Rapid Spatial Analytics
Mohsen Azadbahkt, P2.02
I have proposed an approach for restoration of target response from small-footprint full-waveform airborne LiDAR data (as seen to the right). The obtained high resolution signal eases separation of very closely located targets within the laser path, which is beneficial in many aspects (e.g. DEM enhancement).
The proposed method results in a calibrated signal, independent of the LiDAR instrument. For example, the averaged reflectivity over some asphalt targets is in a close agreement with those from the literature. Compared to the state-of-the-art approaches for full-waveform processing, adjustment of prior information is not required by the user. Regularization methods have been investigated in this research, with the capability of alleviating the noise level by assigning small weights to small singular values. The inclusion of new constraints in such methods allows us to better control the result.
The investigation of the calibrated radiometric features has shown that selected targets can be categorized satisfactorily without making use of geometric information. This is of significant importance, especially in the case of targets with almost the same height level.
Patrizia Russo, P4.53
I have just come back from a study away period of 3.5 months in Italy in which I worked with experts in usability of user-interfaces. It was a very valuable experience as it provided me with knowledge in how to conduct usability evaluations. I will adopt this knowledge in Melbourne to gather information on planning professional requirements and expectations when using Planning Support Systems.
Through visiting planning organisations and conducting interviews with planning professionals I’m aiming at identifying usability criteria for increasing the adoption and use of Planning Support Systems. A further outcome of the study away period is a book chapter related to my research which should allow me to attend a conference in the US this year. I’m slowly seeing the light at the end of the tunnel as I’m writing chapter drafts of my thesis in view of my imminent 2nd year confirmation.
Josh Neville, P4.51
Gosh, time just keeps flying by! At the present point in time my head is constantly dipped towards my screen as I write frantically, producing a scoping report on social housing in Christchurch. The housing portfolio of the local government is being reformed and restructured in a rather innovative way to include NGOs; changing the way it will be funded, maintained, and redeveloped in the future.
I have also been doing some GIS work, looking for ways to identify government properties that are available to be redeveloped into housing. This task is difficult as the cadastral parcels and the rateable units don’t align. There is no joy quite like parcel and property data. As time allows, and when I need a break from this work, I continue on my thesis, with literature reviews, context writing, and working through the database I am building for Auckland city so I can begin to do analysis on willingness to pay, which will open up the potential for more writing. Fortunately, I still love what I am doing!
Chet Bing Tan, P3.02
Current workflows that include the use of geospatial web services are manual and require human analysis of the chaining process and intervention to ensure relevant output. On top of that, current geospatial datasets and web services are disparate, difficult to find and exposed. End users are typically required to know the exact location of these online resources and what they do. Current search capabilities also do not sufficiently expose web services and datasets.
My research aims to provide the ability to chain together multiple geospatial web services (termed Orchestration) automatically and intelligently using Semantic Web (Web 3.0) technologies. This will greatly assist in the efficiency, accuracy and productivity of the end user. Natural language processors and ontologies will be used to build the required Artificial Intelligence to automatically chain together the resources to produce useful output for the end user.
The latest challenge being tackled right now is extracting knowledge from unstructured documents (such as PDF) for use in an ontology to assist in the orchestration process.