Efficient Maize Tassel-Detection Method using UAV based remote sensing

Regular monitoring is worthwhile to maintain a healthy crop. Historically, the manual observation was used monitor crops, which time-consuming and often costly. The recent boom in development of Unmanned Aerial Vehicles (UAVs) has established quick easy way crops. UAVs can cover wide area few minutes obtain useful crop information with different sensors such as RGB, multispectral, hyperspectral cameras. Simultaneously, Convolutional Neural Networks (CNNs) have been effectively for various vision-based agricultural activities, flower detection, fruit counting, yield estimation. However, Network (CNN) requires massive amount labeled data training, not always obtain. Especially agriculture, generating datasets exhaustive since interest objects are typically small size large number. This paper proposes novel method using k-means clustering adaptive thresholding detecting maize tassels address these issues. qualitative quantitative analysis proposed reveals that our performs close reference approaches an advantage over computational complexity. detected counted precision: 0.97438, recall: 0.88132, F1 Score: 0.92412. In addition, tassel detection from UAV images task this paper, we propose semi-automatic image annotation create easily. Based on method, developed tool be conjunction machine learning model provide initial annotations given image, modified further by user. Our tool's performance promising savings time, enabling rapid production datasets.