ABSTRACT

Weather forecasting has become an indispensable application to predict the state of the atmosphere for a future time based on cloud cover identification. But it generally needs the experience of a well-trained meteorologist. In this paper, a novel method is proposed for automatic cloud cover estimation, typical to Indian Territory Speeded Up Robust Feature Transform(SURF) is applied on the satellite images to obtain the affine corrected images. The extracted cloud regions from the affine corrected images based on Otsu threshold are superimposed on the artistic grids representing latitude and longitude over India. The segmented cloud and grid composition drive a look up table mechanism to identify the cloud cover regions. Owing to its simplicity, the proposed method processes the test images faster and provides accurate segmentation for cloud cover regions.

Agriculture industries play an important role in economic development. Many post harvest techniques are performed in these industries. Especially in fruit industries, grading is an important process. As per the grading standard, banana hands must have minimum of twelve banana fingers. In the current scenario, the banana fingers are counted manually which is laborious. In order to reduce manual work, this paper proposes an algorithm for performing an automatic counting system of banana fingers in a bunch/hand using image processing algorithms. The parameters of the Gaussian model is acquired from the samples of fruit distal ends. A multivariate Gaussian modeling is accomplished on HSV color model to detect the distal end of banana fingers using maximum likelihood detection.

Video transition detection (VTD) is a significant topic in the field of video analytics, owing to its useful applications in video indexing, video surveillance, and video understanding. The challenge in VTD is to extract the complex temporal variations caused by the change in illumination, rapid motion of object and camera. To address these challenges, a pyramidal-relative en-tropy based long-short term memory (LSTM) framework is proposed. Initially, the uncompressed video frame is modeled using pyramidal attributes. Then, the temporal signature is generated based on forward-backward ratio of relative entropy measure. The LSTM network is trained to detect the transitions through temporal signature and categorizes them as no transition, abrupt transition and gradual transition. Benchmarks namely TRECVID 2001, TRECVID 2007, VIDEO-SEG2004, RAI and BBC data-sets have been used in evaluation assessments. The simulation results illustrate the efficacy of the proposed framework to achieve an average F 1 score of 95.5±0.06%, 98.51±0.79%, 97.5 ±2.6%, 93.54±5.03%, and 97.58±1.89% on TRECVID 2001, TRECVID 2007 VIDEOSEG 2004, RAI and BBC data-sets, respectively.