Assessing Industrial Development Influence on Land Use and Land Cover Change Detection in Güzelhisar Basin with Cloud-Based Machine Learning Techniques

Şevki Danacıoğlu; Hüseyin Can Öngül

Article Details

Assessing Industrial Development Influence on Land Use and Land Cover Change Detection in Güzelhisar Basin with Cloud-Based Machine Learning Techniques

Authors:

Şevki Danacıoğlu | Hüseyin Can Öngül

Publisher:

Ege University

Main Topics:

Uzaktan Algılama | Beşeri Coğrafya

Journal:

Ege Coğrafya Dergisi Vol.32 Issue 1, pp.135-150

Document Type:

Article

Publication Date:

6/2023

Document Language:

Turkish

Document ID (Org No.):

2001295206

pISSN:

2636-8056

Keywords:

Abstract:

 Industrial activity is the main driving force behind urban growth, influencing the socioeconomic status of an area by creating employment opportunities. Land use and land cover LULC are influenced by various factors such as ecological conditions, geological and geomorphological features, vegetation characteristics, and socioeconomic structure. Monitoring LULC changes, their intensity, direction, and underlying causes provides valuable knowledge for sustainable development planning. Remote sensing RS is widely considered the most cost-effective and practical approach for obtaining regional and global LULC information. The aim of this study is to investigate the impact of industrial activity on LULC in the Güzelhisar Basin. Using satellite imagery and machine learning algorithms, the LULC status from 1995 to 2022 was determined. The LULC classes were classified as Water Surface, Forest Area, Agricultural Area, Bare Surface, and Built-Up Area. The research utilized LANDSAT satellite images with a 30-meter resolution. To enhance accuracy, various indices including the NDVI, SAVI, NDWI, NBLI, BSI, and NDBI were calculated for the years 1995 and 2022. The Random Forest RF machine learning algorithm was employed for satellite image classification. The Google Earth Engine GEE platform was utilized for image acquisition and classification. Classification accuracy was evaluated using the Error Matrix, Users Accuracy, Producers Accuracy, Overall Accuracy, and Kappa Coefficient. The findings indicate a significant increase in built-up areas and a decrease in agricultural and bare areas within the survey area. This demonstrates the impact of industrial operations on urbanization, considering the amount of increase in anthropic surfaces. The study thoroughly evaluates the capabilities of the GEE platform, machine learning-based classification algorithm, and the entire process from image classification to the assessment of obtained findings. These findings are crucial for future studies and the broader implementation of open data sources and cloud-based platforms in RS and Geographic Information Systems research.

 Industrial activity is the main driving force behind urban growth, influencing the socioeconomic status of an area by creating employment opportunities. Land use and land cover LULC are influenced by various factors such as ecological conditions, geological and geomorphological features, vegetation characteristics, and socioeconomic structure. Monitoring LULC changes, their intensity, direction, and underlying causes provides valuable knowledge for sustainable development planning. Remote sensing RS is widely considered the most cost-effective and practical approach for obtaining regional and global LULC information. The aim of this study is to investigate the impact of industrial activity on LULC in the Güzelhisar Basin. Using satellite imagery and machine learning algorithms, the LULC status from 1995 to 2022 was determined. The LULC classes were classified as Water Surface, Forest Area, Agricultural Area, Bare Surface, and Built-Up Area. The research utilized LANDSAT satellite images with a 30-meter resolution. To enhance accuracy, various indices including the NDVI, SAVI, NDWI, NBLI, BSI, and NDBI were calculated for the years 1995 and 2022. The Random Forest RF machine learning algorithm was employed for satellite image classification. The Google Earth Engine GEE platform was utilized for image acquisition and classification. Classification accuracy was evaluated using the Error Matrix, Users Accuracy, Producers Accuracy, Overall Accuracy, and Kappa Coefficient. The findings indicate a significant increase in built-up areas and a decrease in agricultural and bare areas within the survey area. This demonstrates the impact of industrial operations on urbanization, considering the amount of increase in anthropic surfaces. The study thoroughly evaluates the capabilities of the GEE platform, machine learning-based classification algorithm, and the entire process from image classification to the assessment of obtained findings. These findings are crucial for future studies and the broader implementation of open data sources and cloud-based platforms in RS and Geographic Information Systems research.

 Industrial activity is the main driving force behind urban growth, influencing the socioeconomic status of an area by creating employment opportunities. Land use and land cover LULC are influenced by various factors such as ecological conditions, geological and geomorphological features, vegetation characteristics, and socioeconomic structure. Monitoring LULC changes, their intensity, direction, and underlying causes provides valuable knowledge for sustainable development planning. Remote sensing RS is widely considered the most cost-effective and practical approach for obtaining regional and global LULC information. The aim of this study is to investigate the impact of industrial activity on LULC in the Güzelhisar Basin. Using satellite imagery and machine learning algorithms, the LULC status from 1995 to 2022 was determined. The LULC classes were classified as Water Surface, Forest Area, Agricultural Area, Bare Surface, and Built-Up Area. The research utilized LANDSAT satellite images with a 30-meter resolution. To enhance accuracy, various indices including the NDVI, SAVI, NDWI, NBLI, BSI, and NDBI were calculated for the years 1995 and 2022. The Random Forest RF machine learning algorithm was employed for satellite image classification. The Google Earth Engine GEE platform was utilized for image acquisition and classification. Classification accuracy was evaluated using the Error Matrix, Users Accuracy, Producers Accuracy, Overall Accuracy, and Kappa Coefficient. The findings indicate a significant increase in built-up areas and a decrease in agricultural and bare areas within the survey area. This demonstrates the impact of industrial operations on urbanization, considering the amount of increase in anthropic surfaces. The study thoroughly evaluates the capabilities of the GEE platform, machine learning-based classification algorithm, and the entire process from image classification to the assessment of obtained findings. These findings are crucial for future studies and the broader implementation of open data sources and cloud-based platforms in RS and Geographic Information Systems research.

