A Context-Sensitive Clustering Technique Based on Graph-Cut Initialization and Expectation-Maximization Algorithm

This letter presents a multistage clustering technique for unsupervised classification that is based on the following: 1) a graph-cut procedure to produce initial segments that are made up of pixels with similar spatial and spectral properties; 2) a fuzzy c-means algorithm to group these segments into a fixed number of classes; 3) a proper implementation of the expectation-maximization (EM) algorithm to estimate the statistical parameters of classes on the basis of the initial seeds that are achieved at convergence by the fuzzy c-means algorithm; and 4) the Bayes rule for minimum error to perform the final classification on the basis of the distributions that are estimated with the EM algorithm. Experimental results confirm the effectiveness of the proposed technique.     

Ozone in ambient air at a tropical megacity, Delhi: characteristics, trends and cumulative ozone exposure indices

Seven year data of hourly surface ozone concentration is analyzed to study diurnal cycle, trends, excess of ozone levels above threshold value and cumulative ozone exposure indices at a tropical megacity, Delhi. The ozone levels clearly exhibit a diurnal cycle, similar to what has been found in other urban places. A sharp increase in the ozone levels during forenoon and a sharp decrease in the early afternoon can be observed. The average rate of increase in ozone concentration between 09 and 12 h has been observed to be 7.1 ppb h⁻¹. We find that the daily maximum and daytime 8-h (10–17 h) ozone levels are increasing at a rate of about 1.7 (± 0.7) and 1.3 (± 0.56) ppb y⁻¹, respectively. The directives on ozone pollution in ambient air provided by United Nations Economic Commission for Europe and World Health Organization for vegetation (AOT40) and human health protection were used to assess the air quality. The present surface ozone levels in the city are high enough to exceed “Critical Levels” which are considered to be safe for human health, vegetation and forest. The human health threshold was exceeded for up to ~45 days per year. The AOT40 (Accumulated exposure Over a Threshold of 40 ppb) threshold was exceeded significantly during winter (D-J-F) and pre-monsoon (M-A-M) (Rabi crop growing season) season in India. Translating AOT40 exceedances during pre-monsoon into relative yield loss we estimate yield loss of 22.7%, 22.5%, 16.3% and 5.5% for wheat, cotton, soybean and rice, respectively.     

Surface waves on a magnetic cylinder

The problem of surface wave propagation on a magnetised cylindrical plasma is investigated allowing for different axial magnetic fields inside and outside the flux tube. Properties of surface waves for symmetric and asymmetric modes of perturbations are investigated idealising the material inside and outside the cylinder to be insulator or infinitely conducting both in compressible and incompressible approximations for the tube material.     

Evaluating the impact of climate change in threshold values of thermodynamic indices during pre-monsoon thunderstorm season over Eastern India

Natural Hazards - The present study analyses thermodynamic indices variation over three sites of eastern Indian region: Bhubaneswar, Kolkata and Ranchi, associated with pre-monsoon thunderstorms...     

Lapse time and frequency-dependent attenuation characteristics of coda waves in the Northwestern Himalayas

We analyzed the local earthquakes waveform recorded on a broadband seismic network in the northwestern Himalayan Region to compute lapse time and frequency dependence of coda Q (Q _c). The observed Q _c values increase with increasing lapse time at all frequency bands. The increase in Q _c values with lapse time is attributed to an increase in Q _c with depth. This implies that attenuation decreases with increasing depth. The approximate radius of medium contributing to coda generation varies from 55 to 130 km. By comparing the Q _c values with those from other regions of the world, we find that they are similar to those obtained from tectonically active regions. The estimated Q _c values show a frequency-dependent relationship, Q _c = Q ₀ f ⁿ , where Q ₀ is Q _c at 1 Hz and n represents degree of frequency dependence. They represent the level of heterogeneity and tectonic activity in an area. Our results show that northwest Himalayas are highly heterogeneous and tectonically very active. Q ₀ increases from 113 ± 7 to 243 ± 10 and n decreases from 1.01 ± 0.05 to 0.85 ± 0.03 when lapse time increases from 30 to 70 s. As larger time window sees the effect of deeper part of the Earth, it is concluded that Q ₀ increases and n decreases with increasing depth; i.e., heterogeneity decreases with depth in the study area.