Ganesh Kumar Sethi’s research while affiliated with Punjabi University and other places

Agriculture is crucial for economic growth, rural development, and food security. Remote sensing aids in cost-effective agricultural mapping, but challenges like limited resolution, atmospheric errors, and cloud interference in satellite imagery hinder accurate monitoring. To overcome these challenges, this study introduces an image fusion-based framework for detecting agricultural changes with the incorporation of optical and microwave satellite data. It integrates the strengths of multi-source sensors to provide enhanced accuracy in classification and change detection procedures, especially in the detection of agricultural variation. The novelty of the work lies within the incorporation of a fusion process in the post-classification-based change detection method and the newly developed framework named fusion-based post-classification change detection (FPCD). To perform the in-depth analysis of FPCD, various fusion methods i.e., (a) Gram-Schmidt (GS) and (b) PC spectral (PCS), and various machine learning-based classification methods i.e., (a) maximum likelihood classifier (MLC), (b) minimum distance classifier (MDC), and (c) support vector machine (SVM) was utilized. Imagery from the Sentinel-1 with VH and VV polarization bands and Sentinel-2 L2A (Level 2A) satellites were acquired over a region in Punjab, India, known for its fertile soil and significant contribution to wheat production. The experimental outcomes confirmed the effectiveness of the proposed FPCD framework by more than overall accuracy (96.5% and 92.74% in classified and change maps, respectively) as compared to other existing frameworks i.e., SVM-based (94.7% and 88.29% in classified and change maps, respectively) and MDC (88% and 81.19% in classified and change maps, respectively). These outcomes are satisfactory enough to monitor multitemporal agricultural variations at a large scale in an effective manner.

Over the past few years, explainable artificial intelligence (XAI) has become increasingly popular as a result of the demand for AI systems that are simpler to comprehend and with greater interpretability. This study provides a conceptual framework and a quick assessment of the work done in explainable artificial intelligence. Using the Vosviewer application, the researchers analyzed 4781 research publications from the Scopus database, spanning 2004 to 2023. Observations indicate a rapid and exponential growth in the quantity of publications, commencing in 2018. The importance of the study is shown by the analysis of publishing activities according to the year of publication and the geographical area, together with citation analysis, research methodologies, and data analysis techniques. The researchers have highlighted ten interesting areas that require further study from future researchers. Moreover, the work emphasizes the legal, ethical, and social consequences for the researchers.

Image fusion could adopt the optical and microwave images model to obtain more precise agricultural land cover information features by using the optical and microwave remote sensing sensors characteristics. The source of data has decided to take into account Sentinel-2 L2A and Sentinel-1 VH (Vertical-transmit and Horizontal-receive) datasets for testing image fusion classifiers used in the experiments. Different kinds of classifiers were used in this research. Among them, MDC stands for Minimum Distance Classifier, and Maximum Likelihood Classifier represents MLC. Lastly, a Support Vector Machine (SVM) was used during the fusion of the images to classify Land Use and determine the validity of these algorithms. A summary of the outputs will give an upper hand to the classifiers based on Support Vector Machines, with the accuracy reaching 95.67% and a kappa of 0.9191. The SVM classification algorithm during the study showed the dense vegetation, deciduous vegetation, water, and urban zones with shallow errors. MDC performed the task well in detecting the land use category with an accuracy of 91.70% and kappa of 0.8469. The opposite of the MDC classifier was that the accuracy was not the top level, as, in some ways, some water bodies were wrongly assigned as urban regions. MLC has the least promising result among other approaches and is in the last place in terms of accuracy, with 89.52% and a kappa value of 0.8281.

