#Runtime_Terrors

https://youtu.be/QcNXWXemPCo

https://github.com/rohann201/Icar-O-Spector

Extant Scenario 

Nowadays, collecting accurate and meaningful information about the urban localities/environment with the maximum efficiency in terms of cost and time has become more relevant for urban, rural and city-level development planning and administration. Urban decision making increasingly requires urban land-use and land-cover maps generated from very high spatial resolution data. Remote Sensing is well advanced in terms of technologies and methods like multi-sensor, multi-scale and multi-temporal analyses are primarily limited to rooftop views of the buildings which represents an incomplete perspective of understanding the urban systems. During the past 50 years, measurement technologies in surveying and engineering have made significant developments by introducing the surveying methods like Electronic Total Station (ETS), GPS, Robotic total station and laser scanner.

In recent years, Unmanned Aircraft systems/UAVs/Drones have become the most advanced technology developed and a perfect platform for aerial photography, remote sensing studies, topographical surveys and mapping. These UAVs also make use of LIDAR and data derived from these active sensors capable of providing detailed 3D pint clouds from which detailed building structural information can be defined for urban planning and development measures. Many remote sensing and photogrammetry software packages are available, led by Leica, and drone data processing software like Drone Mapper, Photomodeler, Pix4D, Drone2Map from ESRIs platform, Agisoft Photoscan etc., can produce high-resolution orthomosaics,3D information inaccuracy that is equal to or better than conventional aerial photography. Standard GIS software and AutoCAD can be used to stitch and georeference drone aerial photos. Recent developments in Web-based technologies using GIS; the drone derived outputs can be displayed by representing 2D and 3D information through web portals/dashboards.

The United Nations (UN) developed its 2030 Agenda for Sustainable Development to provide a blueprint for peace and prosperity for people and the planet, now and into the future. Options for environmental management have improved dramatically over recent years. Sensors for air and water pollutants, and subsets of the electromagnetic spectrum, have become smaller, cheaper, and more bundled into comprehensive units. Aerial sensor platforms have also expanded in the form of low-altitude unmanned aerial vehicles (micro-drones), but their use in populated

spaces is increasingly restricted for safety and privacy reasons.

Despite the fact that humanity is equipped with state of the art technologies, however, there seems to be a lack of linkage in realizing the spinoffs that make community living better. According to the statistics by google trends, it is clearly evident that underdeveloping countries still could only hope to achieve the apex of development and urban development. Such countries are even the same places devoid of even the basic necessities. 

We know that most of the time, the goal of urban planning issues is to achieve economic prosperity only, but we must also strive to improve the lives of the population. In this article, we will review the most important problems which face urban planning issues, its negative effects on society and how to eliminate them.

Why do we aim to do?

Considering the above factors and the ultimate motive of democratizing development, we aim to develop an easy to use, easy to deploy and effective to implement tool that radically increases accessibility to such technologies, attempting to realize the UN sustainable goals and the global development dream we all have.

Our project aims to meet the following targets set by UN sustainable goals 2030:

1. Clean Water and Sanitation - By effective management of resources and proper segmentation of water resources from usage to wastage, our tool effectively aids in city planning around resources and not the other way around
2. Decent work and economic growth - By aiming at proper spaces in the concrete jungle, we not just are concerned for the environment, but also effectively plan and use the country resources that aids in economic growth on a macroscopic scale
3. Sustainable cities and communities - This is the pith of our study, to live and let our future generations live

• Target 11.1: By 2030, ensure access for all to adequate, safe and affordable housing and basic services and upgrade slums.
• Target 11.2: By 2030, provide access to safe, affordable, accessible and sustainable transport systems for all, improving road safety, notably by expanding public transport, with special attention to the needs of those in vulnerable situations, women, children, persons with disabilities and older persons.
• Target 11.3: By 2030, enhance inclusive and sustainable urbanization and capacity for participatory, integrated and sustainable human settlement planning and management in all countries.
• Target 11.6: By 2030, reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management.
• Target 11.7: By 2030, provide universal access to safe, inclusive and accessible, green and public spaces, in particular for women and children, older persons and persons with disabilities.

