N12 Highway Park
Proof of Concept
This write up represents the findings and insights of the N12 Highway Park informal settlement, from data that is based from drone imagery. The site visit was conducted on Thursday the 21st November 2019 with the purpose of data acquisition required for the testing of a concept regarding the various outputs advance processing could deliver. Of the entire N12 Highway Park, only 25% of the area was covered to build the sample data and common data environment. The feasibility of using unmanned aerial systems as a means to capture data and processing it into a meaningful report will be covered in this “Proof of concept” write up.
This write up will also report on some of the findings that were observed from N12 Highway Park regarding allocation of resources and how those could be optimized for the benefit of the entire community. A comparison of the data and images where compared to Google earth as a start point given there was no previous data collected to compare to.
The data collected is the initial starting point for the common data environment, and the initial base for data harvesting of the area. With more data collected in the future and further processing more outputs will be derived, allowing us to decipher Prescriptive, Predictive and Descriptive results.
"Housing areas in South Africa developed along strict apartheid lines separating whites, Asians, coloureds and blacks into separate geographical areas according to the group areas act of 1950. For the past years South African cities have been 9 growing at an immense rapid rate. Responsible municipal authorities find it impossible to provide standard services above all to provide housing. The urban immigrants build themselves shelters, lacking all services such as water supply, sewage system or electricity, on the least wanted sites usually at the urban periphery.
Informal settlements are residential areas that do not comply with local authority requirements. They are unauthorized and are located upon land that has not been proclaimed for residential use. They exist because urbanization has grown faster than the ability of government to provide land, infrastructure and homes. Informal settlements are characterized by dwellings that are inadequate, Infrastructure that is inadequate, lack of effective government and management, environments that are unsuitable, population densities that are uncontrolled and unhealthy and they are areas of increasingly high risk with regard to health, fire and crime The rising cost of living has seen the increase in informal settlements in South Africa.
Despite the construction of low cost housing by the government the demand of housing is far more than the supply. According to the Census Report of 1996, 1 049 686 households in South Africa lived in informal settlements dwellings/shacks in squatter settlements at the time This implies that policy and professional efforts need to be re-directed towards the needs of the poor rather than the ideals of the middle class"
Extract from "Informal settlements in South Africa 2009 by Mr T. Chikoto
The data collection process was conducted by a drone with a digital camera payload. An array of pictures was taken from directly overhead with the camera at a perpendicular angle relative to the direction of flight. An image was captured every 2 seconds whilst the drone was traveling at 28.8 km/h. These pictures were captured at a height of 40m AGL.
After data collection, the pictures were processed and analyzed with software to provide some meaningful insight of the overall condition and resource allocation of the area.
Date: 21 November 2019
Arrival Time: 10:50
Aircraft: DJI Phantom 4 Pro
Sensor: 1 / 2.3” Effective pixels: 12.4M
Lens: FOV 94 20mm (35 mm format equivalent)
Visual Sensor: 12 MP
Max Video Bitrate: 60MHz
Solar System: Rooftop PV System
Site location: N12 Highway Park
The day presented cloudy skies and very windy conditions. The ambient temperature at the time the flights were conducted was approximately 27˚C. There was a strong wind from the North that was gusting frequently. The average wind was about 11 knots gusting at 19 Knots.
Figure 1a. Aerial view of the area flown for collecting the sample data
Figure 1 is an overview of the area that was flown to collect the sample data which is the eastern section of the settlement. The sample data was then future broken down to focus on a specific area to do advanced analytics to proof the capabilities of deliverables.
Figure 1b. Aerial view of the area flown for collecting the sample data
The individual images captured, where then processed to produce a digital model in a form of a point cloud. The point cloud forms the basis of the common data environment, where the data can be manipulated in a number of ways for desired outcomes. Of the digitized area only 4585.01 sqm was analyzed which was one section of the total sample data that was focused on. The total area was with a total of 75 households covered.
For the purpose of the Proof of concept, the digital data was manipulated to give more insight to the following:
Occupancy area covered
Average shack structure size
Occupancy area coverage
It goes without saying that from the onset, it is notable that the area is densely populated with an average of 4 structures occupying approximately 250 sqm.
Figure 2b. Overhead view highlighting the comparison between total land allocation in an informal settlement and average township household
Average shack structure size
With limited building land, it is only imperative that the total area a shack would cover is limited. An average shack was 26.66 sqm, with some being below the 10 sqm mark. Of the 75 shacks analyzed the biggest was 88.35 sqm (value could be subject to a number of shacks utilizing single roof coverage)
Figure 3. A graph depicting the shack sizes
There was a lack of vegetation throughout the community, with only a few households cultivating trees. Of the 75 houses analyzed, there were 8 trees, which are translated to one tree per 9.78 households.
With regard to the allocation of resources, the items studied are the chemical toilets allocated to the people of N12 Highway Park. Of the 75 shacks that were studied, we can conservatively estimate that an average household comprises of 2 adults and 1 child which translates to 225 individuals. From the data it was noted that there were only 13 toilets allocated to all those 75 shacks. The ratio of every toilet to persons is 18 individuals for a single chemical toilet.
The dumping area spotted on the sample data did not have any volume of dumping, although there was still some rubbish seen in the area. It could be that the refuse had been collected a couple of days prior the flight.
Figure 4: Just to the right of the image what seems to be the common refuse dumping area.
The number of shacks that had access to satellite TV network such as DSTV from the 75 shacks was a total of 13. The majority of the other shacks had normal aerials connected on the roofs suggesting they only had access to free frequency television provided by the South African Broadcasting Corporation
The yards of the shack are inconsistent and therefore encroach on to the roads. In some instances, what was deemed as roads was 4.7m wide. Therefore two cars approaching each other would rather pose a challenge for both vehicles. The traveling efficiency was rather hampered by the width of the roads and other factors as well such as the general condition of the surface. As there was rain the night prior to the flight, accumulation of the mud and some parts of the road covered with puddles suggested that there was lack of watershed or a none existence of a water reticulation system.
With more data collected, changes can be tracked and comparisons conducted. Above is the comparison of Satellite imagery provided by Google Earth compared to collected drone data
With Drone data that has been converted into a point cloud format, a lot of data outputs can be derived to the data for more insight into specific items. Once more data is collected on a frequent basis, it can be harvested into big data where a machine learning element can be applied.
With machine learning, we will be able to learn more about the growth of the informal settlement, pre-existing and post conditions, updated maps, human movement, just to mention a few. All of these insights can be deciphered in a dashboard tracking all pre-set changes.