首页|Investigators at University of Palermo Detail Findings in Robotics and Machine L earning (Comparison of Different Computer Vision Methods for Vineyard Canopy Det ection Using Uav Multispectral Images)
Investigators at University of Palermo Detail Findings in Robotics and Machine L earning (Comparison of Different Computer Vision Methods for Vineyard Canopy Det ection Using Uav Multispectral Images)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Investigators discuss new findings in Robotics and Machine Learning. According to news reporting originating in Palerm o, Italy, by NewsRx journalists, research stated, "In viticulture, the rapid and accurate acquisition of canopy spectral information through ultra-high spatial resolution imagery is increasingly demanded for decision support. The prevalent practice involves creating vigor maps using spectral data obtained from pure vin e canopy pixels." Financial support for this research came from Ministry of Education, Universitie s and Research (MIUR). The news reporters obtained a quote from the research from the University of Pal ermo, "Based Image Analysis (OBIA) among conventional methods exhibits a reasona ble efficiency in canopy classification due to its feature extraction capabiliti es. In recent years, deep learning (DL) techniques have demonstrated significant potential in orchard monitoring, leveraging their ability to automatically lear n image features. This study assessed the performance of different methodologies , including Mask R-CNN, U-Net, OBIA and unsupervised methods, in identifying pur e canopy pixels. The effectiveness of shadow and background detection methods an d the impact of misclassified pixels on NDVI were compared. Results were compare d with agronomic surveys conducted during the 2021 and 2022 growing seasons, foc using on two distinct phenological stages (BBCH65-BBCH85). Mask R-CNN and U-Net exhibited superior performance in terms of Overall Accuracy (OA), F1-score, and Intersection Over Union (IoU). Among OBIA methods, the Gaussian Mixture Model (G MM) proved to be the most effective classifier for canopy segmentation, and Supp ort Vector Machine (SVM) also demonstrated reasonable stability. Conversely, Ran dom Forest (RF) and K-Means yielded lower accuracy and higher error rates. As a result of the limited accuracy, it is noted for vineyard rows with low vigor can opies that NDVI was overestimated, while for high vigor canopies NDVI was undere stimated. Significantly improved determination coefficients were observed for th e comparison between Total Leaf Area (TLA) and NDVI data derived from Mask R-CNN and U-Net. Positive correlations were also found with NDVI data from GMM and SV M algorithms. Regarding leaf chlorophyll (Chl) and NDVI correlations, Mask R-CNN and U-Net methods showed superior performance.
PalermoItalyEuropeRobotics and Mac hine LearningBiological FactorsChlorophyllChlorophyllidesMetalloporphyri nsPorphyrinsUniversity of Palermo
NETL
NSTL
