Welcome A. Prof. Singara Singh Kasana from Thapar Institute of Engineering and Technology Patiala-147004 to be Keynote Speaker!
日期：11/19/2019 点击量： 286次
A. Prof. Singara Singh Kasana, Computer Science and Engineering Department, Thapar Institute of Engineering and Technology Patiala-147004, Punjab, India (click)
Research Area: Machine Learning; Image Processing; Information Security
Title: Deep Learning in Agriculture
Abstract: Deep Learning (DL) based techniques are getting major attention in many areas of agriculture due to their high accuracy in predictions. These techniques require bulky data to train their models. To collect these data from the agriculture fields is a very difficult, time consuming and costly task. On the other hand, data from satellites can be easily collected on the click of a mouse. These data are very necessary to monitor changes occurring in the materials of the soil. By using these changes, a farmer can take necessary steps so that crops production in his fields is not effected. Keeping this in mind, we have developed DL based techniques by using Long Short Term Memory (LSTM) model to predict the amount of components in the soil. Highly hierarchical structure and large learning capacity of LSTM model allows them to perform better predictions. Hyperspectral data has been used as input for the model. Data of hyperspectral images are sequential in nature, due to which it is perfect to use LSTM model to predict the output of the material. Our findings indicate that deep learning based model provides high accuracy, outperforming existing commonly used machine learning techniques. There are better scopes of using Advanced DL techniques in identification of weeds, land cover/use classification, recognition of plants/vegetables, fruits counting and classification of crops etc.
