| With the development of hyperspectral remote sensing technology, the fast detection of cadmium (Cd) pollution has become feasible. Lpomoea aquatica Forsk and Brassica chinesis Lee (Chinese Cabbage) were selected as research objects in this study. Plant leaf hyperspectral characteristic was used to monitor the Cd concentrations in their leaves. Six experimental groups were set up with different concentrations Cd of CK (Omg/kg), T1(0.5mg/kg), T2(1mg/kg), T3(5mg/kg), T4(10mg/kg) and T5(20mg/kg). Growth indicators, chlorophyll concentration in plant leaf and hyperspectral characteristic of the two species were studied. Through the analysis of the chlorophyll contents, Cd concentration in leaf, and hyperspectral spectrums data, we come to the following conclusions:1. Effects of different Cd concentration on the growth indicator of Lpomoea aquatica and Brassica chinesisThe results showed that the growth of plant was restrained in Cd contaminated soil. The plant height and biomass of Lpomoea aquatica and Brassica chinesis under Cd stress were remarkably lower that that in CK. The longer the stress was, the shorter they were, and the less biomass they had. In this research, in the5mg/kg treatment, the plant height and biomass were significantly decreased when compared with CK. As the prolonging of this treatment, inhibitory effect was observed to be stronger. However, low concentration of Cd had promoting effect on these plants. When the Cd concentration in soil was0.5mg/kg, the mean of plant’s height and biomass were found to be higher than that of CK.2. Effects of different Cd concentration on the chlorophyll contents of Lpomoea aquatica and Brassica chinesis L. leaves.The research showed that the chlorophyll-a and total chlorophyll of Lpomoea aquatica were significantly decreased on the15th day, when treated with10mg/kg and20mg/kg Cd concentrations. On the30th day, its total chlorophyll and chlorophyll-a contents became to decrease in the medium concentrations Cd (T3) contaminated soil, while no significant difference of the same index was found between the T1and T2treatment. On the other hand, Brassica chinesis was similar to Lpomoea aquatica. On the30th day, chlorophyll-a and total chlorophyll were significantly different from that in CK in high concentration of Cd (T5), while they were lower than CK in terms of other treatments respectively.3. Effects of different Cd concentration on the sensitive bands to chlorophyll concentration and spectral index of Lpomoea aquatica and Brassica chinesisData between400-899nm were re-sampled, and the correlation between new data in these bands and the total chlorophyll of Lpomoea aquatica and Brassica chinesis were studied. The sensibly bands with a high correlation with Lpomoea aquatica found to be R710, R700, R540, R720, R550and R530, that of Brassica chinesis were R380, R550and R710. Especially, R710very significantly correlated with total chlorophyll of both species with the correlative coefficients of0.867and0.757respectively.Analysis of correlation for spectral parameters and total chlorophyll content in Lpomoea aquatica leaf was studied as well. Spectral indices, which the correlation coefficients came to the top, were Dr, RVI (550) and GNDVI. Specifically, in terms of Brassica chinesis, these indices were NDVI(864), RVI(864) and the slope of the red edge (Dr). Additionally, Dr was obviously correlated with the total chlorophyll of2species, and both of the correlation coefficients were beyond0.7.R710and Dr had closed relationship with chlorophyll concentration for2species. The R2of linear regression model reached to0.545and0.707respectively. Therefore, R710and Dr could be used as spectral index to monitoring the growth of leafy vegetables under different Cd concentration.4. Spectral inverse study of Cd concentration in the leaves of Lpomoea aquatica and Brassica chinesisThe samples from bands of350-1300nm of2species under different Cd stress treatments and different process times were re-performed. The average spectrum of different treatments were got then. Due to the different growth status and leaf structures at different growth period, spectrum data under different times were analyzed separately, to reflect the Cd pollution of Lpomoea aquatica and Brassica chinesis in different growth period. The results showed that leaf hyperspectral characteristic had the best relations with leaf Cd concentration on15th day in terms of Lpomoea aquatica, and Brassica chinesis15th day leaf spectrum was then analyzed. Characteristic bands were extracted, and performed with principal component analysis (PCA). Estimation model was found between principal component F1and leaf Cd concentration through univariate regression analysis. The R2of linear regression model reached to0.714.Similarly, spectral properties of Brassica chinesis leaves and their Cd concentration showed good correlationship as well. R430, R440, R550, R680, R710, R740, R960, R1100, R1150, and R1260were selected based on spectral properties and reflectance spectroscopy. PCAs in these ten bands were performed then and new comprehensive indices were obtained. An linear model of these new data and leaf Cd concentration was built. The univariate regression analysis of F1was the best。... |
PDF Full Text Request