|
Article information
2016 , Volume 21, ¹ 1, p.40-48
Borzov S.M., Uzilov S.B.
Detection of faint anthropogenic changes in vegetation covers by multitemporal satellite images
Purpose. Development and testing of an automatic method for detection of shelterwood cutting areas on medium resolution multispectral satellite imagery. Methodology. The proposed algorithm involves pairwise comparisons of images available for the desired period of time along with the post-processing of changes in the generated map-charts collection for correction of the accepted decisions. Pairwise comparison of images uses a method based on the detection of changes due to blockbased (with bilinear interpolation) matching of the brightness of multi-temporal images and allocation of anomaly pixels based on the analysis for the diagrams of the brightness change. Additionally, the decision has three levels: yes, no, and maybe. The pixels in the direct spread histogram are deviated from the maximum, which is greater than one-and-a-half width of the half-decay of histogram of the reverse spread. They are marked in map-charts of changes as “maybe”, trebled ones are marked as “yes” and others are marked as “no”. At the stage of post-processing of collection of map-charts changes, generated by pairwise comparison of multi-temporal images, we determine the sharpness of changes and whether these changes were kept in the subsequent pictures. Findings. The method was demonstrated on the example of detection of shelterwood cutting areas based on multi-temporal images of satellite SPOT 4. If there are no such areas on a site of observation, all pixels with identical brightness of the basic image have similar brightness on the test image. Upon appearance of shelterwood cutting areas, brightness of channels R and SWIR increases in the corresponding sites so the spread of pixels brightness grows. The width of reverse spread (from the tested image to basic) is determined by hardware errors in registration system, as well as natural changes, and is used as value of tolerance. This allows for the automatization of image processing, eliminating the necessity for experimental selection of threshold values. Conclusions. The estimation of efficiency of the proposed method is made by comparison of detection results with a mask generated on the basis of the visual analysis of images. The example of processing images of satellite SPOT indicates that the use of the proposed post-processing procedures doubles the percentage of correct marks in the resulting map-chart of changes.
[full text]
Keywords: satellite imagery processing, change detection, shelterwood cutting
Author(s):
Borzov Sergey Mikhailovich
PhD.
Position: Head of Laboratory
Office: Institute of Automation and Electrometry SB RAS
Address: 630090, Russia, Novosibirsk, Academician Koptyug ave. 1
Phone Office: (383)330-90-33
E-mail: borzov@iae.nsk.su
SPIN-code: 7504-7810Uzilov Sergey Borisovich
Position: Student
Office: Institute of Automation and Electrometry SB RAS
Address: 630090, Russia, Novosibirsk, Koptyuga 1
Phone Office: (383)330-90-33
E-mail: uzilov@gmail.com
References:
[1] Shimov, S.V., Nikitina, YU.V. Technology of monitoring of deforestation with use of space pictures of the high spatial resolution. Geomatics. 2011; (3):47-52. (In Russ.)
[2] Rouse, J.W., Haas, R.H., Schell, J.A., Deering, D.W. Monitoring vegetation systems in the Great Plains with ERTS. Proceedings of 3rd Earth Resources Technology Satellite-1 Symposium, Greenbelt, NASA SP-351. 1973; (1):309-317.
[3] Paramonov, E.G., Malenko, A.A. Osnovy lesovodstva i lesoparkovogo khozyaystva: uchebnoe posobie [Basics and forest-park management:textbook]. Barnaul: AGAU; 2007: 170. (In Russ.)
[4] Townshend, J., Justice C., Gurney C., McManus, J. The impact of misregistration on change detection. IEEE Transactions on Geoscience and Remote Sensing.1992; (30):1054–1060.
[5] Bromiley, P., Thacker, N., Courtney, P. Non-parametric image subtraction using grey level scattergrams. Image and Vision Computing. 2002; 20(9-10):609–617.
[6] Borzov, S.M., Potaturkin, O.I. Detection of shelterwood cutting areas on multispectral high spatial resolution satellite imagery. Issledovanie Zemli iz kosmosa. 2014; (4):87-93. (In Russ.)
[7] Bartalev, S.A., Kuryatnikova, T.S., Stibig, H.YU. Methods for the analysis of time series of high-resolution satellite images for the assessment of logging in the taiga. Sovremennye problemy distancionnogo zondirovaniya Zemli iz kosmosa: Sbornik statey pod redaktsiey E.A. Lupyana i O.YU. Lavrovoj, T. II. Moscow: «GRANP polygraph»; 2005: 217-227.
Bibliography link:
Borzov S.M., Uzilov S.B. Detection of faint anthropogenic changes in vegetation covers by multitemporal satellite images // Computational technologies. 2016. V. 21. ¹ 1. P. 40-48
|
