجه به نتایج ، روش SVM برای برآرود تغییرات منطقه مورد مطالعه کارآمدتر بوده به گونه ای که نتایج استخراج شده از درصد دقت و ضریب کاپای بالاتری برخوردار است. از نتایج ارزیابی ها می توان چنین استخراج کرد که روند تغییرات کاربری در برخی کاربری ها مانند اراضی زراعی و مراتع بالا است. اراضی زراعی از 33درصد به 37 درصد در طی بازه ی زمانی 28 ساله نوسان داشته و اراضی مرتعی نیز از 51 درصد به 49 درصد کاهش داشته که این دو کاربری بیشترین تغییرات را داشته اند. در نهایت با توجه به اینکه بیشتر تغییرات مربوط در تغییر مرتع به اراضی زراعی و بالعکس بوده است، بنابراین لزوم تمرکز فعالیتهای مدیریت و اصلاح اراضی بر روی این نوع کاربری افزایش مییابد.
Introduction Land is one of the inputs of the manufacturing sector, not only in the agricultural economy and natural resources, but also in the overall economy of the country, and it is necessary to pay attention to the land and the changes that have taken place in it. In recent decades, land use change under the influence of environmental and human factors has caused serious effects on the environment, economy and society. Therefore, having information about the type of land use and its changes over time is one of the important issues in planning and policy making in the country. The present study was conducted to investigate the trend of land use change changes in Ojan Chay catchment area using remote sensing and GIS. Over time, land cover patterns and, consequently, land use change dramatically, and the human factor can play the greatest role in this process. The present study was conducted to investigate the trend of land use change changes in Ojan Chay catchment area using remote sensing and GIS. Based on this, the multi-timer images of TM 1987 and ETM of 2002 and ETM of 2015 were used, and land use maps were prepared based on maximum probability digital processing and backup vector machine. Then, land use maps along with land information entered the GIS environment and the amount and type of changes in land use in the area were obtained Materials and methods In order to study land use changes, the image of Landsat 5 (TM) satellites in 1987 and Landsat 7 (+ ETM) in 2002 and the image and Landsat 8 (+ ETM) in 2015 have been used. The Esther's high-altitude digital model was taken from a global observation by the US Geological Survey with a pixel size of 30 meters to extract the basin and topographically correct the image. For this purpose, pre-processing steps, which include atmospheric, geometric and altitude corrections, were performed on satellite imagery. After cutting the study area from the image, using the 1: 25000 topographic maps and through ground visit and visual interpretation, the training areas for each user class were harvested in two stages before classification and after classification. At this stage, the selection of educational areas was well distributed. In the field visit stage, through face-to-face interviews with the residents of the area, in order to determine the training points using GPS, several points were used that were assured of their use in 1987 and 2001. Taking into account the separability of the bands, the appropriate bands were selected for classification 5, 4 and 7. Then, the images related to the desired years were converted to polygon format, the area of the desired coatings was calculated. Then the images of the cover were re-classified each year, and the final drawings were reduced to four floors each year. Then each image was converted to formatting and its usage area was obtained. By entering descriptive information tables into Excel software each year, the trend of land cover changes between 1987 and 2015 was estimated. Using the land cover maps obtained for each matrix period, the conversion status of the land cover classes between the two time periods has been calculated. From the coverage plan of 1987 and 2002, the first status conversion matrix (Table 4) and from the coverage map from 2002 to 2016, the second status conversion matrix (Table 5) has been calculated. These matrices contain information and convert each class to other classes. The transfer area matrix represents the number of pixels that are converted from one class to another. Results and discussion The present study was conducted to investigate the trend of land use change changes in Ojan Chay catchment area using remote sensing and GIS. Based on this, the multi-timer images of TM 1987 and ETM of 2002 and ETM of 2015 were used, and