TY - JOUR
T1 - Improved Change Detection in Remote Sensed Images by Artificial Intelligence Techniques
AU - Sheoran, Snehlata
AU - Mittal, Neetu
AU - Gelbukh, Alexander
N1 - Publisher Copyright:
© 2021, Indian Society of Remote Sensing.
PY - 2021/9
Y1 - 2021/9
N2 - The remote sensed images carry large amount of crucial information. Image processing, a field of signal processing, helps in analysis of remote sensed data. One of the major processing areas is image segmentation with edge detection, which helps in segmenting an image into various sub regions. These regions identified from images, captured over long span of time can help in identification of change detection. This paper presents an application of nature-inspired algorithms viz.: Ant Colony Algorithm, Particle Swarm Optimization and Genetic Algorithm to optimize edge detection procedure. These methods have been implemented on a set of 15 satellite images and further enhancement is done by application of adaptive thresholding using Python. For qualitative analysis, entropy of each output image is computed. The comparison of computer results revealed that particle swarm optimization outperforms conventional methods, i.e., Sobel, Canny and Prewitt as well as ACO and GA. The PSO-based method is able to find more edges and presents far superior quality output images for further analysis with respect to change detection.
AB - The remote sensed images carry large amount of crucial information. Image processing, a field of signal processing, helps in analysis of remote sensed data. One of the major processing areas is image segmentation with edge detection, which helps in segmenting an image into various sub regions. These regions identified from images, captured over long span of time can help in identification of change detection. This paper presents an application of nature-inspired algorithms viz.: Ant Colony Algorithm, Particle Swarm Optimization and Genetic Algorithm to optimize edge detection procedure. These methods have been implemented on a set of 15 satellite images and further enhancement is done by application of adaptive thresholding using Python. For qualitative analysis, entropy of each output image is computed. The comparison of computer results revealed that particle swarm optimization outperforms conventional methods, i.e., Sobel, Canny and Prewitt as well as ACO and GA. The PSO-based method is able to find more edges and presents far superior quality output images for further analysis with respect to change detection.
KW - Artificial intelligence
KW - Change detection
KW - Particle swarm optimization
KW - Satellite images
KW - Swarm intelligence
UR - http://www.scopus.com/inward/record.url?scp=85107368976&partnerID=8YFLogxK
U2 - 10.1007/s12524-021-01374-x
DO - 10.1007/s12524-021-01374-x
M3 - Artículo
AN - SCOPUS:85107368976
SN - 0255-660X
VL - 49
SP - 2079
EP - 2092
JO - Journal of the Indian Society of Remote Sensing
JF - Journal of the Indian Society of Remote Sensing
IS - 9
ER -