Key stakeholders 

It is required to have a better understanding of the stakeholders who may contribute or influence the prospective goal. According to the author, stakeholder analysis helps to assist management in dealing with stakeholder demands and in increasing the contribution from the stakeholders. Stakeholders of a sustainable city development project help in achieving the goal of implementation of such cities. Therefore there is a growing concern about the identification of stakeholders of an ambitious urban dev project to achieve the goal of success. 

Icar-O-Spector -- Leveraging the power of air and space to make urban living harmonious (Aravind)

Our solution (The Computer vision team)

Waste Distribution Analysis

Waste management is an element of environmental order marking which is one of the areas of sustainable development. The generation of waste is intrinsic to urbanization. According to the statistics cited by Agata et. al in her 2012 publication, Municipal waste in 2012 constituted 9% of all waste in Poland. The communal cleanliness and order maintenance act binds municipalities to organize the system of municipal waste management. Protection against this type of waste is conditioned by environmental, social and economic circumstances and for this reason, it becomes an element of cities' sustainable development. This article draws attention to the problem of municipal waste generated in urban areas in the context of sustainable

Development.

Therefore, it is important to organize in cities the so-called green infrastructure

understood as “a connected network of multifunctional, predominately unbuilt, space that supports both ecological and social activities and processes”. The elements of green infrastructure are: street trees, private and public gardens, parks, riparian zones along urban drainage lines, undeveloped ridges, and a variety of urban agricultural spaces such as food-and community-based gardens. They contribute to the urban ecosystem's good condition.

When it comes to effective urban planning with minimal to zero energy wastage, it is highly substantive that we go ahead with well established “Zero wastage City” Proposal.

In many cities, especially in the developing countries, the environment and, accordingly, the health of dwellers is degraded because the municipal waste management systems are not effective. Most waste is landfilled which challenges the cities to find areas suitable for it. Moreover, landfilling is a costly form of waste neutralization that is not socially accepted

and must meet strict legal and administrative requirements. That is why wherever the prevention of waste is not possible, waste recovery is the desired way of managing it. Municipal waste is where we find resources that can be recovered with the lowest energy and external means expenditure

Our work implements a deep-learning algorithm, known as Masked R-CNN. An important aspect of our work is Transfer Learning. Transfer Learning is very beneficial when a model, which has been trained on a large dataset, is found to be suitable for data similar to one’s own use case. The pre-trained weights are loaded, followed by loading the model. The model is then tested on custom data collected by us.

The result is an image which has masks drawn over the area of the detected waste, along with a tight fit bounding box around it. We calculate the area of the bounding box, which is an almost accurate representation of the amoebic area of the detected waste. The area of the entire image is also calculated, using the dimensions of the image. The resulting area of the tight fit bounding box is divided by the area of the entire image, and multiplied by 100 to get a percentage, which is the percentage of detected waste in a given input image.

Our implementation consists of a Waste Score, out of 100, the entirety of which is divided into two scores. One score is given to the count of the number of waste piles in a given community area. The maximum possible score for the count is taken as 50, for 15 piles of waste. The corresponding score for any number of piles below 15 can be calculated by basic unitary method. The second score is given to the percentage of the waste, divided by 2, to make it out of 50. The addition of the respective scores sums up to give a score out of 100, known as the Waste Score.

The image on the left side is the original image which is inputted to our model for testing. The image on the right is the output from the model, with the mask and the bounding box constructed. The percentage of detected waste in the image is 68.009%. The count of the waste piles is assumed as a placeholder for now, and counting the waste piles is a future prospect for the model. 

The quantity of waste produced per capita during a year serves as an index that can be used in the assessment of the level of sustainable development in terms of environmental governance. The generation of municipal waste is inherent in the functioning of urban communities. Yet it is possible to minimize its negative environmental impact through proper waste management systems which, apart from ecological effects, can also bring economic and social perks. That is why municipal waste management is a significant element of the sustainable development of cities

Water Distribution Study

Urban Water Management involves the fields of water supply, urban drainage, wastewater treatment and sludge handling. An urban water system is typically designed incrementally, along with the development of new urban areas. Amongst professionals within the fields of water supply, urban drainage and wastewater treatment - here summarized as urban water management - sustainable development has become an important topic. It is not clear whether urban water management in its present form is sustainable or not. The nitrate problems in the groundwater, the difficulties to recycle potentially limiting resources such as phosphorus

from wastewater back to agriculture and the obvious problems we encounter in exporting our technology to the third world are only some examples that motivate us to investigate sustainable technology in urban water management.

