Search

In the current scenario, plant disease detection is seeking attention from many agricultural scientists. Plant diseases are deeply influenced by the weather conditions, and each disease has its individual weather requirements. The changes in weather parameters such as humidity, temperature, wind speed, etc., can cause many diseases in tomato plants. In the current empirical study, we have taken specific disease powdery mildew whose fungus is named as Leveillula Taurica which belongs to Leotiomycetes class, and it is responsible for the occurrence of this specific disease in tomatoes. In this research, three weather-based prediction models have been developed using k-nearest neighbor (kNN), decision tree (DT), and random forest (RF) algorithm for powdery mildew disease prediction in tomatoes at an early stage. Results indicate that the proposed model, based on RF algorithm, shows the best accuracy of 93.24% for tomato powdery mildew disease (TPMD) dataset. A real-time version of the proposed model can be used by the agricultural experts to take preventive measures in the most sensitive areas that are prone to powdery mildew disease based on the weather conditions. Hence, timely intervention would help in reducing the loss in productivity of tomato crops which will further benefit the global economy, agricultural production, and the food industry.

In general, AMBTC-based steganography techniques effectively utilize smooth blocks for secret data embedding but keeps complex blocks as it is. It is considered that any modification in the complex blocks may cause a substantiate distortion in an original image. Therefore, conventional techniques do not make any modification in complex blocks. In proposed paper, a technique is disclosed for secret data embedding in complex blocks in addition to smooth blocks. It shows that a cluster of bit-map blocks corresponding to complex blocks of multiple images are formed, and then most frequent 256 bit-map blocks are determined based on histogram analysis. Multiple images can be drawn from any standard image database. The most frequent 256 bit-map are block patterns which are indexed using 8 bits. Now, for secret data embedding, bit-map of each complex block is scanned one-by-one and is compared with the 256 bit-map block patterns. If the bit-map of scanned complex block is matched with any of the 256 block patterns, then first 8 bits of bit-map of scanned complex block is replaced with 8-bit index of matched block pattern and remaining 8 bits of bit-map of scanned complex block are replaced with secret data. Thus, this technique shows a way of embedding secret data in complex blocks using block patterns. Further, for the smooth blocks, present technique utilizes Ou et al. technique which is efficient for secret data embedding in smooth blocks. Results of the proposed scheme show increase in embedding capacity while preserving image quality.

The present study is concerned with the analysis to determine the working range of process parameters in friction stir processing (FSP). Commercial AA6082-T6 rolled plates were subjected to FSP with various process parameters such as tool rotational speed, processing speed, and tilt angle. The experiments were conducted to study the influences of different levels of process parameters on the surface appearance, surface defects, heat generation, and flash formation. The study shows that the operating range of process parameters can be estimated within two critical conditions, i.e., excessive heat and insufficient heat generation. Beyond these two critical conditions, material prone to defect formation during FSP. The tilting of the tool produces lower flash formation and evenly distributes it on the advancing side and retreating side compared to the 0° tilt angle.

Deep Learning is acquiring momentum in the agricultural field for crop disease detection using image processing due to its computational power. Several deep learning techniques have been implemented in different domains and recently introduced in the field of agriculture to classify and predict the diseases of crops. Based on images of banana crops in the early stages of development, the objective of this research study is to create a prediction model using two types of Convolutional Neural Networks (CNN) architectures, namely, AlexNet and ResNet50. In order to carry out the empirical study, the PlantVillage dataset for the Banana plant with 510 images of banana leaves was used to train and test the networks. Results were analyzed using four parameters namely; training accuracy (TA), training loss (TL), validation accuracy (VA), and validation loss (VL). It was observed that ResNet50 outperformed the other one with better results at 88.54% when validation accuracy is considered as a performance evaluation measure. The results of this study will be useful for farmers as they can make timely interventions in the case of Banana Black Sigatoka (BBS) and Banana Bacterial Wilt (BBW) diseases.

Fixture design is the most critical aspect for the success of friction stir welding and processing (FSW/P). Nowadays, the machines that developed exclusively dedicated to FSW/P are very expensive, which motivated to design a new innovative fixture. This paper aims to design and develop a robust fixture that can be easy to use on a conventional vertical milling machine. The designed fixture and clamping devices rigidly hold the workpiece in such a way that it restricts lateral and longitudinal movement during processing. The later part carried out a detailed finite element analysis to check the robustness of fixture using Autodesk Fusion 360 software (Education license) and validated it in the actual condition. The fixture setup is installed and tested successfully on the universal vertical milling machine. The result shows that it withstands high forces generated due to relative motion between the tool and the workpiece.

