X
15541

L3Harris Geospatial Develops Automated Fence-Line Extraction Technique for Cadaster Updates

A pilot project in Queensland, Australia, demonstrated that the automated extraction of fence lines from airborne LiDAR and Imagery data is an effective method of enhancing the spatial accuracy of a digital cadaster database. The process, developed with the ENVI® and IDL® analysis software, shows that the extraction process can be automated to make it significantly faster and more cost effective than manual techniques.

A digital cadaster is a foundational spatial layer in the GISs that is relied upon by governmental agencies for a multitude of land management tasks. The cadaster typically contains the boundary locations for every property parcel owned by private and public entities within the jurisdiction. The accuracy of these boundaries in the official cadastral database can have important legal impact on taxation and property transactions.

Queensland is the second-largest state in Australia, covering more than 715,000 square miles of diverse terrain and home to 4.7 million people, many located in the city of Brisbane. The Queensland Department of Natural Resources, Mines & Energy (DNRME) has built and maintained its cadastral database (DCDB) primarily based on field surveys over the years.


“Depending on the location [within the state], and when the area was surveyed the cadaster can be off by as much as +/- 63 meters in some areas” said Dipak Paudyal of Esri Australia.

The state government sought an efficient means of updating the DCDB. Field surveying, although highly accurate, was often too expensive and time consuming for such a huge land area. Already an experienced user of remote sensing data, the DNRME wanted to compare and contrast the applicability of airborne imagery and LiDAR for the task.

Specifically, the idea behind the pilot project was to identify and map the locations of fence lines in multiple airborne data sets to determine which was more suitable for the application. The extracted fence lines would therefore be used as proxies for official property boundaries, and their mapped locations would be compared against the DCDB for updating where appropriate. The overall goal was to enhance the absolute accuracy of the parcel lines.

 

Extracting Fence Lines

With the basic project concept in mind, DNRME and Cooperative Research Centre for Spatial Information (CRCSI), Australia’s premier agency for spatial research (now called FrontierSI) partnered – to conduct the pilot. A contract was awarded to a three-company team: Esri Australia to provide mapping consultation, L3Harris Geospatial to develop and test the automated extraction process, and RPS Australia to capture airborne LiDAR and image data.

As the developed methodology was expected to operate on both urban and rural areas, Lidar data (2ppsm) and Imagery (10cm GSD) was available for a rural area near Toowoomba and a fresh capture was made for a semi-urban area at Morayfield both in Queensland. The data for Morayfield was captured by RPS over two different times and had different combined point density (24 ppsm and 64 ppsm), while the ground sampling distance (GSD) for the imagery were 10cm and 6 cm respectively.

 

In the first image, we see a LiDAR data surface model showing the presence of chainlink fences in a rural area.

In the second image, a fence line detection algorithm has been applied to the same surface model. The results are shown in red.

The Morayfield Lidar data was verified using differential GPS field survey which demonstrated that the RMS error for vertical accuracy of the Lidar data ranged between 3mm to 3cm. The GPS field survey coordinates was further used for image rectification. Digital cadastral data was available for both the areas. Further Lidar, imagery and cadastral data was made available for Adelaide, South Australia for testing of the algorithm in an urban area with different fence-line characteristics.

L3Harris Geospatial received the data sets for development of automated fence-line extraction techniques at its Broomfield, Colorado headquarters. Atle Borsholm, Senior Software Developer , took the lead in creating separate extraction methodologies for the LiDAR data and imagery, and then comparing the results to determine which worked better.

Borsholm worked in the ENVI image analysis package, an ideal environment for the pilot because it offers a variety of algorithms created for identification, mapping and extraction of specific features in various types of remote sensing data. In addition, ENVI allows the user direct access to the underlying IDL development language used to create the software. This enables the user to customize algorithms and combine them with existing ones to build complex geospatial data analysis workflows.

“The first step was to utilize existing LiDAR point classification tools in ENVI,” said Borsholm.

He explained that some built-in classification algorithms identified buildings, vegetation and terrain. There was no existing algorithm for fences, but the primary attribute of these features – an almost perfectly straight horizontal geometry – helped identify them in the point cloud. ENVI has an existing function for classifying linear features in LiDAR point clouds, which provided a good start for the extraction. The challenge was differentiating fences from other long, straight objects, like kerbs and powerlines.

“We used filters to eliminate these false positives by their height or Z values in the LiDAR data sets,” he said, explaining that any linear feature taller than eight feet was assumed to be a powerline. Features shorter than two feet were likely kerbs or road edges. Likewise, any linear feature wider than six inches was filtered out as non-fence.

The sides of buildings are often also straight lines, but the ENVI classification algorithms had no trouble distinguishing them from fence lines. One source of trouble in the urban/suburban LiDAR data was bushes and other vegetation growing against, and even on top of, fences. Borsholm wrote line fitting algorithms to separate vegetation, which usually has a rough texture in a point cloud, from the smooth fence segments the bushes were obscuring.

“We achieved good results classifying many types of fences – solid wood and chain link,” he said.

Not surprisingly, the denser LiDAR point clouds with 64 ppsm yielded better results in both the rural and suburban/urban applications. In the rural test area where fence lines separating large ranch properties are much longer than in suburban neighborhoods, the automated extraction achieved nearly the same positive results with low-density point clouds as it did with the high-density data set. The lengths of the ranch fences made them extremely easy to identify with the algorithms.

For extraction from the orthorectified imagery, Borsholm created a similar workflow using a combination of built-in and custom algorithms in ENVI. As with LiDAR data, the software had existing tools to automatically classify linear features in digital imagery, which worked well in the pilot project. On the downside, the image data lacked the Z values needed to create filters to eliminate certain features – powerlines and roof edges – by their heights.

“As a result, the automated extraction from the imagery returned more false positives than from the point clouds,” said Borsholm. “The LiDAR technique correctly identified more fences.”


Conclusion

The automated extraction of fence lines from remotely sensed data using customized algorithm workflows in ENVI and IDL showed great promise in the pilot project. In comparing the success of LiDAR versus imagery, the participants agreed the laser scan data was superior. The high-density point clouds correctly identified a significantly higher number of fences than the orthorectified image alone.

“The automated extraction technique was about ten times faster than manual extraction from LiDAR or image data,” said Borsholm.

Results of the project indicate that continued testing of the automated extraction technique should focus on different airborne acquisition parameters for the LiDAR data. Participants recommend capturing data by flying flight lines at different angles and during leaf-on/leaf-off seasons.

In terms of fence-line extraction from data sets covering the entire state, deep learning algorithms may prove more effective than the techniques developed in this pilot.