Introduction:
Encroachment of water bodies is caused by urbanization thus leading to depletion of natural resources thereby causing scarcity and people suffering from lack of food security. The project aims to solve encroachment problems thereby solving urban planning problems using machine learning and remote sensing.
Data collection
Data were collected from Landsat 7 (2000-2014) and Landsat 8 (2015-2022) Surface Reflectance products (Level 2) available on Google Earth Engine for the years (2000-2022). with a spatial resolution of 30 m.
Data preprocessing includes scan line correction that can result in 22% data loss using Focal Averaging and data blending. The figure below shows a comparison of the scan line error present and removed from the Landsat 7 data.
Figure 1. Composite showing with and without scan line error
Clouds are removed using the CF scan algorithm and compositing as Annual Mean from Landsat 7 and Landsat 8.
Water body collection and classification
A machine learning algorithm such as Random Forest was applied to automate the extraction of water bodies, and the classification shows the inaccuracies of the data and Expert System based index approach, which may be more suitable for the extraction of water bodies.
Water-based indices are used, such as the Modified Normalized Difference Water Index (MNDWI), which is a normalization of the Green(G) and Shortwave Infrared (SWIR) bands for the period 2000-2022. MNDWI = (G-SWIR)/(G+SWIR)
Threshold-based classification is performed to extract water bodies such as MNDWI > 0 and MNDWI<0 for water and non-water bodies. Water body trends are then analyzed.
Figure 2: General methodology
Conclusion
Fig. 3. Total areas of lakes in 2000-2022
Chennai lakes started disappearing from 2000 to 2018 and thus showing validation of encroachment over the years is given in the form of statistical analysis in the form of tables and graphs.
Fig 4: Pulicat Lake encroachment
Fig. 5: Encroachment of Puzhal Lake
Fig. 6. Lake Reter encroachment
Fig. 7: Encroachment of Lake Koratur
Fig. 8: Encroachment of Lake Ambatur
Fig 9: Encroachment of Chembarambakam Lake
Figure 10: Mutukadu (Great Salt) lake encroachment
Figure 11: Encroachment of Madurantakam Lake
Fig. 12: Lake Pundi encroachment
It is the northern part of the lakes for Pulikat, Retteri, Madurantakam encroachment at a faster rate and for lakes in southern part of Pujali, Reteri, Ambathur, Chembarambakkam, Pundi are also encroached at a faster rate.
Learn more through: https://omdena.com/chapter-challenges/automatic-lake-encroachment-detection-in-chennai/
Chapter link in the right column. https://omdena.com/local-chapters/ranchi-india-chapter/