MPhil Thesis

Browse

Now showing 1 - 4 of 4

An Application of Deep Transfer Learning To Detect Lychee Leaf Disease
(Daffodil International University, 23-02-18) Jisan, Tareq Rahman
Bangladesh is a primarily agricultural nation. The majority of people depend on agriculture. Our nation of Bangladesh also heavily relies on agriculture. In the current state of affairs, it is crucial that we increase the yields of our crops and fruits in order to grow them. Bangladeshi people and farmers are fighting to grow their crops and fruits in a crucial way despite the country's extreme and changeable climate. Since Bangladesh is an agricultural nation, it is a sad fact that the quality and quantity of our fruits are declining due to various diseases. People in our nation are discovering numerous new rare diseases in our native fruits, but we are failing to recognize these diseases, and the severity of this issue is growing daily. So, in order to combat this issue, proper treatment or recovery is required. As Bangladeshis, it is very difficult for us to identify this rare disease and we require classification of these issues. Since we live in a technological age, it goes without saying that technology can be extremely helpful in identifying these diseases. It is crucial to first identify leaf disease because growing a healthy plant depends on the plant's leaves. As a result, we can maintain a healthy environment for both the leaves and the fruits. In our research, we are trying to identify leaf diseases. Research into litchi leaf disease is something we are very interested in since it is the most popular fruit in Bangladesh. Therefore, by preventing disease in our litchi fruit, we can contribute to the Bangladeshi economy. We use cutting-edge image processing tools that are very beneficial to us in order to guarantee the freshness of the leaves. By simply looking at the leaves, it is very difficult to identify any disease. Our technology uses a cutting-edge method called image processing. We are employing CNN (Convolutional Neural Network) and machine vision- based image processing for this purpose.
Apple and Orange Diseases Detection Using Deep Learniing Techniques
(Daffodil International University, 23-01-18) Faraduzzaman, G M
We know that Bangladesh is an agricultural country where almost all people are dependent on agriculture. In today's world where everyone is health conscious, the ability to identify fruits by quality is very important in the food industry. But farmers produced these fruits without the help of practical rational inventions. This can lead to financial mishaps and reduce profits for drivers. Fruit diseases currently pose many economic and environmental problems. But early detection of fruits diseases can prevent these accidents and keep farmers happy. The market sells different kinds of fruits. However, identifying the best quality fruit is a daunting task. Therefore, we developed an automated system to Detect fruits under natural light conditions that can provide a guideline to detect fruit. Based on Convolutional Neural Networks (CNN), I created an "Apple and Orange detection system" online application that can detect fruits and also determine if they have diseases. Not only images of unhealthy fruits were collected, but also images of infected fruits such as apples and oranges. In this study, we used a fully convolutional neural network (FCNN) for infection order and a convolutional neural network for birth-related neural functions. In this paper I applied different algorithm but I didn’t get my expectation result then I applied CNN which provide 82% accuracy. I think this result is helpful for our research.
Enhancing guava harvest forecasting in Bangladesh through supervised machine learning models
(Daffodil International University, 2024-01-01) Munia, Sahela Khan
Accurate forecasting of guava harvest is essential for efficient resource allocation, market planning, and mitigating post-harvest losses. In Bangladesh, the guava industry faces challenges in predicting harvest yields due to the complex interaction of various environmental factors. This study proposes a novel approach to enhance guava harvest forecasting in Bangladesh through the application of supervised ML models. The research leverages historical guava production data and corresponding meteorological variables, including temperature, humidity, precipitation, and solar radiation. These variables are used as input features for training and testing several supervised ML models, such as linear regression, decision trees, random forests, support vector machines, and artificial neural networks. A comprehensive dataset comprising guava production records and meteorological data from multiple regions in Bangladesh is collected and preprocessed. Feature engineering techniques are employed to extract relevant information from the data and optimize model performance. The dataset is then divided into training and testing sets for model development and evaluation. Performance metrics such as MAE, RMSE, MSE are used to assess the accuracy and reliability of the machine learning models. Where the highest accuracy 84.72% is achieved by DTR. And the lowest accuracy is achieved by LinR accuracy of 43.07%. The models' forecasting capabilities are compared, and the most effective model is identified. The results demonstrate that the supervised machine learning models exhibit promising performance in guava harvest forecasting, outperforming traditional statistical methods. The selected model achieves high accuracy and provides valuable insights into the influence of meteorological variables on guava production. The findings of this study have significant implications for the guava industry in Bangladesh, helping to enhance productivity, reduce wastage, and promote sustainable agricultural practices. Moreover, the methodology presented are extended to other regions and crops, facilitating improved harvest forecasting in diverse agricultural contexts.
Recognizing Bangladeshi Agricultural Insects Using Machine Learning
(Daffodil International University, 23-01-18) Akter, Tapsara
Recently, some work has been done on agricultural insect recognition. But a limited number of works is done on Bangladeshi agricultural insects. Pests decimate crops on a massive scale each year. To achieve high crop output, pest detection and identification are necessary. For efficient pest control management, early pest detection in photographs is absolutely essential. Therefore, it has been difficult to identify the pest in the picture. I gathered the dataset for this study from a variety of sources. To achieve the greatest results in this study, I combined deep learning and transfer learning. I used some Deep Neural Networks here (DNN). ResNet50 and VGG16 produce the greatest results out of all of them. The model's output demonstrated 96.4% accuracy on the testing dataset, which is superior to other previous works. Keywords — Convolutional Neural Network, Transfer Learning, Bangladeshi Agricultural insect Recognition, Bangladeshi Agricultural insects.