natural resource management

Satellite/Ariel Images Classification and Segmentation

Aerial and satellite imagery gives us the unique ability to look down and see the earth from above. It is being used to measure deforestation, map damaged areas after natural disasters, spot looted archaeological sites, and has many more current and untapped use cases. We understand that the enormous and ever-growing amount of imagery presents a significant challenge.

Figure 1: Satellite/Ariel Images
There are not enough people to look at all of the images all of the time. That’s why we are building tools and techniques to allow technology to see what we cannot. This project encapsulates a workflow for using deep learning to understand and analyze geo-spatial imagery. This project is released under an Apache 2.0 license and developed in the open on GitHub. This allows anyone to use and contribute to the project. It can also provide a starting point for others getting up to speed in this area.
This project work on performing semantic segmentation on aerial imagery provided by UC Merced Land Datasets. In this project, we’ll discuss our approach to analyzing this datasets. We’ll describe the main model architecture we used, how we implemented it in Keras and Tensor flow, and talk about various experiments we ran using the Uc Merced Land data. We then discuss how we used other open source tools built this project to visualize our results.
Semantic Segmentation and UC Merced Land Datasets
Figure 2: A ResNet FCN’s semantic segmentation as it becomes more accurate during training.
The goal of semantic segmentation is to automatically label each pixel in an image with its semantic category. Using UC Merced Land Datasets to create a semantic segmentation of high resolution aerial imagery. Part of the datasets had been labeled by hand with 21 21 class UC Merced land-use Datasets (RGB): (a) agricultural, (b) airplane, (c) baseball diamond, (d) beach, (e) buildings, (f) chaparral, (g) dense residential, (h) forest, (i) freeway, (j) golf course, (k) harbor, (l) intersection, (m) medium residential, (n) mobile home park, (o) overpass, (p) parking lot, (q) river, (r) runway, (s) sparse residential, (t) storage tanks and (u) tennis court. The datasets contains 2100 images and is divided into a development set, where the labels are provided and used for training models, and a test set, where the labels are hidden and are used by the client to test the performance of trained models.
Figure 3: UC Merced Land Dataset Classes
Fully Convolutional Networks
There has been a lot of research on using convolutional neural networks for image recognition, the task of predicting a single label for an entire image. Most recognition models consist of a series of convolutional and pooling layers followed by a fully-connected layer that maps from a 3D array to a 1D array of probabilities.
Figure 4: Neural network architecture for recognition
The Fully Convolutional Network (FCN) approach to semantic segmentation works by adapting and repurposing recognition models so that they are suitable for segmentation. By removing the final fully-connected layer, we can obtain a “fully convolutional” model that has 3D output. However, the final convolutional layer will still have too many channels (typically > 512) and too low a spatial resolution (typically 8×8). To get the desired output shape, we can use a 1×1 convolutional layer which squashes the number of channels down to the number of labels, and then use bilinear interpolation to up sample back to the spatial resolution of the input image.
Despite having the correct resolution, the output will be spatially coarse, since it is the result of up sampling, and the model will have trouble segmenting small objects such as cars. To solve this problem, we can incorporate information from earlier, finer-grained layers into the output of the model. We can do this by performing convolution and up sampling on the final 32×32, 16×16, and 8×8 layers of the recognition model, and then summing these together.
Figure 5: Fully convolutional neural network architecture for semantic segmentation
The FCN was originally proposed as an adaptation of the VGG recognition model, but can be used to adapt newer recognition models such as ResNets which we used in our experiments. One advantage of the FCN over other architectures is that it is easy to initialize the bulk of the model using weights that were obtained from training on a large object recognition dataset such as ImageNet. This is often helpful when the size of the training set is small relative to the complexity of the model.
Experiments and Results
We ran many experiments, and the following are some of the most interesting. Each experiment was specified by a JSON file stored in version control, which helped keep us organized and makes it easier to replicate our results.

