Search

Driving scene understanding is an essential technique for realizing autonomous driving. Although many large-scale datasets for autonomous driving have enabled studies in terms of driving scene understanding, unpaved dirt roads have not been considered as their target driving environment. Drivable region identification is important in dirt roads to prevent damage to the vehicle and passengers. Semantic segmentation of camera images and lidar point clouds have been used to recognize the environment around an autonomous vehicle. The road and objects are recognized by classifying image pixels and point clouds into semantic classes. In this study, we introduce a perception method for drivable region identification on dirt roads through fusion of image and point cloud semantic segmentation. Our approach includes an image semantic segmentation algorithm and a point cloud semantic segmentation algorithm, which we combine into a bird's-eye-view grid map format. To transform the point-wise drivable region identification results into the area-wise information, we adopt the alphashape algorithm. Speed improvements are made by using small proportion of drivable points, and the accuracy degradation is compensated by accumulating the perception results along the time.

Maraknya pemberitaan media sosial berjudul “75 Tahun Merdeka Tapi Katong Belum Merdeka” ditampilkan kegiatan warga pedalaman Buru Selatan sebagai pengguna sarana perlintasan menyeberangi sungai Nalbessy. Keselamatan Pengguna sarana dipertaruhkan karena beresiko putus, dan dapat berakibat fatal bila hanyut terbawa derasnya arus sungai. Maraknya pemberitaan media kemudian penulis mengkaji kebutuhan sarana perlintasan bagi warga pedalaman Buru Selatan untuk melintasi sungai Nalbessy. Kajian bertujuan: Menelusuri lokasi dimana kegiatan warga melintasi Sungai Nalbessy, dan mengalisis potensi pengguna sarana perlintasan berdasarkan data kependudukan. Menyajikan history jalan dan jembatan di sekitar Desa Liang dan Neath dalam bentuk stripmap penanganan ruas Namrole – Leksula. Metode penelusuran menggunakan aplikasi Google Earth, tahapan kajian terstruktur dan sistematis sesuai kaidah ilmiah bersumber dari data sekunder. Hasil kajian disimpulkan: Lokasi perlintasan Sungai Nalbessy terletak -3,7628º LS dan 126,5444º BT di Kec. Leksula, Kab. Buru Selatan Provinsi Maluku. Lokasi dimana oleh Dinas PUPR Provinsi Maluku telah direncanakan pembangunan jembatan Wai Nalbessy III. Ruas Namrole–Leksula melintasi 2 desa permukiman warga pedalaman Buru Selatan (Liang dan Neath). Penanganan jalan hingga tahun 2019, kondisi perkerasan masih belum seragam terdiri atas 3 jenis perkerasan: Hotmix 7.0 Km, Timbunan Pilihan 4.0 Km, dan Jalan Tanah 42.0 Km. Penanganan ruas jalan provinsi telah dibangun 3 jembatan dari rencana 33 jembatan. Perkiraan biaya pembangunan, waktu pelaksanaan jembatan bentang 50 meter Wai Nalbessy III, menurut jenis konstruksi, masing-masing: Rp. 4 M, 4 bulan untuk Jembatan gantung pejalan kaki simetris, Rp. 15 M, 10 bulan untuk Rangka Baja, dan Rp. 20 M, 1 Tahun untuk Jembatan beton/GTI.

Bulk transportation of tomatoes is a step with high occurrence of mechanical impacts, leading to increased losses and reduction in the quality of fruits arriving at the industry and reaching the consumer market. The present study aimed to evaluate variations in physical-chemical parameters associated with the quality of tomato fruits, occurred during bulk transportation, using an approach based on principal component analysis. The experiment was conducted in the municipality of Morrinhos - Goiás, Brazil. Firmness, titratable acidity, total soluble solids, pH and mass loss were evaluated in the fruits at three times: during transportation on dirt road; after transportation on dirt road and during transportation on asphalt-paved road; after transportation on dirt and asphalt-paved roads and arrival at the processing yard. Principal component analysis was used to identify which physical-chemical attributes were most affected by transportation and generate quality indices of tomato fruits according to the physical-chemical attributes. Variation in the attributes was explained by the first three principal components, which had cumulative explained variance of 78.37%. Titratable acidity, soluble solids and mass loss had the highest correlations with the most representative principal component, thus being the most affected attributes during the bulk transportation of the fruits.

