Description

ABSTRACT

The project proposes to detect an aircraft in satellite remote sensing image using shape features for accurate detection and tracking. High resolution multispectral satellite images with multi-angular look capability have tremendous potential applications. Here the system involves an object tracking algorithm with three-step processing that includes moving object estimation, target modeling, and target matching. This recognition system involves dimensionality reduction, segmentation and aircraft identification with templates. The satellite remote sensing image dimensionality is reduced using principal component analysis. Here, Histogram probability thresholding is used to detect the desired object from background. Connected component analysis is used here to label the segmented image for grouping similar objects. It is used to extract the local object shape descriptors for identifying desired target. Correlation measurement is used for measuring similarity between two different object region features. This is used to locate the aircraft region for tracking it and it shows that reliable and compatible method for this process. Potentially moving objects are first identified on the time-series images. The target is then modeled by extracting both spectral and spatial features. In the target matching procedure, template will be used as matching model to recognize an aircraft for accurate detection. Final simulated will be demonstrated the capability of object tracking in remote sensing images with help of used approaches.

EXISTING METHOD

• Global thresholding method
• Fourier descriptors and Edge detection
• K-nearest neighbour method for matching

DRAWBACKS

• Detection of object is not optimal for all lighting conditions of an image.
• Lack of representing shape of object regions.
• Low recognition accuracy.

PROPOSED METHOD

Automatic aircraft detection from high resolution remote sensing images using shape analysis based on,

• Otsu(Histogram probability) thresholding with connected component analysis and Correlation measurement

BLOCK DIAGRAM

METHODOLOGIES

• Principal component analysis
• Histogram probability thresholding
• CC analysis and Local region descriptors
• Correlation measurement

ADVANTAGES

• It provides better accuracy of segmenting objects
• Correlation is useful to detect desired object with help of templates rather than euclidean distance.
• Flexibility and better compatible approaches in recognition

APPLICATIONS

• Surveillance
• Defense area and Navigation

SOFTWARE REQUIREMENT

• MATLAB 7.8 or above versions

 RESULT

REFERENCES

[1] S. A. Dudani, K. J. Breading, and R. B. Mcghee, “Aircraft identification by moment invariants,” IEEE Trans. Comput., vol. C-26, no. 1, pp. 39–46, Jan. 1977.

[2] L. Fu, Y. Peng, and L. Kun, “Research concerning aircraft recognition of remote sensing images based on ICA Zernike invariant moments,” CAAI Trans. Intell. Syst., vol. 6, no. 1, pp. 51–56, Feb. 2011.

[3] Z. Yanan and Y. Guoqing, “Plane recognition based on moment invariants and neural networks,” Comput. Knowl. Technol., vol. 5, no. 14, pp. 3771– 3778, May 2009.

[4] J.-W. Hsieh, J.-M. Chen, C.-H. Chuang, and K.-C. Fan, “Aircraft type recognition in satellite images,” Proc. Inst. Elect. Eng.—Vis. Image Signal Process., vol. 152, no. 3, pp. 307–315, Jun. 2005.

[5] S. Dapei, Z. Yanning, and W. Wei, “An aircraft recognition method based on principal component analysis and image model-matching,” Chinese J. Stereol. Image Anal., vol. 14, no. 3, pp. 261–265, Sep. 2009.

[6] L. Ke, W. Runsheng, and W. Cheng, “A method of tree classifier for the recognition of aircraft types,” Comput. Eng. Sci., vol. 28, no. 11, pp. 136– 139, 2006.

[7] N. Dalal and B. Triggs, “Histograms of oriented gradients for human detection,” in Proc. IEEE. Comput. Vis. Pattern Recog., Jun. 2005, pp. 886–893.

[8] G. Peyré, “Sparse modeling of textures,” J.Math. Imag. Vis., vol. 34, no. 1, pp. 17–31, May 2009.