Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
RSA based improved YOLOv3 network for segmentation and detection of weed species
Ist Teil von
Multimedia tools and applications, 2024-04, Vol.83 (12), p.34913-34942
Ort / Verlag
New York: Springer US
Erscheinungsjahr
2024
Link zum Volltext
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
Weeds are among the major risks impacting agricultural production and quality, it is still difficult to create reliable weed identification and detection systems because of the unstructured field circumstances and substantial biological heterogeneity of weeds. The proposed work develops a weed detection model for achieving higher crop yield. Crop and weed images are taken as an input for the proposed method. The collection of data consists of raw-data that cannot produce high accuracy. So, a certain pre-processing technique is used in the proposed method for achieving high accuracy. Adaptive median filter, adaptive gamma correction and high boost filtering techniques are used as pre-processing techniques for noise removal, contrast enhancement and edge sharpening. Then the pre-processed image is segmented and detected according to the features and properties of the pixels in the image. Improved YOLOv3-technique is used in the proposed approach for segmentation and detection of weed species. RSA-optimization is used to select the hyperparameters of YOLOv3-optimally. The proposed method is tested with several metrics which attain better performance like 96% accuracy, 96% precision, 95% recall, 4% error and 95% specificity value. Thus the designed model detects weed-species in an effective manner and it is useful for achieving higher crop production.