1、1The Study of Plateau Lakes Chlorophyll-a Content based on Remote Sensing TechnologyAbstract. Chlorophyll is an important parameter reflecting the degree of water eutrophication. Its extraction using remote sensing is an important way to monitor dynamic of water eutrophication. Since plateau lakes n
2、ormally serous interfered by human activities, most of them belong to case II water. Their optical properties are very complicated, which increase the difficulty to its chlorophyll extracting. At present, the chlorophyll extraction using remote sensing has been achieved prominent advancement. But th
3、e extraction precision still needs to be further improved. And the universality and stability of extraction model needs to be further enhanced. The article results show that: the Images regional green color and the ground monitoring of chlorophyll concentration showed better relationship, but the im
4、age also reflects the effect of graded concentrations, this method can be used as remote monitoring of Dianchi Lake chlorophyll and mapping the quickest way. Key words: Plateau Lakes; chlorophyll-a; remote sensing; information extraction 21. Introduction Remote sensing monitoring of chlorophyll a co
5、mmonly used method is to first establish the ground truth data and remote sensing image channel data regression model, and then according to the model inversion chlorophyll concentration maps. This is because the remote sensing image corresponding to a continuous spectrum for each channel is only a
6、short interval, and the different objects in the reflection spectrum of the different channels differ. Failure to do special processing can not meet the needs of the application. To solve this problem, will focus on appropriate performance the Dianchi chlorophyll distribution MODIS image fusion tech
7、nology. Using the MODIS back-end processing procedures and EOSshop commercial software platform, the completion of the initial the geographical correction and projection of the image of the study area, and spectral values into reflectance values for atmospheric correction, atmospheric correction mod
8、ule provided by ENVI4.3 data pretreatment. 2. Image fusion Remote sensing image fusion of different spatial and spectral resolution images from different sensors is processed by a particular algorithm, so that the new image of the newly generated simultaneously with multi-spectral characteristics of
9、 3the original image and the high ground resolution,that realize different application needs. Image fusion HSV transform Fusion, PCA transform fusion, the Brovey transform fusion and wavelet transform fusion. This paper attempts to choose the the HSV transform fusion method to achieve image fusion p
10、rocessing of the Dianchi chlorophyll remote sensing monitoring, in order to get a better visual effect. 2.1 HSV Principles of Remote Sensing Image Fusion Due to the correlation of the image channel is not clearly expressed, in general, the RGB color space of the image is not used for fusion treatmen
11、t. And HSV opposite, which clearly illustrates the characteristics of the particular color channel: H (Hue), said hue, S (Saturation) represents saturation, H and S represents the spectral resolution of the image; V (Value) indicates the intensity of the color , corresponding to the image of the gro
12、und resolution. According to this principle, the fusion MODIS images, first 250m resolution generated after the appropriate treatment the V high component, and then replaced after the decomposition of the low-resolution image HSV (H, S, V) in the channels 1,2 V low component constitute a new HSV spa
13、ce (H, S, V), then the inverse transform to the RGB space, to generate the fused image. 42.2 Channel Selection and false color composite Chlorophyll in the blue band of 1-800-552-170 nm and red bands of 678nm near have significant absorption, when the algal density higher water spectral reflectance
14、curves in the vicinity of the two bands appear absorption peak, water absorbed on the near-infrared band comparison strong, when the water-containing chlorophyll, the near-infrared bands significantly uplift. In order to take full advantage of the chlorophyll in the blue band, red band and near-infr
15、ared spectral characteristics, the choice of visible light combinations (RGB: 143) as a low-resolution RGB image. Although the image can be completely natural color reproduction (Figure 1), but the disadvantage is that the atmospheric disturbances, shortwave scattering serious, Vegetation Informatio
16、n weak larger image texture detail loss; 1,2 channel corresponding to the red channel and near-infrared channel, two RGB channels: 121 false color composite effect RGB: 143 (Figure 2) is complementary, 250m channels for improving the resolution of the image is particularly important. 250m resolution
