Weed detection and removal using robotic system | [ARPN Journal of Engineering and Applied Sciences • 2023]

Author(s):

1. Puvaneswari G.: Department of Electronics and Communication Engineering, Coimbatore Institute of Technology, Coimbatore, Tamil Nadu, India

2. Arulselvan S.: Department of Civil Engineering, Coimbatore Institute of Technology,Coimbatore, Tamil Nadu,India

Abstract:

Weeding is one of the most significant practices in agricultural production. Weeds are unwanted plants that grow along with the crops and compete with the crops for space, light, water, and soil nutrients. Weeds propagate themselves either through seeding or creeping rootstalk and decrease yields, increase production costs, interfere with the harvest, and lower the product quality. The use of herbicides reduces labor requirements for weed control by up to 60 percent but affects environmental quality and can be toxic to a wide range of organisms. Hence it is necessary to develop an automated system to identify and remove weeds from the vegetable fields. The objective of the proposed work is to develop a mobility level tracked bot that identifies the weeds and removes them with the help of a robotic end effector and to develop a machine learning model to identify the weeds. This functional module will be processed in a Raspberry Pi processor and by using a Raspberry Pi camera module the bot will detect the weeds in vegetable fields. We performed weed detection with different machine learning models like Haar cascade, YOLOv5, and CNN. To evaluate the performance of the machine learning models used, the performance metrics accuracy, precision, recall, and F-measure are estimated and it has been found that CNN has better accuracy, precision, and recall as compared to YOLOv5 and Haar cascade. CNN has the highest F-measure among the three algorithms at 98%. The weed removal is done using a robotic end-effector which is controlled by the Arduino UNO based on the signal from Raspberry Pi.

Page(s): 2424-2435

DOI: DOI not available

Published: Journal: ARPN Journal of Engineering and Applied Sciences, Volume: 18, Issue: 21, Year: 2023

Keywords:

machine learning , Convolutional Neural Network , Robot , microcontroller , Weed Detection

References:

[1] Aharon Shlomi,Lati Ran N. .2020 .Image-Based High-Throughput Phenotyping of Cereals Early Vigor and WeedCompetitiveness Traits. Remote Sensing, 12(23) : 3877-3894.

[2] De Castro Ana I.,Jorge I.,López-Granados Francisca .2018 .An automatic random forest-OBIA algorithm for early weed mapping between and within crop rows using UAV imagery. Remote Sensing, 10(2) : 285-306.

[3] Holman Fenner H.,Andrew B.,Riche Martin J.,Wooster Malcolm J.,Hawkesford Malcolm J. .2019 .. , : .

[4] Huang Danfeng .2014 .A review of imaging techniques for plant phenotyping. Sensors., : .

[5] . .Discriminating Crops/Weeds in an Upland Rice Field from UAV Images with the SLIC-. , : .

[6] . .. Plant Production Science, 24(2) : 198-215.

[7] Hassan Anys,Dong-Chen He Anys .1995 .Evaluation of textural and multipolarization radar features for crop classification. IEEE Transactions on geoscience and remote sensing, 33(5) : 1170-1181.

Citations

Downloads

Views