Considering the Water distribution study in an urban setting, we have identified the following key influencing factors

Entropy - Earth - The Earth is a closed system that only exchanges energy, but no material with the surroundings. So, our local actions have to sum up global sustainability. In regard to this, our local actions have to sum up to global sustainability.E.g. in a cold, humid climate, water might be an excellent sun-powered and sustainable means for transporting pollutants whereas the direct use of solar energy for transport might be

more appropriate in hot, water-scarce climates.

In order to create sustainable innovation in technology, we must not focus on the technology itself, but rather investigate the function or services provided by a given technology.To mitigate such problems, water managers and city planners often rely on implementing ‘sustainable urban drainage systems. Such measures must often be supplemented with large-scale measures to ensure sufficient levels of flood and environmental protection, particularly in already built-up areas.

Key Assumptions before study and modelling:

• Assuming that the large-scale infrastructures for stormwater management, wastewater treatment, and drinking water provision are designed in a structured decision-making process, the planning process will involve a visioning or brainstorming phase where design alternatives are developed, a socio-economic and environmental performance assessment, and an iterative improvement
• Given that the technical lifetime of urban water infrastructures is on the order of 30 to 100 years (even longer for the spatial layout), the key drivers for so-called deep uncertainties include uncertain projections of climatic change
• The existing urban drainage infrastructure was assumed to remain unchanged. In reality, urban development would lead to an upgrade of the existing water infrastructure
• At the locations where runoff from newly introduced areas was introduced into an existing combined sewer system, the runoff was assumed not to affect the peak flows in the system

Factors under consideration for our study

• Urban Hygiene
• Drinking-Water and Personal Hygiene
• Prevention of Flooding I Draining of Urban Areas
• Integration of Urban Agriculture into Urban Water Management
• Providing water for pleasure and for recreational aspects of urban culture

Traditionally, we have been used to express our possibilities of action in terms of money. With the notion of sustainable development, it has become clear that this concept is inadequate as long as the consumption of the environment is not included in the prices. It is necessary to determine the scale of the impervious areas and wastewater flows, as well as the parts of the sewer and natural water systems that are affected by the new developments

(Include in a very detailed manner all the technical aspects)

• We use openCV to detect and draw a contour around the water bodies present near the potential hotspots in a given location. Potential hotspots are any densely populated area present in the regions surveyed by the drone.The coordinates of this region are acquired from municipal data which is fed into the algorithm.
• A model is then built to calculate a score out of 100. This score includes proximity, availability, and utilization of that water body. Proximity is calculated based on the Euclidean distance of the nearest water resources from that given prime hotspot coordinates. Availability is the presence of a water body, or the count of it. 
• Utilization is calculated by assuming that an individual uses roughly 30 pounds of water in a day. Hence this could be multiplied by the total population of that hotspot, which is the total utilization of water in that region. The ratio of this value to the amount of water present in its nearby water bodies is taken into consideration.
• The input and the output to our code are as follows:

Traffic Distribution study

Transportation management is an important aspect of urban development and planning. Globally, urban areas are rapidly growing, as is the urban population. Urban planning has a big influence on how much a city encourages corporate growth, and transportation management plays a big part in that. Per capita expenditures and revenues, as well as a country's GDP, are heavily influenced by city transportation. If I may use my own city as an example, Chennai owes much of its commercial capital position to well-planned public transit. On the contrary, a city like Pune is also notorious for its inability to cope with a rapid and widespread expansion of IT companies due to poor city planning. Typically over 2 hours are spent in traffic, traffic dangers, and pollution levels, after which livability is determined by traffic modes and sizes in the city. 

