Improved automatic impact crater detection on Mars based on morphological image processing and template matching

Abstract

Impact craters help scientists to understand the geological history of planetary bodies. The aim of this paper is to improve the existing methodology for impact craters detection in images of planetary surfaces using a new approach based on morphological image processing (MIP). The improved methodology uses MIP followed by template matching based on fast Fourier transform (FFT). In this phase, a probability volume is generated based on the correlation between templates and images. The analysis of this probability volume allows the detection of different size of impact craters. We have applied the improved methodology to detect impact craters in a set of images from Thermal Emission Imaging System onboard the 2001 Mars Odyssey Space probe. The improved methodology has achieved a crater detection rate of 92.23% which can be considered robust, since results were obtained based on geomorphological features, different illumination conditions and low spatial resolution. The achieved results proved the viability of using MIP and template matching by FFT, to detect impact craters from planetary surfaces.

Delineamento Automático de Crateras de Impacto em Imagens HRSC da Superfície de Marte

“…Após o refinamento da metodologia e a extração das características das estruturas, foi efetuada a avaliação do desempenho do processo de detecção e delineamento de crateras. A avaliação do desempenho foi realizada por meio de métricas, quantitativas e qualitativas, baseadas em pixels, propostas por Shufelt (1999), a qual tem sido muito utilizado em pesquisas de detecção de crateras DING;STEPINSKI, 2012;PEDROSA et al;. Dessa forma, foram consideradas as seguintes métricas: True Detection Rate ( ) ou taxa de detecção verdadeira, False Detection Rate ( ) ou taxa de detecção falsa, Branching Factor ( ) ou fator de ramificação e Quality Percentage ( ) ou percentual de qualidade, as quais são obtidas por meio das Eq.…”

“…Pesquisadores do mundo todo têm proposto diferentes metodologias com a finalidade de detectar e extrair diversas feições em corpos planetárias (BANDEIRA et al, 2011;STATELLA;SILVA, 2012;PALAFOX et al, 2017;WARDLAW et al, 2017). Dentre as diversas feições de interesse e as superfícies que vêm sendo alvo de investigações pela comunidade científica, destacam-se as crateras de impacto em Marte DING;STEPINSKI, 2012;PEDROSA et al, 2017;XIN et al, 2017). De forma geral, as abordagens de detecção e extração de feições, encontradas na literatura, se divergem em três pontos principais.…”

“…Se por um lado, muitos trabalhos com o objetivo de detectar crateras em superfícies planetárias já foram desenvolvidos, por outro, nota-se a necessidade de mais investigações e integrações de técnicas para sanar os problemas encontrados nos métodos desenvolvidos até o presente momento (SALAMUNIĆCAR et al, 2011;DING;STEPINSKI, 2012;PEDROSA et al, 2017). As dificuldades se referem principalmente ao fato de as crateras serem feições de difícil distinção em relação às demais estruturas.…”

Detection of Small Impact Craters via Semantic Segmenting Lunar Point Clouds Using Deep Learning Network

“…Di et al [22] adopted a novel variant of Haar-like features and a boosting algorithm to detect the square regions containing one crater. Pedrosa et al [23] used the watershed algorithm in infrared images, and they used the Template Matching to match the rim features of the craters.…”

A combined method of crater detection and recognition based on deep learning

“…The main methods include: Hough transform and its improved algorithm (Emami et al, 2019), genetic algorithm (Hong et al, 2012), radial consistency method (Earl et al, 2005) and template matching method (Cadogan, 2020; Lee et al, 2020). For example, one crater detection method based on morphological image processing method was proposed by Pedrosa et al (2017). This method mainly includes three steps: firstly, the candidate regions were determined by the morphological method, and then the possible crater areas were found by removing noise; secondly, the template matching method based on fast Fourier transform was used to CONTACT Haibo Li haibo021@sues.edu.cn establish the association between candidate regions and templates; finally, probability analysis was carried out to determine crater areas.…”