Scholarly record
WEED SPECIES IMAGE RECOGNITION USING DEEP LEARNING TECHNIQUE FOR SELECTIVE SPOT SPRAYING
Abstract
Application of image processing recognition in the agricultural field is a difficult and intensive task. Recent years developments of technologies such as GPUs (Graphics Processing Units) and the fast development of artificial intelligence algorithms allows the reliability of using computer vision technology to be used for recognition of plants, including weeds, ensuring the efficiency of intelligent agricultural systems. Computer vision technology uses a camera and a computer to identify and measure objects in the scene. With the development of computer vision algorithms , such technology has been taken into account for the use in agriculture. Intelligent weed control systems are used to reduce herbicide usage as it allows more efficient selective spraying to weed targets. Minimizing the use of chemicals would translate in a low quantity of chemicals used and also in important economic impact. Development of intelligent agricultural spray will reduce use of herbicides and improve productivity. Detection and classification of weed using computer vision algorithm is one important step in developing industry acceptance of intelligent weed control technology. In this paper a deep learning technique is used in order to be used along with computer vision to identify weeds. The emerging technology of deep learning for object detection and classification needs to consider several factors such as the optical system, scene variability, negative samples, training data set, position relative to the weeds positions and 3d dimensions, labelling. Image processing application to agriculture implies special technical difficulties such as light variation, dust presence. The trained data set must match the target application morphology concerning leafs as the same weeds have different morphology depending on its age. The majority of current weed species classification pose unique challenges concerning computer vision and training using big data sets .
Publication Impact Profile
- Captures
- Mendeley - Readers: 22
Publication details
References4
H. Tian, T. Wang, Y. Liu, X. Qiao, Y. Li, Computer vision technology in agricultural automation —A review, Information Processing in Agriculture, vol. 7, Issue 1, pp. 1-19, 2020.
M. H. Asad and A. Bais, Weed detection in canola fields using maximum likelihood classification and deep convolutionalneural network, Information Processing in Agriculture,DOI: 10.1016/j.inpa.2019.12.002.
X.- E. Pantazi, D. Moshou, C. Bravo, Active learning system for weed species recognition based on hyperspectral sensing, Biosystems Engineering, vol. 146, pp. 193-202, 2016.
S. Amatya, M. Karkee, A. Gongal, Q. Zhang, M. Whiting, Detection of cherry tree branches with full foliage in planar architecture for automated sweet-cherry harvesting, Biosystems Engineering, vol. 146, pp. 3-15, 2016.
View or Download full articleAccess options
SWS access login
Login as SWS Scientific CommitteeLogin as SWS Scientific PartnerLogin as SWS AuthorAuthors and approved SWS contributors will read and export their own linked papers after identity matching by SWS profile, email and SGEM GlobalID.
For librarian assistance: [email protected]
Purchase Instant Access
- Article can be downloaded after successful payment.
- Article may be used according to SWS library access terms.
- Article cannot be redistributed.