Documentation, preservation, maintenance and restoration of cultural heritage as well as their buffer zones are considered as important tasks of the people and the government. It is imperative for administrators and managers of civil and development projects to adhere to it. To do this, having the exact engineering maps is essential. The detailed maps are the basis for the maintenance, identification, rehabilitation and archiving of the National Cultural Heritage. Today, with the rapid growth of urbanization and the expansion of modern technologies that accelerated and facilitated construction, more attention was paid to the identification and preservation of historical monuments. A principal approach for monitoring and ensuring the preservation of the cultural heritage and its buffer zone is the scientific documentation. The main focus of the present research is on the development and design of an optimal method for the automatic detection and documentation of the Qanat (Kariz) and its environment, which is one of the engineering feats, including the unique cultural heritage of Iran, by extracting and recording spatial information. In this research, data fusion methods were used for the integration of aerial and satellite imagery in order to identify automatic water wells. Two types of integration have been made to achieve the appropriate data for Kariz detection and documentation: 1. Integration of aerial and satellite imagery; 2. Integration of the extracted features from the fused image into decision-making. Satellite and aerial images of the region in Eslamshahr have been merged into Ehler's method. After analyzing each of the different fusion methods and the histogram of the images before and after fusing, and examination of the quantitative criteria, different radiometric characteristics of the wells of the aqueduct are extracted. These methods include applying the TC3 indices (with 62% success in identifying the desired pixels) and NDWI (with 62% success in identifying the desired pixels) and SAVI (with 52% success in identifying the desired pixels) and applying the segmentation algorithm on different image bands, the Ehler algorithm was 76% successful in identifying desirable complications. In the next step, to integrate at the decision level, two other layers of information (the layer of slope of the region and the layer obtained from the template matching algorithm with 54% success in determining the desired pixels) are extracted from the geometric properties, and along with the features obtained in the previous step, the stage of decision-making will begin. Fuzzy method has been used to integrate the results at the decision-making level. Finally, the properties of the Kariz system were detected with a better accuracy than 90%.
Based on the results obtained, this method is not optimal in all conditions. Therefore, it is recommended to use the fusion method at hybrid levels and instead of using a single procedure, in each layer of information, an optimal method to input to the next step is feature extraction. Ultimately, merging the extracted properties in different layers is applied. In the present study, integration was performed at the decision-making level based on fuzzy logics. To achieve optimal performance as a result of fusion, different layers, each with their own weight, come in with different coefficients. To decide, a combination of multi-layered neural network algorithm and fuzzy logic can be used which will be tested and evaluated in the next stages of the research.