Klasifikasi Mangrove Berbasis Objek dan Piksel Menggunakan Citra Satelit Multispektral di Sungai Kembung, Bengkalis, Provinsi Riau
Siregar, Vincentius P
Prasetyo, Lilik Budi
MetadataShow full item record
Mangrove that generally found in coastal regions plays important roles such as (1) coastal protection from hurricane, tsunami, wind and wave, (2) spawning ground for many fishes and other faunas, (3) a place for recreational, (4) source of nutrients for organisms, and (5) source of wood. Recently, mangrove ecosystems in Kembung River, Bengkalis Island Riau Province tend to get pressure from antrophogenic. Protection and management of mangrove ecosystems needs to be a serious concern of various parties. Such efforts require data acquisition techniques and information mangrove spatially and accurately. Field data were collected in June and December 2012. Satellite imageries were used in this study consisted of Landsat 5 TM, Landsat 7 ETM +, Landsat 8 OLI, SPOT 6 multispectral, SPOT 6 panchromatic, ALOS PALSAR and SENTINEL-1. Pre-processing in the satellite imageries were applied including atmospheric correction, radiometric calibration, geometric correction, and spectral transformation. Field data observations and measurements were conducted on mangrove vegetation and land cover. Scheme development of mangrove community classification was conducted by using clusters and similarity percentage (SIMPER) analyses. Furthermore, the scheme was used to charactirized satellite images spectral reflectance by using spectral analysis. Satellite data were classified by using object-based approach which applied random forest algorithm. Object-based classification results was then compared to pixel-based classification technique which used maximum likelihood algorithm. The succesfull object-based classification then applied on magrove change detection in the study area. Kembung River mangrove community was assembled by 69 mangroves species which consists of 22 true mangrove species and 47 species of associate mangrove. Xylocarpus granamun, Rhizophora. apiculata, Lumnitzera. racemosa and Scyphiphora hidrophyllaceae were dominant species and always found in the study area compared with other species, for all strata. Based on standard criterias and guidelines for the determination of mangrove destruction which issued by the Minister of the Environment No. 201 of 2004, Mangrove at Kembung River was categorized as good and very dense mangrove ecosystem. Based on the composition of mangrove species, one level classification scheme was succesfully developed. The scheme consisted of 12 mangrove community classes, however most of the mangrove species belong to the two main classes of Rhizophora apiculata (Ra) composed of 33 samples and Xylocarpus granatum (Xg) composed of 93 samples Other 10 classes had limited sample number, therefore the 10 other classes were groupped into one class (La) that composed of 28 samples. Both classification techniques were able to identify all land cover classes and mangrove communities. Some misclassifications were still found to produce salt and pepper effects on both classifications, however, the random forest algorithm v could reduced such errors than the maximum likelihood algorithm. The best result of land cover classification using object-based and pixel-based classifications was obtained through input image layer M6. The object-based classification approach was better than pixel-based and it can improved the classification results by 1-24.5%. Change detection analyses showed that the mangrove area in Kembung River was relatively stable. For nearly two decades, we found mangrove loss about 197.2 ha, gain of 251.1 ha, and unchanged of 2904.9 ha. Changes in mangrove covers were generally caused by anthropogenic factors such as mangrove replanting, logging, changes over the function of mangrove regions into the road, embankment, settlement, shrimp farms, and natural growth. Serious attention from various parties are needed to maintain the existence and sustainablility of mangrove ecosystems in Kembung River.
- DT - Fisheries