The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
352 
enhance weak heavy metal stress information, and Gaussian 
membership function is adopted because of its non-constant 
differentiable character. Tow standards, output error and the 
width of Gaussian membership function, are used to decide 
whether a new fuzzy rule should be added into this system. The 
significance of each rule is evaluated to decide whether a rule 
should be deleted. A training data set which is composed of 
250 samples obtained from MODIS data is applied to adjust 
network structure and to generate fuzzy rules. On the basis of 
seven fuzzy reasoning rules, this system can achieve 95% 
accuracy. According to the result of experiment, the advantages 
of this DFNN model are summarized as follows: 
1. It offers faster convergence and is less sensitive to 
both training and testing datasets; 
2. It substantially decreases the number of hidden 
neurons which is crucial in the optimization of network 
structure, as fuzzy rules are generated or deleted 
according to the network performance and the significance 
of each rule; 
3. It is capable of extracting crop heavy metal stress 
information with reasonable accuracy, and thus it could be 
used as an effective tool in monitoring and managing 
agricultural environment. 
ACKNOWLEDGMENT 
This work is under the auspice of National High-tech R&D 
Program of China (863 program) (2007AA12Z174) and 
National Natural Science Foundation of China (40771155). 
REFERENCES 
Atkinson P.M., Tatnall, A.R.L, 1997. Neural networks in 
remote sensing. Int. J. Remote Sens, 18, pp. 699-709. 
D.P. Kanungo, M.K. Arora, S. Sarkar, R.P. Gupta, 2006.A 
comparative study of conventional, ANN black box, fuzzy and 
combined neural and fuzzy weighting procedures for landslide 
susceptibility zonation in Darjeeling Himalayas. Engineering 
Geology, 85, pp. 347 -366. 
Gong Peng, 1996. Integrated analysis of spatial data for 
multiple sources: using evidential reasoning and artificial 
neural network techniques for geological mapping. 
Photogramm. Eng. Remote Sensing, 62, pp. 513-523. 
Hasi Bagan, et al., 2003. Study of ASTER data classification 
using self-organizing neural network method. Advance in 
Earth Sciences, Vol. 18, No.3, pp. 345-350. 
H. Noh, Q. Zhang, B. Shin, S. Han, L. Feng , 2006. A neural 
network model of maize crop nitrogen stress assessment for a 
multi-spectral imaging sensor. Biosystems Engineering, 94(4), 
pp. 477-485. 
Jesus Favela, Jorge Torres., 1998. A two-step approach to 
satellite image classification using fuzzy neural networks and 
the ID3 learning algorithm. Expert System with Application, 
14, pp. 211-218. 
Jiao Yunqing, Wang Shixin, Zhou Yi, Fu Qinghua, 2007. 
Super-resolution target identification from remotely sensed 
imagery using Hopfield neural network. Journal of System 
Simulation, Vol.19, No.14, pp. 3223-3225. 
Kwokwing Chau, 2006. A review on the integration of 
artificial intelligence into coastal modeling. Journal of 
Environmental Management, 80, pp. 47-57. 
Lehmann, F., Rothfuss, H., Werner K., 1991. Imaging 
spectroscopy data used for geological and environmental 
analysis in Europe. In: Summaries of the Third Annual JPL 
Airborne Geoscience Workshop, 20 and 21 May 1991. JPL, 
Pasadena, pp. 62-71. 
L. Kooistr, E.A.L. Salas, 2004. Exploring field vegetation 
reflectance as an indicator of soil contamination in river 
floodplains. Environment Pollution, 127, pp. 281-290. 
Martin Hellmann, Gunther Jäger, 2002. Fuzzy rule based 
classification of polarimetric SAR data. Aerospace Science and 
Technology, 6, pp. 217-232. 
Min Han, Yannan Sun, Yingnan Fan, 2008. An improved 
fuzzy neural network based on T-S model. Expert Systems 
with Applications, 34, pp. 2905-2920. 
Mohammed, G.H., Noland, T.L., Irving, D., Sampson, P.H., 
Zarco Tejada, P.J., Miller, J.R., 2000. Natural and Stress- 
Induced Effects on Leaf Spectral Reflectance in Ontario 
Species. Forest Research Report No. 156, Ontario Ministry of 
Natural Resources, Canada. 
Ralf Wieland, Wilfried Mirschel, 2008. Adaptive fuzzy 
modeling versus artificial neural networks. Environmental 
Modelling & Software, 23, pp. 215-224. 
Shao Dong, Zhou Zhihua, 1999. Research of fuzzy neural 
network. Application Research of Computers, Vol.7, pp. 
1,2,21. 
Sommer, S., Hill, J., Megier, J., 1998. The potential of remote 
sensing for monitoring rural land use changes and their effects 
on soil conditions. Agriculture Ecosystems and Environment, 
67, pp. 197-209. 
Sucharita Gopal, Curtis E. Woodcock, Alan H. Strahler, 1999. 
Fuzzy Neural Network Classification of Global Land Cover 
from a 1° AVHRR Data Set. REMOTE SENS. ENVIRON., 61, 
pp.230-243. 
Wu Yifan, 2004. A data mining algorithm based on fuzzy 
neural network’s pattern classification. Journal of Changsha 
Communications University, Vol 20, No.l, pp. 45-49. 
Yunzhao Wu, Jun Chen, Xinmin Wu, Qingjiu Tian,Junfeng Ji, 
Zhihao Qin, 2005. Possibilities of reflectance spectroscopy for 
the assessment of contaminant elements in suburban soils. 
Applied Geochemistry, 20, pp. 1051-1059. 
Zhang Qiang, Zhou Qiusheng, 2006. Remote sensing image 
classification and study using a kind of fuzzy neural networks. 
Engineering of Surveying and Mapping, Vol 15, No.5, pp. 42- 
46.