The Western Himalayas receive significant snowfall, and it's essential to monitor the snow cover for various purposes like managing water resources, conducting hydrological research, and predicting avalanches using remote sensing techniques. With the pan-sharpening of different satellite sensor images, the essential features can be obtained to understand the snow cover dynamics. The resampling algorithms play a very important role in the pan-sharpening procedure to highlight the important features of the pan-sharpened dataset. However, the performance of various resampling algorithms for pan-sharpening is very rarely validated with optical and microwave datasets. In this article, the spatial and spectral analysis of different resampling algorithms, i.e., Nearest Neighbour (NN), Bilinear (BI), and Cubic Convolution (CC) resampling, have been performed using the fusion of optical and microwave satellite images. In this study, two datasets, i.e., MODIS (optical) and SCATSAT-1 (microwave) were used over the western Himalayas (i.e., Ladakh, Jammu and Kashmir, Himachal Pradesh, and Uttrakhand). To evaluate the performance of each resampled pan-sharpened dataset, the output is classified using a Support Vector Machine (SVM) classifier and a Spectral Angle Mapper (SAM) classifier. Root Mean Square Error values were computed to quantify the level of agreement between the pan-sharpened datasets and the ground truth data. The result of statistical analysis showed that NN-based pan-sharpened performed better than BI-based and CC-based pan-sharpened classified images with both classifiers i.e., SVM and SAM. This study is important in terms of the effective utilization of the resampling techniques along with pan-sharpening algorithms.

Rice is an essential staple food for human nutrition. Rice varieties worldwide have been planted, imported, and exported. During production and trading, different types of rice can be mixed. Due to rice impurities, rice importers and exporters may lose trust in each other, requiring the development of a rice variety identification system. India is a significant player in the global rice market, and this extensive study delves into the importance of rice there. The study uses state-of-the-art deep learning and TL classifiers to tackle the problems of rice variety detection. An enormous dataset consisting of more than 600,000 rice photographs divided into 22 different classes is presented in the study to improve classification accuracy. With a training accuracy of 96% and a testing accuracy of 80.5%, ResNet50 stands well among other deep learning models compared by the authors. These models include CNN, Deep CNN, AlexNet2, Xception, Inception V3, DenseNet121, and ResNet50. Finding the best classifiers to identify varieties accurately is crucial, and this work highlights their possible uses in rice seed production. This paper lays the groundwork for future research on image-based rice categorization by suggesting areas for development and investigating ensemble strategies to improve performance.

The biggest contribution to the Indian economy is made through agriculture. It is one of the main sources of income for farmers. The major concern for the farmers during cultivation is to produce crops of good quality and quantity. For a wide range of productivity and protection from various diseases, several techniques must be revised. Machine learning techniques were utilized for the categorization of leaf diseases, because earlier methods were time-consuming and labor-intensive. Better results are achieved using machine learning techniques, along with certain limitations. Deep Convolutional Neural networks are now one of the finest methods for categorizing leaf diseases in crops. There are still issues with getting access to trustworthy models for identifying a variety of diseases and segmenting lesion areas for assessing severity in real-world field conditions, despite the growing popularity of deep learning approaches for detecting various diseases. Under Deep Convolutional Neural Networks, various architectures perform well for image classification. The quantitative evaluation of features that increase a plant’s ability to withstand disease is crucial in the selection of plant breeders. As a result, it is vital to take advantage of the disease-affected regions’ ability to determine how badly diseased the leaves are. To address these issues, CNN-based segmentation model was proposed to separate the wheat leaf diseases from the wheat leaf image dataset at the pixel level. This is superior to the segmentation algorithms currently in use. Our ultimate goal is to assist farmers in detecting and learning about early-stage illnesses in wheat leaves. To classify the disorders, a Deep Convolutional Neural Network was used. The investigation made use of a dataset including 4000 images of wheat leaves infected with three different types of leaf diseases: powdery mildew, leaf rust, and spot blotch. This manuscript includes the proposed method for feature extraction using the Point Rend deep segmentation model, followed by classification through the EfficientNet architecture. The results reveal that the proposed model is more accurate, showing a classification accuracy of 99.43% in comparison with classification performed on wheat leaf diseases without employing segmentation networks.