Níckolas Santana

Dr. Níckolas Santana

Universidade de Brasília - UnB
orcid

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 February 2022 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
CSCristiano Rosa e Silva
Cristiano Rosa e Silva
Anesmar Olino de AlbuquerqueAnesmar Olino de Albuquerque
Anesmar Olino de Albuquerque
Níckolas SantanaNíckolas Santana
Níckolas Santana
RGRenato Guimaraes
Renato Guimaraes
Panoptic segmentation combines instance and semantic predictions, allowing the detection of countable objects and different backgrounds simultaneously. Effectively approaching panoptic segmentation in remotely sensed data is very promising since it provides a complete classification, especially in areas with many elements as the urban setting. However, some difficulties have prevented the growth of this task: (a) it is very laborious to label large images with many classes, (b) there is no software for generating DL samples in the panoptic segmentation format, (c) remote sensing images are often very large requiring methods for selecting and generating samples, and (d) most available software is not friendly to remote sensing data formats (e.g., TIFF). Thus, this study aims to increase the operability of panoptic segmentation in remote sensing by providing: (1) a pipeline for generating panoptic segmentation datasets, (2) software to create deep learning samples in the Common Objects in Context (COCO) annotation format automatically, (3) a novel dataset, (4) leverage the Detectron2 software for compatibility with remote sensing data, and (5) evaluate this task on the urban setting. The proposed pipeline considers three inputs (original image, semantic image, and panoptic image), and our software uses these inputs alongside point shapefiles to automatically generate samples in the COCO annotation format. We generated 3400 samples with 512 × 512 pixel dimensions and evaluated the dataset using Panoptic-FPN. Besides, the metric analysis considered semantic, instance, and panoptic metrics, obtaining 93.865 mean intersection over union (mIoU), 47.691 Average (AP) Precision, and 64.979 Panoptic Quality (PQ). Our study presents the first effective pipeline for generating panoptic segmentation data for remote sensing targets.
1

21 April 2022 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
Anesmar Olino de AlbuquerqueAnesmar Olino de Albuquerque
Anesmar Olino de Albuquerque
Níckolas SantanaNíckolas Santana
Níckolas Santana
RGRenato Guimaraes
Renato Guimaraes
DBDíbio Leandro Borges
Díbio Leandro Borges
Vehicle classification is a hot computer vision topic, with studies ranging from ground-view to top-view imagery. Top-view images allow understanding city patterns, traffic management, among others. However, there are some difficulties for pixel-wise classification: most vehicle classification studies use object detection methods, and most publicly available datasets are designed for this task, creating instance segmentation datasets is laborious, and traditional instance segmentation methods underperform on this task since the objects are small. Thus, the present research objectives are as follows: first, propose a novel semisupervised iterative learning approach using the geographic information system software, second, propose a box-free instance segmentation approach, and third, provide a city-scale vehicle dataset. The iterative learning procedure considered the following: first, labeling a few vehicles from the entire scene, second, choosing training samples near those areas, third, training the deep learning model (U-net with efficient-net-B7 backbone), fourth, classifying the whole scene, fifth, converting the predictions into shapefile, sixth, correcting areas with wrong predictions, seventh, including them in the training data, eighth repeating until results are satisfactory. We considered vehicle interior and borders to separate instances using a semantic segmentation model. When removing the borders, the vehicle interior becomes isolated, allowing for unique object identification. Our procedure is very efficient and accurate for generating data iteratively, which resulted in 122 567 mapped vehicles. Metrics-wise, our method presented higher intersection over union when compared to box-based methods (82% against 72%), and per-object metrics surpassed 90% for precision and recall.

29 November 2018 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Níckolas SantanaNíckolas Santana
Níckolas Santana
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
RGRoberto Arnaldo Trancoso Gomes
Roberto Arnaldo Trancoso Gomes
RGRenato Guimaraes
Renato Guimaraes
Fires associated with the expansion of cattle ranching and agriculture have become a problem in the Amazon biome, causing severe environmental damages. Remote sensing techniques have been widely used in fire monitoring on the extensive Amazon forest, but accurate automated fire detection needs improvements. The popular Moderate Resolution Imaging Spectroradiometer (MODIS) MCD64 product still has high omission errors in the region. This research aimed to evaluate MODIS time series spectral indices for mapping burned areas in the municipality of Novo Progresso (State of Pará) and to determine their accuracy in the different types of land use/land cover during the period 2000–2014. The burned area mapping from 8-day composite products, compared the following data: near-infrared (NIR) band; spectral indices (Burnt Area Index (BAIM), Global Environmental Monitoring Index (GEMI), Mid Infrared Burn Index (MIRBI), Normalized Burn Ratio (NBR), variation of Normalized Burn Ratio (NBR2), and Normalized Difference Vegetation Index (NDVI)); and the seasonal difference of spectral indices. Moreover, we compared the time series normalization methods per pixel (zero-mean normalization and Z-score) and the seasonal difference between consecutive years. Threshold-value determination for the fire occurrences was obtained from the comparison of MODIS series with visual image classification of Landsat Thematic Mapper (TM), Enhanced Thematic Mapper Plus (ETM+), and Operational Land Imager (OLI) data using the overall accuracy. The best result considered the following factors: NIR band and zero-mean normalization, obtaining the overall accuracy of 98.99%, commission errors of 32.41%, and omission errors of 31.64%. The proposed method presented better results in burned area detection in the natural fields (Campinarana) with an overall accuracy value of 99.25%, commission errors of 9.71%, and omission errors of 27.60%, as well as pasture, with overall accuracy value of 99.19%, commission errors of 27.60%, and omission errors of 34.76%. Forest areas had a lower accuracy, with an overall accuracy of 98.62%, commission errors of 23.40%, and omission errors of 49.62%. The best performance of the burned area detection in the pastures is relevant because the deforested areas are responsible for more than 70% of fire events. The results of the proposed method were better than the burned area products (MCD45, MCD64, and FIRE-CCI), but still presented limitations in the identification of burn events in the savanna formations and secondary vegetation.

