Paper
29 January 2007 Research on crop and weed identification by NIR spectroscopy
Jiazhi Pan, Yueming Tang, Yong He
Author Affiliations +
Proceedings Volume 6279, 27th International Congress on High-Speed Photography and Photonics; 62797D (2007) https://doi.org/10.1117/12.725951
Event: 27th International congress on High-Speed Photography and Photonics, 2006, Xi'an, China
Abstract
Crop and weed identification is very importance in precision farming field. As spectroscopy can reflects the contents of object tested, so it is possible to identify crop and weed with high correct rate. ASD FieldSpec recorded the spectrum of crops and weeds. Its waveband is 325-1075nm and with resolution of 3.5nm. One crop seedling and three kinds of weeds living together were tested. Each species has at least 30 sampling spectrum taken down. As one sample spectrum has too much data, wavelet transform reduced the data volume firstly, which compressed source signals to tens of floating numbers from 751 floating numbers. Totally 160 samples were used to build a radial basis function neural network, the object output was a 4 by 1 dimension vector. Those left 43 samples used to check the identifying capability. As neural network model has huge power in solving these pattern recognition problems. It can approach to giving finite function at any approximation. Nearly all these predicting samples classified right. Therefore, by using spectroscopy in the identification is possible, and having high correct rate. Further more, the computation is very fast. Whereas the spectrometer is expensive and easily affected by shaking and variation of light shine, it cannot installed directly on vehicles at present time. In the future, it may be possible to recognize crop and weed in real time by using spectroscopy.
© (2007) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jiazhi Pan, Yueming Tang, and Yong He "Research on crop and weed identification by NIR spectroscopy", Proc. SPIE 6279, 27th International Congress on High-Speed Photography and Photonics, 62797D (29 January 2007); https://doi.org/10.1117/12.725951
Lens.org Logo
CITATIONS
Cited by 1 scholarly publication.
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Spectroscopy

Neural networks

Neurons

Wavelet transforms

Near infrared

Data modeling

Wavelets

Back to Top