Seeds and propagating materials of plants are the primary source of pathogen inoculum to cause diseases. These materials also transmit the pathogen from one place to other places and establish the disease in a new area. Hence, it is an utmost requirement to detect the pathogen of a particular crop before transportation and sowing to ensure that no potentially damaging pathogens are introduced in the field through seeds and planting materials. This can be most effectively accomplished by keeping out pathogens from seed lots by either discarding or treating seeds with chemicals. Various conventional methods for the detection of pathogens such as visual examination, selective growth media, serological methods, and bioassay have been used commonly. But these methods have disadvantages like inefficiency, less specificity, less sensitivity, and more time-consuming. Now-a-days, polymerase chain reaction (PCR) has more potential to improve bacterial pathogen detection in seeds as well as planting materials. There are advanced techniques like BIO-PCR, immunomagnetic separation-PCR (IMS-PCR), and magnetic capture hybridization-PCR (MCH-PCR) which reduce inhibitory compounds during PCR, which further improve the detection level of bacterial pathogens from seeds and planting materials. IMS-PCR and MCH-PCR are more attractive due to their simple and universally applicable methods to test seeds for different culturable and non-culturable bacterial pathogens. However, it is difficult to adapt their applicability for routine testing of seed under the laboratory. It should be ensured that these methods should work and these methods must be validated in multi-laboratory tests thoroughly and these tests should be reproducible and repeatable before their commercialization.

Pulmonary infections have always been a cause of high morbidity and mortality, particularly in the pediatric and geriatric population and in immunocompromised hosts [1]. Pulmonary infections have various etiologies and have variegated patterns on radiographs and computed tomography (CT). Imaging plays an important role in the initial diagnosis and follow-up of various lung infections. Radiographs can be normal or non-specific during the initial evaluation, and CT findings may be more definitive. CT not only helps with the diagnosis but can also aid in management by guiding the diagnostic and therapeutic procedure. The pulmonary infections spread by direct or indirect contact with the infected host, droplet transmission, or an airborne spread. In rare cases, some infections can also be transmitted by vectors, namely, insect or animal hosts, and rarely by direct invasion from nearby infected organs. Pulmonary infections may have typical imaging patterns and distribution based on the mode of spread. There are a number of well-described imaging patterns of alveolar infections. The localization and morphological features on imaging may help in the diagnosis of infection and identification of mode of infection and, in certain cases, the microorganism responsible for the infection.

Chest wall and pleural pathologies have always been a diagnostic challenge for the clinicians and the radiologists. These pathologies range from various congenital chest wall deformities, inflammatory or infectious lesions, as well as benign and malignant tumors. Radiographs are often the first-line imaging modality and are useful in the follow-up assessment. Ultrasound is helpful in the assessment of soft-tissue lumps involving the chest wall and pleural pathologies. Its role is much more important in guiding drainage procedures. Cross-sectional imaging techniques, namely, computed tomography (CT) and magnetic resonance imaging (MRI), are used in the localization and extent assessment of the chest wall and pleural abnormalities.

Vehicular Ad Hoc network (VANET) is a promising technology mainly used to increase the safety of the vehicles, passengers, and etc. on the road. The safety and convenience services of VANET are supported through different type of messages. The high speed of vehicles results in the short time period available to exchange messages between the vehicles. All sorts of messages are broadcasted by the peer vehicles in the network, and those messages may causes channel overload, when vehicle density increases on the roads. In consequence of these, our channel gets highly congested which results in packet loss. The packet loss inside network is dangerous for many of VANET applications specially emergency noti cation services. It raises not only safety concerns but also results into degraded performance of VANETs. In our proposed algorithm Trust Based Congestion Control Algorithm (TBCCA), the priorities are computed independently in each vehicle for each message. Here a new trust parameter is considered in computation of message priority. The trust parameter is a ratio of sent and received messages, which determines how much a vehicle is contributing in congestion. The trust parameter not only aids in priority calculation but also helps in further broadcast of messages sent by trusted vehicles. Thus our algorithm controls congestion on the roads more efficiently. Our proposed algorithm has improved efficiency of VANET by controlling the congestion which reduced end to end delay by 7.778% and reduced packet loss by 8.333%. Our algorith also increased the throughput by 13.726%.

So far Electroencephalography (EEG) has been analyzed by the re- search community for interaction with the computers. Studies regarding EEG signals has gained attention in the recent past as it gives an alternate way of com- munication for the persons suffering from partially or fully paralytic disability. Every second different activities are performed by millions of neurons. Decoding and detecting such complex activity of the brain while analyzing the EEG signals is a challenging task. In this paper, we have proposed an activity recognition system using EEG signals. The two activities, namely, eyes open (EO) and eyes close (EC) have been considered in this work. The recorded signals are then decomposed using Discrete Wavelet Transform (DWT) to analyze the impact of both the activities. The recognition of activities has been performed using Support Vector Machine (SVM) classifier. For experimentation, a publicly available dataset i.e. PhysioNet consisting data of 109 users while performing one minute EO and EC activity has been used. A notable activity recognition rate of 86.08% has been recorded using gamma band feature. The paper further proposes that the system can be used as a reference to detect different types of activities performed at different instance of time and for rehabilitation purposes also.