Optimization of Parallel K-means for Java Paddy Mapping Using Time-series Satellite Imagery (original) (raw)
Related papers
Optimization of Parallel K-means for Java Paddy Mapping Using Time-series Satelite Imagery
TELKOMNIKA (Telecommunication Computing Electronics and Control)
Spatiotemporal analysis of MODIS Vegetation Index Imagery widely used for vegetation seasonal mapping b oth on forest and agricultural site. In order to provide a long-terms of vegetation characteristic maps, a wide time-series images analysis is needed which require high-performance computer and also consumes a lot of energy resources. Meanwhile, for agriculture monitoring purpose in Indonesia, that analysis has to b e employed gradually an d endlessly to provide the latest condition of paddy field vegetation information. This research is aimed to develop a method to produce the optimized solution in classifying vegetation of paddy fields that diverse b oth spatial and temporal characteristics. The time-series EVI data from MODIS have b een filtered using wavelet transform to reduce noise that caused b y cloud. Sequential K-means and Parallel K-means unsupervised classification method were used in b oth CPU and GPU to find the efficient and the rob ust result. The developed method has b een tested and implemented using the sample case of paddy fields in Java Island. The b est system which can accommodate of the extend-ab ility, affordab ility, redundancy, energy-saving, maintainab ility indicators are ARM-b ased processor (Raspb erry Pi), with the highest speed up of 8 and the efficiency of 60% .
Spatial change analysis of paddy cropping pattern using MODIS time series imagery in Central Java
IOP Conference Series: Earth and Environmental Science, 2017
Central Java had the diverse paddy field cropping patterns and it was influenced by several factors such as water availability, land condition, paddy fields ownership, and local culture. This research was aimed to analyze dynamic changes of paddy cropping pattern using MODIS imagery (MOD13Q1 16-day composite from 2001 to 2015). This research used kmeans clustering algorithm for classified cropping pattern in Central Java based on similarity pattern of annual data from vegetation index. The result of this research classified cropping pattern become a main class and produced 15 maps of distribution cropping patterns (from 2001 to 2015). The result also divided Central Java's paddy fields become 2 section (constant and change) based on cropping pattern that majority was caused by water availability. This research got the better accuracy (77.67%) of cropping pattern than long time series analysis from previous research. Although some classes successfully obtained upon annual time series analysis, MODIS still difficult to detect mixed crop pattern. Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Classification and Mapping of Paddy Rice by Combining Landsat and SAR Time Series Data
Remote Sensing
Rice is an important food resource, and the demand for rice has increased as population has expanded. Therefore, accurate paddy rice classification and monitoring are necessary to identify and forecast rice production. Satellite data have been often used to produce paddy rice maps with more frequent update cycle (e.g., every year) than field surveys. Many satellite data, including both optical and SAR sensor data (e.g., Landsat, MODIS, and ALOS PALSAR), have been employed to classify paddy rice. In the present study, time series data from Landsat, RADARSAT-1, and ALOS PALSAR satellite sensors were synergistically used to classify paddy rice through machine learning approaches over two different climate regions (sites A and B). Six schemes considering the composition of various combinations of input data by sensor and collection date were evaluated. Scheme 6 that fused optical and SAR sensor time series data at the decision level yielded the highest accuracy (98.67% for site A and 93.87% for site B). Performance of paddy rice classification was better in site A than site B, which consists of heterogeneous land cover and has low data availability due to a high cloud cover rate. This study also proposed Paddy Rice Mapping Index (PMI) considering spectral and phenological characteristics of paddy rice. PMI represented well the spatial distribution of paddy rice in both regions. Google Earth Engine was adopted to produce paddy rice maps over larger areas using the proposed PMI-based approach.
Procedia Environmental Sciences, 2016
Accurate and up-to-date information of paddy fields over wide areas is essential to support sustainable agricultural and a food security program. It is an urgent need to develop near-real time paddy field monitoring, which can be used by policy maker for handling the food problems directly. This study explored the use of multi temporal MODIS EVI 16-day composite data, which provided the seasonal dynamics for the paddy field patterns from 2000 to 2014. We characterized seasonal vegetation dynamics from MODIS satellite datasets in order to analyze the dynamics change in paddy field. The results indicate that the methodology employed in this research distinguished many specific uses in paddy fields as means of their cropping intensity. Moreover, the seasons were the most important factor affected the dynamics change in the agricultural system. Indeed, characterizing the longterm vegetation dynamics provides information about the characteristic and trends in paddy field area, either caused by natural factors or human activities, also to be a guidance of water resources management due to improving its effectiveness.
CIGR-AgEng 2012
The classification of land use in agriculture usually requires high resolution satellite, therefore the cost tends to be expensive and that is the barrier of remote sensing use in agriculture. The author Utilize the time frequency analysis to the NDVI data series observed by Moderate Resolution Imaging Spectroradiometer (MODIS), and then presents an automatic and cost free method for the classification of paddy fields in this paper. This research is conducted in Mekong Delta, in the southern part of Vietnam. 10-days composite images (250m spatial resolution) from the MODIS sensor onboard the NASA EOS Aqua and Terra satellite were used. The wavelet analysis is used as the method for time frequency analysis. The calculated wavelet powers of the waves of NDVI are used to get the characteristics of paddy area which corresponding to the rice cropping calendar. For example in the level -3 at 80 days period, the wavelet powers of paddy area become very high in contrast with other area because the rice cropping calendar is approximately 90 days. After getting features of paddy field Linear Discriminant Analysis (LDA) is assess for automatic land use classification using statistical NDVI data series and calculated wavelet powers. The result shows that the combination of wavelet power and statistic of NDVI data for discriminant analysis work well for the classification even a low resolution satellite images were used. Furthermore by using the wavelet analysis it is possible to distinguish whether the paddy field is single or multiple cropping by utilizing the value of wavelet powers in time.