21 December 2018 in Fire
Verified
Journal Article
Open access
PDF file
Níckolas SantanaNíckolas Santana
Níckolas Santana
Fire is one of the main modeling agents of savanna ecosystems, affecting their distribution, physiognomy and species diversity. Changes in the natural fire regime on savannas cause disturbances in the structural characteristics of vegetation. Theses disturbances can be effectively monitored by time series of remote sensing data in different terrestrial ecosystems such as savannas. This study used trend analysis in NDVI (Normalized Difference Vegetation Index)–MODIS (Moderate Resolution Imaging Spectroradiometer) time series to evaluate the influence of different fire recurrences on vegetation phenology of the Brazilian savanna in the period from 2001 to 2016. The trend analysis indicated several factors responsible for changes in vegetation: (a) The absence of fire in savanna phytophysiognomies causes a constant increase in MODIS–NDVI, ranging from 0.001 to 0.002 per year, the moderate presence of fire in these areas does not cause significant changes, while the high recurrence results in decreases of MODIS–NDVI, ranging from −0.002 to −0.008 per year; (b) Forest areas showed a high decrease in NDVI, reaching up to −0.009 MODIS–NDVI per year, but not related to fire recurrence, indicating the high degradation of these phytophysiognomies; (c) Changes in vegetation are highly connected to the protection status of the area, such as areas of integral protection or sustainable use, and consequently their conservation status. Areas with greater vegetation conservation had more than 70% of positive changes in pixels with significant tendencies. Absence or presence of fire are the main agents of vegetation change in areas with lower anthropic influence. These results reinforce the need for a suitable fire management policy for the different types of Cerrado phytophysiognomies, in addition to highlighting the efficiency of remote sensing time series for evaluation of vegetation phenology.

Verified
Journal Article
Open access
PDF file
Níckolas SantanaNíckolas Santana
Níckolas Santana
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
RGRoberto Arnaldo Trancoso Gomes
Roberto Arnaldo Trancoso Gomes
RGRenato Guimaraes
Renato Guimaraes
Monitoring of fire-related changes is essential to understand vegetation dynamics in the medium and long term. Remote sensing time series allows estimating biophysical variables of terrestrial vegetation and interference by extreme fires. This research evaluated fire recurrence in the Amazon and Cerrado regions, using Moderate Resolution Imaging Spectroradiometer (MODIS) albedo time series, enhanced vegetation index (EVI), gross primary productivity (GPP), and surface temperature. The annual aggregated time series (AAT) method recognized each pixel’s slope trend in the 2001–2016 period and its statistical significance. A comparison of time trends of EVI, GPP, and surface temperature with total fire recurrence indicates that time trends in vegetation are highly affected by high fire recurrence scenarios (R2 between 0.52 and 0.90). The fire recurrence and the albedo’s persistent changes do not have a consistent relationship. Areas with the biggest evaluated changes may increase up to 0.25 Kelvin/Year at surface temperature and decrease up to −0.012 EVI/year in vegetation index. Although savannas are resistant to low severity fires, fire regime and forest structure changes tend to make vegetation more vulnerable to wildfires, reducing their regeneration capacity. In the Amazon area, protection of forests in conservation units and indigenous lands helped in the low occurrence of fires in these sensitive areas, resulting in positive vegetation index trends.

7 August 2020 in International Journal of Wildland Fire
Verified
Journal Article
Closed access
PDF file
Níckolas SantanaNíckolas Santana
Níckolas Santana
Osmar Carvalho JúniorOsmar Carvalho Júnior
Osmar Carvalho Júnior
RGRoberto Arnaldo Trancoso Gomes
Roberto Arnaldo Trancoso Gomes
RGRenato Fontes Guimarães
Renato Fontes Guimarães
The Moderate Resolution Imaging Spectroradiometer (MODIS) products are the most used in burned-area monitoring, on regional and global scales. This research aims to evaluate the accuracy of the MODIS burned-area and active-fire products to describe fire patterns in Brazil in the period 2001–2015. The accuracy analysis, in the year 2015, compared the MODIS products (MCD45/MCD64) and the burned areas extracted by the visual interpretation of the LANDSAT/Operational Land Imager (OLI) images from the confusion matrix. The accuracy analysis of the active-fire products (MOD14/MYD14) in the year 2015 used linear regression. We used the most accurate burned-area product (MCD64), in conjunction with environmental variables of land use and climate. The MCD45 product presented a high error of commission (>36.69%) and omission (>77.04%) for the whole country. The MCD64 product had fewer errors of omission (64.05%) compared with the MCD45 product, but increased errors of commission (45.85%). MCD64 data in 2001–2015 showed three fire domains in Brazil determined by the climatic pattern. Savanna and grassy areas in semi-humid zones are the most prone areas to fire, burning an average of 25% of their total area annually, with a fire return interval of 5–6 years.