Research on Lettuce Canopy Image Processing Method Based on Hyperspectral Imaging Technology.

Area of Science:

• Agricultural remote sensing
• Plant imaging and analysis

Background:

• Accurate segmentation of lettuce canopy images is challenging due to uneven illumination and background interference.
• Hyperspectral imaging offers a potential solution for detailed plant analysis.

Purpose of the Study:

• To develop and verify an accurate method for segmenting lettuce canopy images using hyperspectral imaging.
• To address challenges of uneven illumination and background noise in plant image analysis.

Main Methods:

• Hyperspectral imaging was used to capture lettuce images across different growth stages.
• The spectral ratio method identified characteristic wavelengths (553.8 nm, 702.5 nm, 731.3 nm).
• Image denoising, fusion, filtering, and multi-threshold segmentation were applied.

Main Results:

• The optimal characteristic wavelengths for lettuce segmentation were identified as 553.8 nm, 702.5 nm, and 731.3 nm.
• The proposed method achieved a high area overlap degree (AOM) of 0.9526 and a low misclassification rate (ME) of 0.0477.
• The segmentation method demonstrated excellent stability with variance less than 0.01.

Conclusions:

• Hyperspectral imaging combined with multi-wavelength and multi-threshold segmentation provides accurate lettuce canopy segmentation.
• This technique effectively overcomes challenges posed by illumination variations and background interference.
• The validated method offers a robust tool for precise agricultural image analysis and crop monitoring.