  Overall agricultural airplane baseball beach buildings
Validation 89.3 88.5 87.1 89.9 84.1 88.6
Test 92.7 89.7 90.7 91.8 86.8 90.4


  chaparral dense residential forest freeway golf course harbor
Validation 88.0 86.9 86.9 88.2 87.1 83.1
Test 91.2 89.8 87.8 91.9 87.9 86.9


  intersection medium residential Mobile home park overpass parking lot river
Validation 89.9 84.1 88.6 86.9 86.9 88.2
Test 89.7 90.7 91.8 91.8 86.8 91.9


  runway sparse residential storage tanks tennis court
Validation 86.8 89.1 88.6 85.7
Test 89.9 93.1 91.6 88.9

Scouting/Monitoring Plant Health and Field Conditions

The use of drones in almost every sector of the economy is growing fast, but drone usage in the agricultural industry is booming. According to some reports, the agricultural drone market is expected to grow from a $1.2 billion(USD) industry in 2019 to $4.8 billion in 2024. From scouting to monitoring, drone use will become more ubiquitous on large and small scale farms in a few short years. The information gathered by drones on farms is often used to better inform agronomic decisions and is part of a system generally referred to as ‘precision agriculture’.

Figure 1: Farm Image
In many areas, drone use has become an essential part of large scale precision farming operations already. The data collected from drones recording fields help farmers plan their planting and treatments to achieve the best possible yields. Some reports indicate that using precision farming systems can increase yields by as much as 5%, which is a sizable increase in an industry with typically slim profit margins.
In this project we will tell about how FukatSoft help to improve drone technologies are already being used on farms.
Scouting/Monitoring Plant Health

  • FukatSoft get data from drone imagery for monitoring plant health.
  • Drones equipped with special imaging equipment called Normalized Difference Vegetation Index (NDVI) use detailed color information to indicate plant health.
  • This allows farmers to monitor crops as they grow so any problems can be dealt with fast enough to save the plants. This Figure 2 illustrates simply how NDVI works.
  • FukatSoft also get data from drones using ‘regular’ cameras are also used to monitor crop health.
  • FukatSoft also use satellite imagery to monitor crop growth, density, and coloration, but accessing satellite data is costly and not as effective in many cases as closer drone imaging.
  • Drones fly close to fields, cloud cover and poor light conditions matter less than when using satellite imaging.
  • Satellite imaging may offer to the meter accuracy, but drone imaging is capable of producing accurate image location to the millimetre.
  • This means that after planting, areas with stand gaps can be spotted and replanted as needed, and disease or pest problems can be detected and treated for right away.

Figure 2: The Basic Principle of NDVI
Monitoring Field Conditions

  • Fukatsoft get data from drone for field monitoring that also being used to monitor the health of soil and field conditions.
  • Drones can provide accurate field mapping including elevation information that allow growers to find any irregularities in the field.
  • After getting information on field elevation FukatSoft analyses these data and determining drainage patterns and indicate wet/dry spots which allow for more efficient watering techniques.

Figure 3: Indicate of wet/dry spots
FukatSoft also working on precise application of fertilizers, eliminating poor growing spots and improving soil health for years to come.

Lahore Orange-line Project

Lahore, the capital of Punjab, Pakistan, is a dynamic center of commerce and government located in the middle of Punjab. Amongst the most rapidly growing cities in the Pakistan, Lahore is home to more than 12 million people.

Figure 1: Lahore City
Traffic congestion in Punjab’s capital is a growing concern for residents and visitors and a potential barrier to continued economic growth. A state-of-the-art, fully automated rapid transit system is a primary component of the authority efforts to facilitate robust economic growth while providing Lahore’s population with a high level of service.
Some statistics of Orange-line are mention below.

  • The Orange line is the first of the three proposed rail lines proposed for the Lahore Metro.
  • The line will span 27.1 km (16.8 mi) with 25.4 km (15.8 mi) elevated and 1.72 km (1.1 mi) underground.
  • The line will be served by 26 stations.
  • Is expected to handle 250,000 passengers daily.
  • Anarkali and Central stations will be underground
  • While the remaining 24 will be elevated.

Figure 2: Lahore Orange-line
FukatSoft was awarded a contract by the Lahore Development Authority to provide following solutions.

  • Program and collecting, reviewing and preparing spatial datasets.
  • These datasets are crucial to numerical groundwater model.
  • Designing thematic maps focused on geology, piezo metric contours, watertable status.
  • Also monitoring underlying aquifer dynamics over time.

Figure 3: Collecting, Reviewing and Preparing Spatial Datasets
Four main phases of workflow as shown in the workflow chart

  • Selection of datasets types.
  • Data collection and preparation.
  • Data harmonization.
  • Technical validation

Figure 4: Work-flow Chart

Geographic Information Center for Metropolitan Corporation Lahore

In the developed world spatial technologies such as Geographic Information Systems (GIS) and Remote Sensing are playing a pivotal role in natural resource management and its efficient utilization.

