תפריט נגישות
ניגודיות עדינהניגודיות גבוהה
מונוכרום
הדגשת קישורים
חסימת אנימציה
פונט קריא
סגור
איפוס הגדרות נגישות
להורדת מודול נגישות חינם תפריט נגישותניגודיות עדינהניגודיות גבוההמונוכרוםהדגשת קישוריםחסימת אנימציהפונט קריאסגוראיפוס הגדרות נגישותלהורדת מודול נגישות חינםEarly detection of non-optimal weed control is now a priority to ensure herbicide efficacy. This study aimed to evaluate the potential of hyperspectral imaging (HSI) for early detection of the effects of glyphosate and glufosinate on weeds. Specific features (bands and vegetation indices [VIs]) were extracted as indicators for glyphosate and glufosinate efficacy. Black nightshade (Solanum nigrum L.) was used as the model weed and treated with glyphosate or glufosinate at the fourth-leaf stage. Plants were imaged in the laboratory at 6, 24, 48, 72, and 96 hours after treatment (HAT) with a 204-wavelength (400–1000 nm) hyperspectral camera. The impact of the herbicide treatments on the spectral reflectance values was analyzed using the two-sided Mann–Whitney U test, followed by classification with a machine learning (ML) model applied on the full spectrum and on 12 VIs. In addition, the contribution of the different wavelengths (features) to classification accuracy was assessed using a feature selection process. For glufosinate, 95% classification accuracy was observed as early as 6 HAT, with four features from the green region required. For glyphosate, four features from the red, red-edge, and green regions were used to achieve 88% classification accuracy at 24 HAT. The accuracy of VIs-based classification was generally lower than that of the full spectrum classification accuracy. Above 85% classification accuracy was achieved only at later imaging campaigns, starting at 48 HAT. This study thus demonstrates that non-optimal application of glyphosate and glufosinate can indeed be detected using spectral imaging.
Early detection of non-optimal weed control is now a priority to ensure herbicide efficacy. This study aimed to evaluate the potential of hyperspectral imaging (HSI) for early detection of the effects of glyphosate and glufosinate on weeds. Specific features (bands and vegetation indices [VIs]) were extracted as indicators for glyphosate and glufosinate efficacy. Black nightshade (Solanum nigrum L.) was used as the model weed and treated with glyphosate or glufosinate at the fourth-leaf stage. Plants were imaged in the laboratory at 6, 24, 48, 72, and 96 hours after treatment (HAT) with a 204-wavelength (400–1000 nm) hyperspectral camera. The impact of the herbicide treatments on the spectral reflectance values was analyzed using the two-sided Mann–Whitney U test, followed by classification with a machine learning (ML) model applied on the full spectrum and on 12 VIs. In addition, the contribution of the different wavelengths (features) to classification accuracy was assessed using a feature selection process. For glufosinate, 95% classification accuracy was observed as early as 6 HAT, with four features from the green region required. For glyphosate, four features from the red, red-edge, and green regions were used to achieve 88% classification accuracy at 24 HAT. The accuracy of VIs-based classification was generally lower than that of the full spectrum classification accuracy. Above 85% classification accuracy was achieved only at later imaging campaigns, starting at 48 HAT. This study thus demonstrates that non-optimal application of glyphosate and glufosinate can indeed be detected using spectral imaging.
