Matches in SemOpenAlex for { <https://semopenalex.org/work/W4296340309> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4296340309 abstract "It is beneficial for growers and breeders to optimize the planting process, to make better decision on the production process, and to improve the fruit quality by predicting and estimating the yield of Xanthoceras sorbifolium Bunge fruit accurately. Due to the lack of automatic counting of Xanthoceras sorbifolium Bunge, this paper proposed a counting method based on deep learning, thus predicting and estimating the fruit yield of Xanthoceras sorbifolium Bunge quickly and effectively, and promoting the development of Xanthoceras sorbifolium Bunge planting industry. Experiments were conducted with deep learning models including SSD, YOLOv2, YOLOv2-tiny, YOLOv3, YOLOv3-tiny, YOLOv4,YOLOv5s and YOLOv5x, respectively, and the optimal model was selected. In view of the similar color of Xanthoceras sorbifolium Bunge fruit and leaves, in order to improve the accuracy of detection and counting, first of all, the experiment was to improve the contrast of the original data set twice. Secondly, the data set of green Malus pumila Mill and Pyrus spp cultivars with similar fruit type was mixed with the data set of Xanthoceras sorbifolium Bunge for data augmentation. In order to further study the influence of different flight modes of UAV on the prediction results, a comparative test was conducted on three flight modes. There is the optimal performance of YOLOv3, with the IOU 98.84%, and there is the highest counting accuracy of YOLOV3-tiny, up to 98.05%. Moreover, the YOLOv3-tiny model is suitable for the counting statistics of Xanthoceras sorbifolium Bunge data not only in large area collected by UAV, but also in local area collected by daily mobile phone, so that it is suitable for predicting and estimating the yield of Xanthoceras sorbifolium Bunge. Appropriately increasing the image contrast can improve the accuracy of the detection and counting of Xanthoceras sorbifolium Bunge fruit, but if the contrast increases too much, the detection accuracy will decrease. The accuracy of detection and counting can also be improved by training with mixed fruit data sets of similar color and fruit shape. The best value of recognition and counting accuracy can be achieved by the two modes of lateral flight and vertical flight, respectively. The identification and counting method based on deep learning can realize the automatic, accurate and efficient estimation of the fruit yield, and overcome the shortcomings of artificial fruit counting, such as missing number, repeated counting and heavy workload." @default.
- W4296340309 created "2022-09-20" @default.
- W4296340309 creator A5038760334 @default.
- W4296340309 creator A5055185648 @default.
- W4296340309 creator A5071963696 @default.
- W4296340309 creator A5076255361 @default.
- W4296340309 creator A5082533055 @default.
- W4296340309 date "2022-07-26" @default.
- W4296340309 modified "2023-10-05" @default.
- W4296340309 title "Research on Fruit Counting of Xanthoceras Sorbifolium Bunge Based on Deep Learning" @default.
- W4296340309 cites W2019168270 @default.
- W4296340309 cites W2026992687 @default.
- W4296340309 cites W2109934232 @default.
- W4296340309 cites W2144898326 @default.
- W4296340309 cites W2183320018 @default.
- W4296340309 cites W2501369945 @default.
- W4296340309 cites W2578363764 @default.
- W4296340309 cites W2611227133 @default.
- W4296340309 cites W2793323341 @default.
- W4296340309 cites W2798415800 @default.
- W4296340309 cites W2798814941 @default.
- W4296340309 cites W2802344721 @default.
- W4296340309 cites W2888488776 @default.
- W4296340309 cites W2894904576 @default.
- W4296340309 cites W2904751442 @default.
- W4296340309 cites W2909494862 @default.
- W4296340309 cites W2936307272 @default.
- W4296340309 cites W2945051190 @default.
- W4296340309 cites W2963523428 @default.
- W4296340309 cites W2967173272 @default.
- W4296340309 cites W2973318735 @default.
- W4296340309 cites W4239147634 @default.
- W4296340309 doi "https://doi.org/10.1109/icivc55077.2022.9886298" @default.
- W4296340309 hasPublicationYear "2022" @default.
- W4296340309 type Work @default.
- W4296340309 citedByCount "0" @default.
- W4296340309 crossrefType "proceedings-article" @default.
- W4296340309 hasAuthorship W4296340309A5038760334 @default.
- W4296340309 hasAuthorship W4296340309A5055185648 @default.
- W4296340309 hasAuthorship W4296340309A5071963696 @default.
- W4296340309 hasAuthorship W4296340309A5076255361 @default.
- W4296340309 hasAuthorship W4296340309A5082533055 @default.
- W4296340309 hasConcept C108583219 @default.
- W4296340309 hasConcept C144027150 @default.
- W4296340309 hasConcept C153180895 @default.
- W4296340309 hasConcept C154945302 @default.
- W4296340309 hasConcept C2777306635 @default.
- W4296340309 hasConcept C41008148 @default.
- W4296340309 hasConcept C58489278 @default.
- W4296340309 hasConcept C86803240 @default.
- W4296340309 hasConceptScore W4296340309C108583219 @default.
- W4296340309 hasConceptScore W4296340309C144027150 @default.
- W4296340309 hasConceptScore W4296340309C153180895 @default.
- W4296340309 hasConceptScore W4296340309C154945302 @default.
- W4296340309 hasConceptScore W4296340309C2777306635 @default.
- W4296340309 hasConceptScore W4296340309C41008148 @default.
- W4296340309 hasConceptScore W4296340309C58489278 @default.
- W4296340309 hasConceptScore W4296340309C86803240 @default.
- W4296340309 hasFunder F4320321001 @default.
- W4296340309 hasLocation W42963403091 @default.
- W4296340309 hasOpenAccess W4296340309 @default.
- W4296340309 hasPrimaryLocation W42963403091 @default.
- W4296340309 hasRelatedWork W2731899572 @default.
- W4296340309 hasRelatedWork W2738221750 @default.
- W4296340309 hasRelatedWork W2939353110 @default.
- W4296340309 hasRelatedWork W2948658236 @default.
- W4296340309 hasRelatedWork W3009238340 @default.
- W4296340309 hasRelatedWork W3215138031 @default.
- W4296340309 hasRelatedWork W4312962853 @default.
- W4296340309 hasRelatedWork W4321369474 @default.
- W4296340309 hasRelatedWork W4327774331 @default.
- W4296340309 hasRelatedWork W4360585206 @default.
- W4296340309 isParatext "false" @default.
- W4296340309 isRetracted "false" @default.
- W4296340309 workType "article" @default.