• Version
• Download 8
• File Size 620.42 KB
• File Count 1
• Create Date 9 May 2023
• Last Updated 9 May 2023

FARMER MOVEMENT ANALYSIS AND ALERT SYSTEM USING MACHINE LEARNING

FARMER MOVEMENT ANALYSIS AND ALERT SYSTEM USING MACHINE LEARNING

R.THENMALAR , ASHWIN.C.KUMAR, ABHINAV.V.NAMBIAR, N.SANJAY, C.SUKUMAR

Dept of Computer Science & Engineering
RVS College of Engineering & Technology,
Coimbatore

Abstract— Agriculture is one of the major and the least paid occupation in India, where most of the workers are elderly people and most of them work remotely at rubber estates, teak plantations, tea plantations  at high altitudes and unsteady terrain. The objective of agriculture fall detection using machine learning is to improve the safety of agricultural workers and reduce the risk of injury or death due to falls. This can help to reduce severity of injuries and potentially save lives.An automated fall detection system will provide timely assistance and hence, it will reduce medical care costs significantly. The paper presents a machine learning framework consisting of data collection, pre-processing of data, feature extraction and machine learning classifiers. They comprise Random Forest, Naïve Bayes & Decision Tree Classifiers. The Random Forest algorithm combines multiple decision trees to improve accuracy of the classification model, it can handle noisy and complex data. Naïve Bayes handles uncertain and missing data in the input. Decision Tree Classifier is a class capable of performing multi-class classification on a dataset. We analyse original acquisition datasets of values obtained from two accelerometers and on gyroscope performing falls and Activities of Daily Living (ADL) from three dimensional axis (x, y, z). Datasets are pre-processed using pandas is split to train and test model. The accuracy, precision and recall values are calculated using accuracy_score from Sci-Kit learn. Spatial characteristics were used to train the machine learning classifiers to distinguish between fall and non-fall event. The activity can be monitored by respective farm controller using a web application and alert sound is given if the model detects fall or unusual activity.

Keywords—Fall Detection, Machine Learning, Classifiers, Activities of Daily Living, Gait analysis, Monitoring, Health system