**Detection of Water Pipes and Leakages in Rural Water Supply Networks Using Remote Sensing Techniques**

Diofantos G. Hadjimitsis, Athos Agapiou, Kyriacos Themistocleous, Dimitrios D. Alexakis, Giorgos Toulios, Skevi Perdikou, Apostolos Sarris, Leonidas Toulios and Chris Clayton

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/39309

**1. Introduction**

Water leakages have been a major problem for many regions around the world (Weifeng et al. 2011). However, monitoring such leakages is a difficult task since traditional field survey methods are costly and time consuming (Huang et al. 2010). Researchers from diverse scientific fields have studied this problem through the development of several techniques including radar technique, geophones, gas filling, and many others. Different conventional techniques such as acoustics, radioactive, electromagnetic, ground penetrating radar and linear polariza‐ tion resistance have been used over the years for water pipeline leakage detection (Skolnik, 1990; Heathcote and Nicholas, 1998; Hunaidi and Giamou, 1998; Eyuboglu *et al*., 2003; Burn *et al.*, 2001; Hadjimitsis, *et al*., 2009).

Remote sensing has been used for a wide range of applications including water management. Studies have shown promising results from its use for water leakage detection (Sheikh Naimullah, 2007). The uses of remote sensing techniques for water leakage detection are time and cost effective compared with traditional, intrusive methods, but their use is restricted due to their spatial resolution. The pipeline leakages occur along the length of the pipeline and the area affected may not be detectable by the satellite sensor as it depends on the pixel size and the density of the vegetation developed due to the presence of water.

Vegetation indices (VI) are the main form of satellite spectral data used for several applications. According to Agapiou *et al.* (2012a), VIs can be divided into five main categories according to equation or the use of each index, which include broadband indices, narrowband indices

(hyperspectral), leaf pigment indices, stress indices and water stress indices. They reported that VI can be simply divided according to the wavelength characteristics used in their formula (broadband and narrowband indices). Using airborne remotely sensed imagery, Pickerill and Malthus (1998) analyzed two known water leaks and found that different vegetation indices and single bands were required in order to identify each leak. The spectral profile of one leak responded best to a ratio of NIR to red reflectance, while in the other, NIR to red reflectance ratio was not useful in differentiating it from its surroundings.

**2. Study areas**

**2.1. "Southern Conveyor Project"**

**2.2.** *"Lakatameia"* **pipeline**

2010.

In this section, three different case studies are presented. In the first case study, a part of the *"Southern Conveyor Project"* is described; following, two case studies for the "*Lakatameia* " and the "*Choirokoitia -Frenaros*" water pipes are presented. In the first case study, the authors have focused to the detection of the actual footprint of the pipe while in the next two case studies, remote sensing techniques have been evaluated for the detection of water leakages. The "*Lakatameia*"is a pipeline which is currently not in use while the "*Choirokoitia -Frenaros*" is a major pipeline of Cyprus where three major leakages have been recorded between 2007 to

Detection of Water Pipes and Leakages in Rural Water Supply Networks Using Remote Sensing Techniques

http://dx.doi.org/10.5772/39309

157

Water resources development in Cyprus initially focused on groundwater and, until 1970, groundwater was the main source of water supply for both drinking and irrigation purposes. As a result, almost all aquifers were seriously depleted because of over pumping. In addition, seawater intrusion was observed in most of the coastal aquifers. The increase of population as well as the increase in the tourist and industrial activities have led to an increase in the demand

Under these conditions, the implementation of the *"Southern Conveyor Project"* was a necessity and a basic prerequisite for the further agricultural and economic development of the island. The *"Southern Conveyor Project"* is the largest water development project ever undertaken by the Government of Cyprus. The basic objective of the project is to collect and store surplus water flowing to the sea and convey it to areas for both domestic water supply and irrigation. Essentially, the project aims to support the agricultural development of the coastal region between Limassol and Famagusta, as well as to meet the domestic water demand of Limassol, Larnaca, Famagusta, Nicosia, and a number of villages. In addition it supports the tourist and industrial demand of the southern, eastern and central areas of the island. The project is able to supply 33 million cubic metres of water for the irrigation of 13 926 hectares and another 33 million cubic metres of water for domestic purposes (Cyprus Water Development Department,

An existing pipeline in the area of Lakatameia (central Cyprus) was selected to be used for the pilot study (Figure 2). The existing pipeline, with a length of less than 5 km, has been system‐ atically reported as problematic due to several leakages and is therefore no longer in use by local authorities. The waterpipe runs through both urban and rural areas (see Figure 2). A section of the pipeline with a length of over 2km and located in a rural area, has been used to apply the different remote sensing techniques for the detection of leakages. Since the existing waterpipe is not currently used, it was necessary to fill the pipe with water periodically in

for water and have created an acute shortage of potable water.

2000). In this case study, a part in the SE of Cyprus was examined (Figure 1).

order to observe the effectiveness of such remote sensing techniques.

Huang et al. (2009) used airborne multispectral remote sensing imagery with high-resolution imaging sensors in the visible, NIR and thermal infrared wavelengths and found that airborne multispectral imaging is a useful tool in the detection of irrigation canal leakage in distribution networks. They concluded that the analysis of the processed image data from red, NIR and thermal bands is highly consistent with the observations from field investigation. Images from individual bands, particularly from the thermal band, can help detect leakage from irrigation canals. The NDVI image, which combines the data from the red and the NIR bands, can help detect and correct errors observed on the thermal imagery.

On-site observation, which consists of data collection, periodical observations, and multivari‐ ate risk assessment analysis, is the most common technique of monitoring the water pipe network in Cyprus. However, this is difficult to accomplish with traditional methods since it is time consuming, expensive and monitoring is localized. Furthermore, part of the water network tends to be located in inaccessible areas, away from the main road network and urban areas. A complete geoinformation system providing the exact location, characteristics and relevant data for the water mains does not exist, making the leakage monitoring procedures even more challenging.

This paper presents the results from a project which combines different remote sensing technologies for the detection and monitoring of water leakages for water utility systems located in open fields in Cyprus. Two case studies areas were evaluated using freely distrib‐ uted Landsat 7 ETM+ satellite images and ground spectroradiometric data. In addition, a low altitude system was deployed to observe these pipelines from different heights.

Finally, different remote sensing techniques have been used evaluated as in the detection of leakage from a major water pipe in Cyprus ("*Southern Conveyor Project*"). Although significant efforts have been made to detect possible water leakages, as shown above, the detection of the water pipe itself it still problematic . This is because such water pipes networks are commonly mapped in a digital form (e.g. GIS environment). However, in most cases the digital location of the water pipe does not fully correspond with the real world, since many obstacles during the construction can be arise and therefore the route of the proposed pipe can change.

In order to explore further the capabilities of remote sensing –beyond the detection of water leakages- the authors have applied several algorithms for the detection of buried water pipes. The detection of buried features is well established procedure in archaeological research since buried anthropogenic remains can be found using remote sensing techniques (Agapiou *et al*., 2010, 2012b; Sarris *et al*., 2013). Indeed, soil marks or crop marks related with water pipes can be used, in a similar approach, for mapping the real footprint of a pipe network.