The Tanah Baru Depok Road is 4.12 km in length consisting of two lanes connecting the Jakarta-Depok route. Current conditions with a road width of 5.2m during the rainy season on the right side of the new dirt road, there are certain parts of the STA, there are certain road conditions that are cracked and landslides. rainy season. The purpose of this research is to conduct monitoring and evaluation of the handling that has been done on the highway, as well as providing these results to make planning for the prevention of landslide prone roads. Stages of methods to be carried out during monitoring are identification of landslide hazard locations (not / moderate / already handled), identification of causes of landslides, identification of consequences, morphological analysis, geology, hydrology, and land use. Whereas in the evaluation stage is the analysis of the handling that has been done and the determination of new treatments. The results of monitoring and evaluation are there are 5 locations on STA Km 3 + 240 and STA Km 3 + 650 landslides, STA Km 1 + 970 and STA Km 2 + 450 potential landslides, STA Km 3 + 840 have dangerous roadside because there is no road divider. Countermeasures for landslide areas are carried out by installing a retaining construction on the right side of the road that intersects the river in the form of a cantilever wall. +840 mounted guardrail to prevent vehicles from entering the river bank..

Land degradation is a major environmental and social issue in temperate steppes. It is commonly determined from vegetation cover using remote sensing techniques. Steppes in eastern Mongolia are subject to resource extraction activities, such as mining and oil extraction, which affect land degradation. Recent technological progress in remote sensing has facilitated the acquirement of high-resolution data by, for example, the CubeSat satellite or unmanned aerial vehicles (UAV), providing data for detailed maps of vegetation cover and plant functional groups (PFGs). Traditional methods for monitoring vegetation cover often face typical scale issues, such as the upscaling of vegetation parameters if plot-scale field measurements are integrated to satellite data. Here, we studied the spatial distribution of PFG using machine learning and a combination of field measurements, UAV imagery (spatial resolution: 2 cm), and PlanetScope multi-temporal imagery. We provide two products at two spatial resolutions: one for UAV data, which is restricted to comparatively small areas around field measurements, and one for PlanetScope, which covers large parts of northeastern Mongolia. The results showed that the overall accuracies of UAV classification were 91–95%, whereas those of PlanetScope were 78–95%. In integrating the classified UAV data to the PlaneScope data, our proposed model minimized the scale issue that often impedes classification. Importantly, our findings revealed that the ecological effects of dirt road and railroad extended up to 60–120 m into the adjacent, otherwise less degraded steppe vegetation. A comparison between burned and unburned areas in different years indicates that wildfires affect the composition of PFG in reducing the fractional cover of graminoids and forbs, and that increasing cover of bare ground leads to a distinct and patchy mosaic of different vegetation types.

Understanding how exposure to particulate matter impacts human health is complex. Personal exposure is a function of the pollution concentrations measured at any given place and time. The health impacts of this exposure are, amongst other factors, determined by how high pollutant concentrations are and what enters the body. This study considered data gathered in the winter of 2013 in a low-income community on the Mpumalanga Highveld, South Africa, which is a geographical area known for its high air pollution levels. Time-activity data collected by GPS monitors worn by individuals in the community were used to understand in which microenvironments people spend most of their time. Eight days’ worth of ambient, indoor and personal particulate matter measurements were paired with individual GPS positioning data for one study participant. We identified pollutant concentrations where the person spent time and how much particulate matter was potentially inhaled in specific micro-environments. Participants spent time in five main micro-environments: (highest rank first) inside a house, directly outside a house, on a dirt road, on a tar road, and on an open field. Exposure to particulate matter concentrations in these micro-environments exceeded the National Ambient Air Quality Standards. Highest exposure was measured inside the dwelling and directly outside the dwelling. When comparing directly- and indirectly derived time-weighted potential intake doses, directly derived intake doses were higher and more likely to represent particulate matter concentrations inhaled by the participant. This study suggests that people living in communities on the Mpumalanga Highveld are exposed to unacceptably high air pollution levels in places in which they spend most of their time. Direct exposure and intake dose assessments are an essential element of environmental health studies to supplement data collected by stationary monitors.