17、 1,2 channel can be used for small water area of Dianchi Lake, and thus high-resolution image to do texture structure (Figure 3) by the formula (1), which both have a high spatial resolution, 5along with the near-infrared spectral characteristics of the channel. 1-7 the channel characteristics are s
18、hown in Table 1. A high-resolution image texture = (b1 b2) * B2 + (B2 B1) * B1 Where b1 is 1 channel of 250m resolution, b2 is 2 channel of 250m resolution. 2.3 Image change RGB to HSV and anti-HSV to RGB change ENVI4.3 platform using RGB-HSV change the above resolution images and high-resolution im
19、ages to the changes in the RGB-HSV, HSV-RGB function against changes to the high spatial resolution image V (H) to replace the low-resolution the rate images HSV exploded V (low), to constitute a new HSV space, the RGB space, the inverse transform to generate the fused image (Figure 4). Table 1 chan
20、nel characteristics 2.4 Spectral characteristic parameters and chlorophyll correlation Studies have shown that chlorophyll a, chlorophyll a content increased reflectance spectra of the water change, in the blue band 440 nm red band 678nm near significant absorption spectral reflectance curves of the
21、 water when the algae density is higher near the two bands The absorption peak appears. In the vicinity of 685nm, the chlorophyll a fluorescence peak, 6containing the most significant spectral characteristics of algae water body is the peak of reflectance appears in the vicinity of 685 715nm, the pe
22、ak position and the peak of its reflection is a chlorophyll a concentration of directions. For inland body of water the green reflection peak in the vicinity of 550 to 570 nm as the chlorophyll a quantitative signs. Remote sensing monitoring of chlorophyll a content for different lakes commonly used
23、 method is the near-infrared and red band reflectance ratio, the reflectance of the red band and blue band ratio of 660 to 680 nm and 685 715 nm band near various the combination of the test, to identify the best inversion band combination. According to the results of correlation analysis, select 25
24、0m data ratio combination r 2 / r 1 as a factor built MODI S-based remote sensing monitoring model of chlorophyll a concentration. Through a linear regression analysis, the model is as follows: Cchl- a = 79. 386( r 2/ r 1) - 16. 092 ( 1) Wherein Cchl-a is concentration value of chlorophyll a; r 1, r
25、 2 is MODIS 250m reflectance of the bands 1,2. 500m and 1000m data is not highly correlated with the chlorophyll a concentration factor, these two sets of data, 7multiple regression modeling attempts. Of the 500m data 1 7 band and chlorophyll a concentration Backward law multivariate regression anal
26、ysis, the establishment of the model are as follows: Cchl-a = - 46824. 4r1 - 6954. 2 r 2 + 44719. 6 r 4 + 11582. 7r 7 - 10. 5 Wherein r1, r 2, r 4, r 7 is 1,2,4,7 reflectance of . MODIS 500m resolution band 3. Accuracy validation of image processing results Compared with the chlorophyll a concentrat
27、ion of ground truth value in 2006.4.3 and transit remote sensing image fusion in 2006.4.2. As can be seen from Figure 5, the green area of the waters of the ultimate fusion of image color depth and chlorophyll concentration monitoring values showing a good correlation between the high and low, and t
28、he images also reflect the effect of cyanobacteria concentration gradient. Higher chlorophyll concentration distribution at the edge of the area on the east coast, north shore and central West Bank region (green area), southwestern and central regions of the lower regions of the chlorophyll concentr
29、ation (blue area). The image fusion method has better visual effects of chlorophyll concentration monitoring. 84. Conclusion The green area of the waters of the ultimate fusion of image color depth and chlorophyll concentration monitoring value showing a good correlation between the high and low, an
30、d the images also reflect the effect of cyanobacteria concentration gradient. The fusion process, different times of the image fusion results comparable advantages dopant will be carried out in the remote sensing monitoring business, as the Dianchi chlorophyll remote sensing monitoring and mapping s
31、hortcut method. The fusion method not only has good visual effects, but also can better reflect changes in chlorophyll concentration effect, as the chlorophyll a concentration of Remote Sensing Quantitative image model. Reference 1 YE Yong. Preliminary Study on Water Quality Monitoring of the Disper
32、sed Water Source Based on Remote Sensing TechnologyJ. ，Journal of Anhui Agri Sci. 2012，40(2):1206 1208 2 ZHU Ling- ya,WANG Sh-i xin,ZHOU Yi,YAN Fu- li,YANG Long- yuan. Determination of Chlorophyl-l a Concentration in Taihu Lake Using MODIS Image Data J 2006. 84:25-29 93 ZHAGN Yun1, FENG Xuezhi1, MA Ronghua2, YONG Bin3. Advances in Inland Water Chlorophyll-a Extraction by Remote Sensing J. Ecological environment.139-143. 4 Lahet F, Ouillon S. A Three-component model of ocean colour and its application in the Ebro River Mouth area J. Remote Sensing of Environment, 2000(72): 181-190.