Traffic being one of the key aspects when it comes to urban planning, some of the key points that we kept in mind for this study includes:

• Density of the road:
• This comprises of leveraging the google maps API to know the kind of road we are aiming at, the roads have been classified as 
• National Highways (NH); he main highways running through the length and breadth of a country
• State Highways (SH): The highways linking up with the national highways of adjacent states, district headquarters and important cities 
• Major District Highways (MDR): The important roads within a district serving areas of production and markets and connecting these places with each other
• Other District Roads (ODR): The roads serving rural areas of production and providing them with outlet to market centers, tahsil headquarters, block development headquarters, railway stations etc.
• Village Roads (VR): The roads connecting villages or group of villages with each other or with the nearest road of higher category are known as village roads.
• The roads in urban areas are further classified as:
• Arterial Roads.
• Sub- Arterial Roads.
• Collector Streets.
• Local Streets.
• Once the type of road is identified, we get the dimensions of the road and considering a length of 500-1000m depending on the height at which the drone is flying. 
• We use MOG based background subtraction based algorithms to extract the foreground from the background thereby calculating the number of vehicles on the road per hour and thereby also calculating the density of the road. In MOG algorithm, the pixels are in the same position in the video and are regarded as a single time segment, and the subsequent probability of the current pixels can be expressed with the weighted sums of the K-MoG function.
• Mathematically:

Wkt represents the weight of the kth Gaussian distribution at the time of t.

Mkt and skt represent are the mean and standard deviation of the kth Gaussian function respectively.

h(eta) represents the Gaussian probability-density function.

For a new frame, the observed value of each pixel is acquired and then compared with the K Gaussian model. The D standard deviation range of the kth Gaussian is considered and matched with the kth submodel.Then the first B distributions are considered as the background model and the rest is the foreground model.

foreground_objects = current_frame - background_layer

But in some cases, we cant get static frame because lighting can change, or some objects will be moved by someone, or always exist movement, etc. In such cases we are saving some number of frames and trying to figure out which of the pixels are the same for most of them, then this pixels becoming part of background_layer. Difference generally in how we get this background_layer and additional filtering that we use to make selection more accurate.

• A rough output from our model:
• Training the background subtractor:

• Post Filtering
• 
• 

• Based on the factors such as the type of road, the dimensions of road, the number of vehicles on the road, we average down the total to 100 and store the values on a per hour basis throughout the time period.   

Zephyrus Web Tool - An assistive platform for visualizing and analyzing data

Traffic, Waste and Water management is often implemented and studied in Urban and Developed countries. These projects usually take millions of dollars to get implemented. But what about the smaller parts of the world? What about the places that cannot afford such an amount? They also deserve the same tools and services that could make their environment better, safer and hygienic. Zephyrus was created to equally help the rich and poor demographic alike.

Zephyrus is a web tool that integrates the above mentioned Waste, Water and Traffic management algorithm and displays the graphical or pictorial output. As the location and tool required are selected, the website receives the aerial images and sends them to the corresponding algorithm, returning the output as graphs or pictures.

Future Prospects

Future prospects regarding the waste detection module include counting the number of waste piles in an image with the help of our model. Currently, the entire waste is detected as 1 pile and a mask and bounding box is drawn around the entire area. When the method to count the number of waste piles is included, the current placeholder waste pile count can be replaced by the actual counted value, and an accurate Waste Score can be determined. The formula for detecting the accurate Waste Score out of 100 is given by 

Waste Score = (Count*50/15) + (percentage(out of 100) / 2)

Future prospects of the Zephyrus:

1. Use the Google Earth API to get an accurate aerial view of a location and use that data to compute water and waste distribution and traffic analysis.
2. Add a tool to generate 3D graphs of output

NASA Worldviewer data

ESA data

Sentinel Datasets for drones

Online drone datasets - opensource

We started by clearly analyzing the problem statement and using the industry relevant stakeholder study to demarcate the key influencing factors to develop state of the art Computer vision tool and hosted in a website

NASA dataset

#sustainable#future#water#waste#drones#traffic#urbanplanning

This project has been submitted for consideration during the Judging process.