Page 406 - NGTU_paper_withoutVideo

Modern Geomatics Technologies and Applications

Submission Template for 1st international conference and 2nd national conference on
“Modern Geomatics Technologies and Applications”, in honor of Professor Abbas
Rajabifard

CROP ROW DETECTION AUTOMATICALLY USING LOW-COST UAV
IMAGERY SYSTEM
2
Abolfazl Sharifi , Reza Naimaee , Mohammad SaadatSeresht 3
1*

1 Master student in School of Surveying and Geospatial Engineering, College of Engineering,
University of Tehran
2 Master student in School of Surveying and Geospatial Engineering, College of Engineering,
University of Tehran
3 Assistant Professor in School of Surveying and Geospatial Engineering, College of Engineering,
University of Tehran
* abolfazl.sharifi@ut.ac.ir

Abstract: Precision Agriculture (PA) systems have received a lot of attention in recent decades. Precision Agriculture is a
combination of different technologies such as navigation and imaging systems to control and manage the process of
planting to harvest and aims to increase the quality and quantity of production while reducing costs and There have been
possible injuries. Using Image processing sciences, remote sensing, etc to detect complications in various applications is
constantly expanding. The field of agriculture is no exception to this, and the automatic identification of plant planting
lines is an essential part of the agricultural process and helps a lot in making decisions in this field. So far, various methods
and algorithms for automatic extraction and identification of plant planting lines have been proposed. These algorithms
typically use data such as RGB or orthophoto image, NIR image, and digital elevation model (DSM). This paper tried to
identify crop lines in orthophoto images using digital image processing operators, Mathematics, and remote science. To
this goal, the initial crop lines were identified from the Hough algorithm and the final lines were extracted using a series
of rules. High execution speed, high accuracy, ability to detect direct lines in one direction are advantages of this method
compared to other algorithms presented so far. Implementing this algorithm on different parts of an aerial orthophoto
image showed 92% accuracy in identifying crop lines.

Keywords: image processing, agriculture, plant planting lines, UAV.

1. Introduction
The agriculture industry plays an important role in the life and development of any community; therefore, it was essential
to develop a management system to enhance the outcome while controlling the use of the agriculture process inputs for economic
and environmental purposes. For such needs, Precision Agriculture (PA) was introduced as a smart management system that
aims to distribute the different agriculture inputs like water, fertilizers, herbicides, etc. based on the needs of each spot in the
agriculture field while fitting the environmental and economic requirements[1]. Precision farming is a new, integrated, and
internationally standardized method that aims to increase resource efficiency, reduce its detrimental effects on the environment,
and reduce uncertainty in farm management decisions. Precision agriculture is one of the most scientific and modern production
methods in the 21st century. It focuses on the balanced method between traditional knowledge and information of contemporary
management technology[2].
Meanwhile, UAVs proved to be an exceptional platform for imagery and remote sensing applications during the last two
decades. The significant development of drones, which have been widely used in the past few years, has bridged the gap between
remote and terrestrial platforms. They can deliver imagery data with suitable spatial and temporal resolution for different
applications [3]. One of the most important precision farming programs is identifying and extracting plant planting lines and is
an essential step for other precision farming programs such as crop classification, weed identification, and intelligent
management and planning [4]. Currently, planting lines can be easily identified through many of the world's most advanced
technologies, including remote sensing techniques, aerial platforms, and sensors installed on agricultural machinery [5]. UAV
imagery systems were used for these advantages for different applications, including PA applications as vegetation segmentation
[6]. Weed management [6-8] and crop row detection [9-11] one of the important PA applications is crop row detection, especially

1