land use maps were prepared based on maximum probability digital processing and backup vector machine. Then, land use maps along with land information entered the GIS environment and the amount and type of changes in land use in the area were obtained. According to the obtained results, SVM method is more efficient for estimating the changes in the study area, so that the extracted results have a higher percentage of accuracy and capability coefficient. From the results of the evaluations, it can be deduced that the trend of land use changes in some land uses, such as agricultural lands and pastures, is high. Agricultural lands fluctuated from 33 percent to 37 percent over a 30-year period, and rangelands fell from 51 percent to 49 percent, with the two most variable uses. There are few land use changes in orchards, irrigated agricultural lands and settlements. The results suggest that primarily the use of remote sensing technology with GIS can enhance the ability to access land use information. However, due to the similarity between the range of use of pastures and abandoned lands, there is a need to implement more accurate methods. On the other hand, due to the close relationship between population status and land use change, it can be concluded that the process of change found can be a result of population change that requires further research. Finally, due to the fact that most of the changes have been related to the change of rangeland to agricultural lands and vice versa, therefore, the need to focus on management activities and land improvement on this type of use increases Conclusion The findings of the present study suggest that data from remote sensing and digital interpretation, along with educational sampling, provide excellent information to users. Due to the fact that the aim of the study was to compare the two algorithms of the backup vector machine and the maximum probability, so in the support vector machine method, a multi-stage kernel with 5 degrees was used for classification due to its greater accuracy than other grades (Andriani, 1393, 103). In this study, changes in land cover in the use of gardens and irrigated agricultural lands and residential areas, rainfed lands and pastures were used using satellite images in the period of 1366-1395 with maximum similarity algorithm and support vector machine. The transfer of land uses, respectively, in the winter of 2007-2008 is related to agricultural lands, pastures, orchards, irrigated agricultural lands and residential areas, and it has been determined that the area of agricultural lands has increased in 2016, but in the year 1361, 1381 is not. . The area of residential areas during the mentioned years has not changed significantly and the highest increase has been from 1987 to 2002. In 2001, with 910,8592 hectares, the maximum increase was possible. Rangelands have also experienced a declining trend during the period 1361-1381 and its area has decreased every year compared to the following year, but the area of pastures during the years 1381-1395 has gone up. Considering the outputs of the maps in the three time periods of 1987, 2002 and 2015, it has been observed that most of the changes in employment have been related to rainfed and rangeland lands, which have resulted in soil degradation and pasture degradation.
خلاصه ماشینی:
3 Gumeh, 4 Conway & Lathrop 5 Torrens & Osullivan, 6 Khorn&et al 7 Post Classification Method 8 Radke, R.
در اين تحقيق جهت بررسي تغييرات مساحت پوشش زمين حوضۀ آبريـز اوجـان چـاي از تصاوير ماهواره اي لندست ٥-٧ و ٨ بعد از اعمال تصحيحات لازم استفاده شد و به بررسـي دقـت الگـوريتم SVM بـا کرنل چندجمله اي و روش حداکثر احتمال در استخراج تغييرات کاربري اراضي در اين محدوده پرداختـه شـده اسـت که تبديل اراضي و تغيير کاربري در سطح آن اتفاق افتاده که مغاير با اصول زيست محيطي و توسـعه پايـدار مـي باشـد.
جدول ١- کلاس هاي قديم و جديد لندست {مراجعه شود به فایل جدول الحاقی} (مأخذ: يافته هاي نويسندگان ) شکل ١ مراحل استخراج کاربري اراضي با روش SVM و MLC را نشان مي دهد: (رجوع شود به تصویر صفحه) شاخص پوشش گياهي ١ (NDVI)، شاخص خاک باير٢ ( BI) و تحليل مؤلفه اصلي (PCA) بـه عنوان بانـد ورودي در طبقه بنديها شرکت داده شد و دقت آن ها موردبررسي قرار گرفت .
٣- نتايج در پژوهش حاضر سعي بر آن است تا تغييرات کاربري اراضـي در بـازه ي زمـاني سـال هاي ١٣٩٥-١٣٦٦(١٩٨٧- ٢٠١٥) با روش هاي ماشين بردار پشتيبان و حداکثر مشـابهت و ارزيـابي ايـن تغييـرات و درنهايـت بـه دسـت آوردن ماتريس انتقال وضعيت مد نظر قرار گرفته شود.