Figure 1: What is Geographic Information Systems (GIS)

To keep up with the pace of modern technologies, FukatSoft is offering short term training courses in GIS. If you are interested in environmental management, information technology, surveying, business management or just interested in geography you must consider enrolling in these courses. This will not only give you an edge over others but will also assist you in finding employment in this exciting field.
GIS and remote sensing rely on acquiring, analyzing and processing spatial data obtained from satellite images as well as from traditional sources such as aerial photographs and ground based surveys. FukatSoft is well equipped with the state of the art hardware, software & lab facilities. These courses will assist you not only to learn the theoretical aspects of remote sensing & GIS technologies but will also provide “hands on experience” by applying and working on real life problems.

Figure 2: GIS Courses by FukatSoft

Under Punjab Local Government Act 2013, Lahore is a metropolitan area and under the authority of the Metropolitan Corporation Lahore. The district is divided into 9 zones, each with its own elected Deputy Mayor. The Metropolitan Corporation Lahore is a body of those 9 deputy, as well as the city’s mayor – all of whom are elected in popular elections. The Metropolitan Corporation approves zoning and land use, urban design and planning, environmental protection laws, as well as provide municipal services.

Figure 3: Metropolitan Corporation Lahore

Metropolitan Corporation Lahore is the supervising ministry for all municipalities around the Lahore. Metropolitan Corporation Lahore contract with FukatSoft and we currently offering:

  • Design development and implementation of a centralized GI Centre is in progress.
  • Study of the current system is in progress.
  • Which will follow the actual development of the system.

Spray on Crops using AI Drone

FukatSoft, a company, now endeavors to address these crucial issues that the agricultural sector is facing. Already, it has covered about 5,000 acres of crops by targeted spraying of pesticides and fertilizers in Lahore and other cities.

Figure 1: Crops Image
“The farmers are suffering yield losses due to inadequate or improper fertilizer application. If you do more of this, it could impact soil health as well. We can tell apart a healthy plant from an unhealthy one.

Founded by a team of Pakistani UET Lahore Alumni, the company developed an intelligent, autonomous spraying drone targeted at agriculture. Using drones that can cover 40 km in an hour in low altitudes, the company gathers data and takes a few hours to analyses and map the crops for pesticide and/or fertilizer sprays. Another drone, taking the payload with it, does the spraying, using a predefined route.

The company used similar techniques to help the Metropolitan Corporation Lahore (MCL) identify larvae population in lakes and helped reduce spraying instances, saving time and money.

With RGB (red, green blue), hyper, multi-spectral cameras and powerful sensors tucked under its belly, the drone gathers huge amounts of data, analyses it to map nutrition deficiencies and diseases in a particular field. It then makes a guided sortie with the required payload and sprays the required amounts in targeted areas. And we take images of that area and send to application. With the help of application, we can see these targeted areas and also we shown these statistics in tabular or graphical view. The drone can make a low-altitude (as low as 5-6 ft.) flight to make targeted spray in a particular area.

Figure 2: Spray on crops via AI drone
Figure 3: Data gathering from crops
For input manufacturers, the company is also selling data that they gather while working on the fields. “What we need is large amounts of data and machine learning tools to make the solution work better,” CEO FukatSoft said.

CEO Fukatsoft says the data gathered by drones can be very useful to assess the risks and losses on farms that are insured or those who have applied for insurance.

“The Prime Minister of Pakistan makes it mandatory to use exponential technology, including remote sensing, in addition to drone imaging, to detect fraudulent claims and discrepancies,” he pointed out.

“Farmers can make insurance claims by capturing drone feeds as evidence. The data is useful for insurance companies too to estimate damages and cross-check the claims,” CEO FukatSoft said.

Ground water monitoring / Water Analysis for Lahore, Pakistan

For analysis of ground water level for the city of Lahore by utilizing current groundwater condition that is relying on underlying geology and aquifer system. By viewing the previous and current pattern of groundwater situation also produce some prediction for the future. Designing and developing hydro databases. The core part is the development and design of automated workflows using ETL and by writing queries which are optimized fully , it also includes the business workflows automation regarding GISystems. Also deployed telemetry modules for the continuous logging and transmission of data. We also deployed a server and set all of its infrastructure.