Mae Cham district is located only 156 kilometers from the center for Chiang Mai, but the road trip would take at least 3 hours on the meandering mountains road. The first dirt road that provides access between Mae Cham in Chiang Mai was completed in 1963. The result of the road network and access to the global market has both positive and negative impacts on the people pf Mae Cham. The access to Mae Cham has transformed the food production system in Mae Cham from being self sufficient to becoming a part of the national and international food chain. The crops on the field have changed from rice to cash crops like animal feed corn and from local vegetables to commercial crops like cabbages, to feed into the national and international food chains. The consumption of local food has been transformed as well. The local dishes remain the same, while the added ingredient such as monosodium glutamate, fish sauce, seasoning powders that have increase the chance of high blood pressure, heart disease and diabetes.

In the last decade, a significant activation of water erosion was detected in the Priol’khonie (Lake Baikal). The comparison of satellite and aerial images of 1982–1986 and 2016 years shows that the length and density of the dirt road network have increased by 12.8 times during the period of 35 years. Recreational development, unsystematic laying, and exploitation of dirt roads led to the significant increase of linear erosion and the formation of numerous erosion forms. Linear erosion develops in most cases along roads laid along the thalwegs of the ravine network. Weak substrate and sparse vegetation (sparseness is aggravated by intensive grazing) predetermine rapid development of water erosion. Thus, in most cases, the increase of the erosion network is caused by anthropogenic impact. The development of forms of linear erosion leads to the disturbance of the integrity of modern landscapes of the Olkhon region, which significantly reduces their stability and recreational potential.

In the Eastern Mongolian Steppe Ecosystem, one of the major driving factors of human-induced land degradation is the expansion of road networks mainly due to intensifications of oil exploration and exploitation. Modern remote sensing data and machine learning techniques provide the possibility to detect and monitor long-term changes of land use such as dirt road and linear infrastructure in oil exploitation area. This study determined changes of land use/cover in the period 2005–2018 and estimated effects of dirt roads on surrounding grasslands ecosystems using multi-temporal satellite data with medium to high spatial resolution covering 20,000 km2 in eastern Mongolia. The ecological impacts of dirt roads and oil extraction sites were investigated using field vegetation photo-monitoring data, existing land use/cover maps and additional NDVI analysis. As a result, this paper describes current and historical land disturbance and its impact on grasslands ecosystem in Eastern Mongolia Steppe.

The majority of Africa's parks and conservation areas have a vast road network, facilitating motorized vehicle game viewing. These roads have an influence that is both road type‐ and species‐specific, on the surrounding ecosystem. Due to their higher traffic volumes, we hypothesized that tar roads and their immediate surrounds within the Kruger National Park, South Africa, are avoided to a greater extent by medium‐to‐large mammals than comparable dirt roads in the park. We systematically recorded the presence of medium‐to‐large mammal species from our vehicle, recording data at 401 tar and 369 dirt road stops in the Kruger National Park. In addition to species presence, we also estimated the proximity of animals to the road, as well as herd sizes. Our results indicate an equal likelihood of viewing the commonly recorded medium‐to‐large mammal species from both road types. The likelihood of observing larger herd sizes was also similar between tar and dirt roads for the three most commonly observed species, African elephant (Loxodonta africana), impala (Aepyceros melampus), and plains zebra (Equus quagga), and the likelihood of viewing impala and zebra close to the road also did not differ between tar and dirt roads. However, elephant was observed more often close to tar roads, compared to dirt roads. We interpreted this as the result of potentially increased woody cover associated with more water runoff in close proximity to tar roads compared with dirt roads. Our results not only have ecological significance, supporting the notion that many of the park's species are habituated to human infrastructure, but also management implications, informing park officials about the influence of road traffic and road type on wildlife distributions., The likelihood of viewing the common mammal species in the Kruger National Park (KNP), South Africa, is equal from tar and dirt roads. The likelihood of viewing large herds of elephant (Loxodonta africana), impala (Aepyceros melampus), and zebra (Equus quagga) is equal from tar and dirt roads in KNP. The likelihood of close‐up sightings of impala and zebra is equal from tar and dirt roads, but close‐up elephant sightings are more likely next to tar than dirt roads at KNP.