Application of Wavelet Transform for Paddy Area Classification Using MODIS NDVI Data Series
2013
An algorithm for the classification of paddy area was proposed by utilizing wavelet transform as time frequency analysis of Normalized Difference Vegetation Index (NDVI) data series derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. The calculated wavelet powers of the NDVI wave correspond to the rice cropping calendar, and were used to obtain the characteristics of paddy area. After determining features of the paddy area, Linear Discriminant Analysis (LDA) is assessed for land-use classification using statistical values of NDVI data series and calculated wavelet powers. This research was conducted in the Mekong Delta, southern Vietnam. Three years, NDVI data series (January 2009 up to December 2011) of 10-day composite images with 250 m spatial resolution were applied. The result shows that the combination of wavelet powers and statistical values of NDVI data for LDA worked well for the classification, even when low resolution satellite images like MODIS were used. Furthermore, paddy classification can obtain the most appropriate result if the starting point of the calculation is adjusted to the rice planting period. However, verification of accuracy has not yet been done due to lack of the latest land-use data of the study area, and still remains as a future subject.
Analysis of Paddy Productivity Using NDVI and K-means Clustering in Cibarusah Jaya, Bekasi Regency
IOP Conference Series: Materials Science and Engineering, 2019
Information about rice productivity is one of the references for government to maintain food availability. With remote sensing technology, rice productivity can be known faster. This research was conducted using UAV (Unmanned Aerial Vehicle) and Sentinel-2 Satellite. Sentinel-2 NDVI which has a low resolution with high resolution UAV images, both variables have similarity values and regression reaches 0.8. NDVI are grouped into 8 classes using kmeans clustering based on the similarity of the waveforms of each data retrieval point. Based on characteristic of k-means classes, field which has earlier planting times and the location closer to the water source, allowing a higher paddy productivity. Further analysis was also carried out to get the best period to estimate paddy productivity using Sentinel-2 imagery. Sentinel-2 was chosen because it has a distance between data as far as 5 days, allowing it to be more accurate. The best time is obtained at 63 DAP (Days After Planting), which is when NDVI reaches its maximum state. The estimation model of rice productivity based on UAV has a high coefficient of determination compared to Sentinel-2 so that the relationship between maximum NDVI UAV and rice productivity is better than Sentinel-2.
Segmentation and classification of high resolution satellite imagery is a challenging problem due to the fact that it is no longer meaningful to carry out this task on a pixel-by-pixel basis. The fine spatial resolution implies that each object is an aggregation of a number of pixels in close spatial proximity, and accurate classification requires that this aspect be subtly considered. K-means clustering algorithm is a better method of classifying high resolution satellite imagery. The extracted regions are classified using a minimum distance decision rule. Several regions are selected as training samples for region classification. Each region is compared to the training samples and is assigned to its closest class. The procedure significantly reduces the mixed pixel problem suffered by most pixel based methods. In this paper, we used K-means clustering algorithm to classify satellite imagery into specific objects within it for cadastral and environmental planning purposes, thereby ...
Efficient parallel algorithm for pixel classification in remote sensing imagery
Ujjwal Maulik, Anasua Sarkar
An important approach for image classification is the clustering of pixels in the spectral domain. Fast detection of different land cover regions or clusters of arbitrarily varying shapes and sizes in satellite images presents a challenging task. In this article, an efficient scalable parallel clustering technique of multi-spectral remote sensing imagery using a recently developed point symmetry-based distance norm is proposed. The proposed distributed computing time efficient point symmetry based K-Means technique is able to correctly identify presence of overlapping clusters of any arbitrary shape and size, whether they are intra-symmetrical or inter-symmetrical in nature. A Kd-tree based approximate nearest neighbor searching technique is used as a speedup strategy for computing the point symmetry based distance. Superiority of this new parallel implementation with the novel two-phase speedup strategy over existing parallel K-Means clustering algorithm, is demonstrated both quantitatively and in computing time, on two SPOT and Indian Remote Sensing satellite images, as even K-Means algorithm fails to detect the symmetry in clusters. Different land cover regions, classified by the algorithms for both images, are also compared with the available ground truth information. The statistical analysis is also performed to establish its significance to classify both satellite images and numeric remote sensing data sets, described in terms of feature vectors.
Agriculture data analysis using parallel k-nearest neighbour classification algorithm
A cost-effective and effective agriculture management system is created by utilizing data analytics (DA), internet of things (IoT), and cloud computing (CC). Geographic information system (GIS) technology and remote sensing predictions give users and stakeholders access to a variety of sensory data, including rainfall patterns and weather-related information (such as pressure, humidity, and temperatures). They have unstructured format for sensory data. The current systems do a poor job of analysing such data since they cannot effectively balance speed and memory usage. An effective categorization model (ECM) on agriculture management system is proposed to address this research difficulty. First, a classification technique called priority-based k-nearest neighbour (KNN) is provided to categorize unstructured multi-dimensional data into a structured form. Additionally, the Hadoop MapReduce (HMR) framework is used to do classification utilizing a parallel approach. Data from real-time IoT sensors used in agriculture is the subject of experiments. The suggested approach significantly outperforms previous approaches that are computing time, memory efficiency, model accuracy, and speedup