Evaluating of applying image fusion and optimizing techniques for improving the scale parameter of segmentation in object-based image analysis approach

Segmentation is one of the basic method of the information extraction within the object-based image analysis (OBIA) approach. This process separates initial and main objects which are basis for OBIA. According to this, generating appropriate segments plays an important role for performing high accurate object-based classification. Within this research, we aimed to employ multi spectral and spatial satellite images including: IRS, Quick Bird and Spot5, for the purpose of image fusion and optimizing the scale of segmentation. For this to happen, Multi-resolution segmentation approach was performed based on various satellite images with different spatial resolution. As that, spatial information of Quick Bird and panchromatic band of IRS and Spot5 images, alongside spectral resolution of Spot5 (red band, especially) and Quick Bird, have a significant impact in increasing the contrast of image and improve the quality of segmentation, subsequently. The results of this research, indicate the importance of applying spatial information for optimizing the scale of segmentation. In addition, results confirmed that object based image fusion techniques can be employed for integrating different spatial resolution of satellite images. It also turned out that integrating lower spatial resolution with high spatial resolution is an efficient procedure for improving segmentation quality. The results of research, are great of importance for identifying different segmentation approach of object-based classification. The achieved results are also important for executive departments such as Natural resource, agriculture, etc. in light of presentation appropriate approach for rapid extraction of information from satellite image.