Search

Soil texture using a hydrometer or pipette method requires expertise, although these are accurate. A soil expert may help the farmer to detect the soil texture by analyzing the visual texture of the soil, which is not always accurate. This paper presents the smartphone image-based sand and clay soil classification in wet and dry humid conditions using Self Convolution Neural Network (SCNN) and finetuned MobileNet.A soil dataset of 576 soil images was prepared using a low-cost smartphone under natural light conditions. Different augmentation techniques such as shift, range, rotation, and zoom were applied to the soil dataset to increase the number of images in the soil dataset. The best performance of the MobileNet was reported at epoch 15 with a testing and training loss of 0.0091 and 0.0194, respectively. Though the SCNN model performed best at epoch 10 with a testing accuracy of 99.85%, the MobileNet reported less computation time (167.8s) than the SCNN (273.2s). The precision and recall of the models were 99.62 (MobileNet) and 99.84 (SCNN). The accuracy of the SCNN reported itself as the best model, whereas the computing time of the MobileNet reported itself as the best model in different humid conditions. The model can be used to replicate the traditional soil texture analysis method and the farmers can use it for better productivity.

The citrus family provides healthy fruits to humans. The quality and quantity of the citrus fruits depend on the quality of the citrus leaves. Due to the diseases of citrus leaves, the quality and productivity of citrus fruit are degraded. This paper provides a computer automation system to detect the diseases of the citrus leaves using machine learning and deep learning techniques. In this paper, images of citrus leaves are captured using an Android Smartphone in natural environmental light. Three different classes of citrus leaves are collected using the Smartphone and destructive method. These are citrus healthy, citrus greening, and citrus CTV (Citrus Tristeza virus). Citrus images are processed for image resizing, image noise removal, image enhancement, and feature extraction. Features of the images are calculated using Gray Level Co-occurrence Matrix in the color and gray domain of the images. Finally, the detection and classification are done using K- Nearest Neighbor (KNN) and Deep Neural Network (DNN) classifier. The accuracy of the KNN classifier is 89.9% for the K value 3 whereas the accuracy of DNN is 99.89% with an error of 0.0219. The proposed system helps the farmer to detect diseases at an early stage.

Identifying race, which is a major physical feature in humans, is still a challenging task owing much to the lack of a concrete definition of race and the diversity of population across the globe. In this paper, we try to address the problem of race identification of four broad racial groups namely Caucasian, African, Asian and Indian. The newly developed BUPT Equalised Face dataset bearing about 1.3 M images in unrestricted environment is used to train our deep convolution network (R-Net) which achieves a state-of-the-art accuracy of 97%. To test the validity of this model it is evaluated on other datasets namely UTK and CFD. R-Net is also compared with fine-tuned VGG16 model for race estimation. Experimental results prove the robustness of this model for use in unconstrained environments. And finally, Gradient-weighted Class Activation Mapping (Grad-CAM) is applied to get a visual explanation of the deep learning model.

Immature and tender tea leaves always produce high-quality tea than mature tea leaves. Depending on the maturity and age of the leaf, the colour and texture of the tea leaf are different. The photosynthesis capacity of the tea leaf also changes with the change of leaf maturity. Though the tea farmer plucks, classifies, and recognizes the best tea leaves (immature and tender) by viewing the visual symptoms and position of the leaves, the method is not authentic all time and leads to the overall degradation of the tea quality. The present study presents a smartphone assist tea leaf recognition system by analyzing the colour and texture properties of the tea leaf. The six different colour features and 4 Haralick texture features were extracted in the colour and grey domain of the leaf images. Three types of tea leaves, i.e., mature, immature, and tender, were classified using Deep Neural Network (DNN) with ADAM (Adaptive Moment Estimation) optimizer. With an accuracy of 97%, the DNN outperformed the Support Vector Machine (SVM) and K Nearest Neighbor (KNN). The SVM and KNN reported a total of 94.42% and 95.53% accuracy, respectively. The investigated system using DNN with an average precision and recall value of 98.67 and 98.34, respectively, may detect and classify the tea leaf maturity status. The system also can be used in AI-based tea plucking robotic systems or machines.

The objective of this study is to process the soil images to generate a digital soil classification system for rural farmers at low cost. Soil texture is the main factor to be considered before doing cultivation. It affects the crop selection and regulates the water transmission property. The conventional hydrometer method determines the percentage of sand, silt, and clay present in a soil sample. This method is very cost and time-consuming process. In this approach, we collect 50 soil samples from the different region of west Guwahati, Assam, India. The samples are photographed under a constant light condition using an Android mobile of 13 MP cameras. The fraction of sand, silt, and clay of the soil samples are determined using the hydrometer test. The result of the hydrometer test is processed with the United State Department of Agriculture soil classification triangle for the final soil classification. Soil images are processed through the different stages like pre-processing of soil images for image enhancement, extracting the region of interest for segmentation and the texture analysis for feature vector. The feature vector is calculated from the Hue, Saturation, and Value (HSV) histogram, color moments, color auto Correlogram, Gabor wavelets, and discrete wavelet transform. Finally, Support Vector Machine classifier is used to classify the soil images using linear kernel. The proposed method gives an average of 91.37% accuracy for all the soil samples and the result is nearly the same with the United State Department of Agriculture soil classification.

Automatic detection of citrus leaves disease is very much essential for the better productivity of citrus. Citrus leaves are affected by bacteria, fungus and virus respectively. Farmer detects the diseases of the plant using laboratory, naked eyes or using expert’s view. The rural farmers often face difficulties to detect these diseases due to the non availability of the laboratories in their area. Here in this paper, a computer automation system is proposed to detect the diseases of citrus leaves on an early stage. Citrus leaves images are captured using Smartphone. Captured images are used to extract the different features of the citrus leaves samples using Gray Level Co-occurrence Matrix. Finally, citrus greening and citrus CTV images are classified from citrus healthy images using Gaussian kernel based support vector machine. Accuracy of the kernel is evaluated for the different values of Gamma parameter of kernel. The Gaussian kernel gives maximum accuracy (95.5%) with Gamma value 1.

Weather Forecasting is a serious issue in agriculture, especially in Assam and Meghalaya. The productivity of agriculture is dependent on rainfall. This paper forwards a comparative assessment of Auto-Regressive, Moving Average, and Auto-Regressive Moving Average Model for rainfall in Assam and Meghalaya. A total of 117 years of rainfall data of Assam and Meghalaya division is collected from data.gov.in [11]. The models are implemented by visualizing the time series components of rainfall. The necessary investigations such as ACF, PACF, rolling mean, and ducky fuller tests are reported in the paper for the analysis of stationarity of time series. The evaluating parameters such regression score (0.73), mean absolute error (75.70), median absolute error (61.43), mean squared error (9396.09), mand root mean square error (96.93) select the ARMA model as the best model for the time series forecasting of Assam and Meghalaya Division.

This paper presents the design and development of an expert system that aim to provide for suitable diagnosis of some of the diseases of rice plants. An expert system can be defined as a computer program that uses encoded knowledge to solve problems in a specific domain that normally requires a specialized human expert. The proposed system composed of knowledge base, inference mechanism or control structure and user interface. It facilitates for diagnosis the common rice diseases on the basis of the user responses made against the queries related to a particular disease symptoms. Here, in this system users or farmers come with the problem or diseases occur in the rice plants during life span. Then these problems are input by the users through a user's interface and compared to the Knowledge base which is in the form of rules and identifies the probable disease.