 Endüstriyel faaliyetin varlığı, kentsel büyümenin ana itici gücüdür ve istihdam fırsatları yaratarak bölgenin sosyoekonomik durumunu etkilemektedir. Arazi Örtüsü ve Arazi Kullanımı AÖAK, ekolojik koşullar, jeolojik ve jeomorfolojik özellikler, bitki örtüsü özellikleri gibi biyotik ve abiyotik faktörler ile sosyoekonomik yapı tarafından etkilenmektedir. AÖAK değişimlerini, bunların yoğunluğunu, değişim yönünü, etkenlerini ve izlemek, sürdürülebilir kalkınma planlaması için önemli bilgiler sağlamaktadır. Uzaktan Algılama UA, bölgesel ve küresel AÖAK bilgisi elde etmek için en ekonomik ve uygulanabilir yaklaşım olarak kabul edilmektedir.. Çalışmanın amacı Güzelhisar Havzasında sanayi faaliyetlerinin AÖAK durumu üzerindeki etkisini araştırmaktır. Bu bağlamda uydu görüntüleri kullanarak makine öğrenme algoritması ile 1995-2022 yıllarına ait AÖAK durumu tespit edilmiştir. Sınıflandırmada AÖAK sınıfları Su Yüzeyi, Orman Alanı, Tarım Alanı, Açık Yüzey ve Beşeri Yüzey olarak belirlenmiştir. Araştırmada 30 m çözünürlüğü ile LANDSAT uydu görüntüleri kullanılmıştır. Normalize Edilmiş Fark Bitki Örtüsü İndeksi NDVI, Toprakla Düzeltilmiş Bitki Örtüsü İndeksi SAVI, Normalize Edilmiş Fark Su İndeksi NDWI, Normalize Edilmiş Açık Yüzey İndeksi NBLI, Çıplak Toprak İndeksi BSI, Normalize Edilmiş Fark Yerleşim Alanı İndeksi NDBI indeksleri 1995 ve 2022 yılları için hesaplanarak doğruluğu artırmak amacıyla kullanılmıştır. Uydu görüntülerinin sınıflandırmasında Rastgele Orman RF makine öğrenme algoritması tercih edilmiştir. Görüntülerin elde edilmesinde ve sınıflandırma işlemlerinde Google Earth Engine GEE platformu kullanılmıştır. Sınıflandırma doğruluğu hata matrisi, kullanıcı doğruluğu, üretici doğruluğu, genel doğruluk ve Kappa Katsayısı ile hesaplanmıştır. Sonuç olarak araştırma sahasında beşeri yüzeylerde önemli miktarda artış meydana gelirken, tarım alanlarında ve açık yüzeylerde azalma olduğu tespit edilmiştir. Beşerî yüzeylerdeki artış miktarı dikkate alındığında bölgede sanayi faaliyetlerine bağlı istihdam potansiyelinin kentleşme üzerindeki etkisini göstermektedir. Araştırma kapsamında GEE platformunun yetenekleri, makine öğrenmesine dayalı sınıflandırma algoritması, sınıflandırma süreçleri ve elde edilen bulguların değerlendirilmesine kadar olan tüm süreç performansları değerlendirilmiştir. Bu açıdan çalışmanın tüm sonuçları, gelecekte yapılacak çalışmaların geliştirilmesi, ayrıca UA ve Coğrafi Bilgi Sistemleri araştırmalarında açık veri kaynaklarının ve bulut tabanlı platformların yaygınlaşması açısından önem arz etmektedir.

Article Details

Assessing Industrial Development Influence on Land Use and Land Cover Change Detection in Güzelhisar Basin with Cloud-Based Machine Learning Techniques

Authors:

Şevki Danacıoğlu | Hüseyin Can Öngül

Publisher:

Ege University

Main Topics:

Uzaktan Algılama | Beşeri Coğrafya

Journal:

Ege Coğrafya Dergisi Vol.32 Issue 1, pp.135-150

Document Type:

Article

Publication Date:

6/2023

Document Language:

Turkish

Document ID (Org No.):

2001295206

pISSN:

2636-8056

Google Earth Engine

Remote Sensing

Random Forest

Land Use and Land Cover Change

Güzelhisar BasinGüzehisar Havzası

Uzaktan Algılama

Google Earth Engine

Arazi Kullanımı ve Arazi Örtüsü

Makine Öğrenmesi

Issues

Article Details

Authors: Şevki Danacıoğlu | Hüseyin Can Öngül

Publisher: Ege University

Main Topics: Uzaktan Algılama | Beşeri Coğrafya

Journal: Ege Coğrafya Dergisi Vol.32 Issue 1, pp.135-150

Document Type: Article

Publication Date: 6/2023

Document Language: Turkish

Document ID (Org No.): 2001295206

pISSN: 2636-8056

Issues

Authors:

Şevki Danacıoğlu | Hüseyin Can Öngül

Publisher:

Ege University

Main Topics:

Uzaktan Algılama | Beşeri Coğrafya

Journal:

Ege Coğrafya Dergisi Vol.32 Issue 1, pp.135-150

Document Type:

Article

Publication Date:

6/2023

Document Language:

Turkish

Document ID (Org No.):

2001295206

pISSN:

2636-8056