An aging man ambles down a dirt road, walking slowly away from the camera. He seems to be alone, until we realise he’s stalking an emu, which crosses to the other side of the track. He stops and tu...

It is a sweltering June day when we drive through the Bears Ears National Monument. The dry desert scrub oak of the western slope gives way to ponderosa, grassland, and aspen groves. Flowers line the curving dirt road and the car thermometer shows a 10 °F dip. We begin our backpacking trip at the top of Woodenshoe Canyon and will end near the bottom of Dark Canyon. Accompanied by five of our student athletes and my sister, my partner and I begin our trek in the deep shade of hundred-year-old aspens and ponderosas. The cicadas sing and, with their transparent wings, alight on our backpacks.

Understanding images by predicting pixel-level semantic classes is a fundamental task in computer vision and is one of the most important techniques for autonomous driving. Recent approaches based on deep convolutional neural networks have dramatically improved the speed and accuracy of semantic segmentation on paved road datasets, however, dirt roads have yet to be systematically studied. Dirt roads do not contain clear boundaries between drivable and non-drivable regions; and thus, this difficulty must be overcome for the realization of fully autonomous vehicles. The key idea of our approach is to apply lightweight non-local blocks to reinforce stage-wise long-range dependencies in encoder-decoder style backbone networks. Experiments on 4,687 images of a dirt road dataset show that our transitional asymmetric non-local neural networks present a higher accuracy with lower computational costs compared to state-of-the-art models.

The demand for a variety of situational data from the traffic environment and its participants has intensified with the development of applications in Intelligent Transport Systems (ITS). Among these data, the road surface type classification is one of the most important and can be used in the entire ITS domain. For its widespread application, it is necessary to employ a robust technology for the generation of raw data and to develop of a reliable and stable model to process these data in order to produce the classification. The developed model must operate correctly in different vehicles, under different driving styles and in different environments in which a vehicle can travel. In this work we employ inertial sensors, represented by accelerometers and gyroscopes, which are a safe, non-polluting, and low-cost alternative, ideal for large-scale use. We collect nine datasets with contextual variations, including three different vehicles, with three different drivers, in three different environments, in which there are three different road surface types, in addition to variations in the conservation state and presence of anomalies and obstacles such as potholes and speed bumps. After data collection, these data were used in experiments to evaluate various aspects, such as the influence of the vehicle data collection point, the analysis domain, the model input features, and the data window. Afterwards we evaluated the learning and generalization capacity of the models for unknown contexts. In a third step, the data were used in three Deep Neural Network (DNN) models: LSTM-based, GRU-based, and CNN-based. Through a multi-aspect and multi-contextual analysis, we considered the CNN-based model as the best one, which obtained an average accuracy between the data collection placements of 94.27% for learning and 92.70% for validation, classifying the road surface between asphalt, cobblestone or dirt road segments.

The technological growth and accessibility of Unoccupied Aerial Systems (UAS) have revolutionized the way geographic data are collected. Digital Surface Models (DSMs) are an integral component of geospatial analyses and are now easily produced at a high resolution from UAS images and photogrammetric software. Systematic testing is required to understand the strengths and weaknesses of DSMs produced from various UAS. Thus, in this study, we used photogrammetry to create DSMs using four UAS (DJI Inspire 1, DJI Phantom 4 Pro, DJI Mavic Pro, and DJI Matrice 210) to test the overall accuracy of DSM outputs across a mixed land cover study area. The accuracy and spatial variability of these DSMs were determined by comparing them to (1) 12 high-precision GPS targets (checkpoints) in the field, and (2) a DSM created from Light Detection and Ranging (LiDAR) (Velodyne VLP-16 Puck Lite) on a fifth UAS, a DJI Matrice 600 Pro. Data were collected on July 20, 2018 over a site with mixed land cover near Middleton, NS, Canada. The study site comprised an area of eight hectares (~20 acres) with land cover types including forest, vines, dirt road, bare soil, long grass, and mowed grass. The LiDAR point cloud was used to create a 0.10 m DSM which had an overall Root Mean Square Error (RMSE) accuracy of ±0.04 m compared to 12 checkpoints spread throughout the study area. UAS were flown three times each and DSMs were created with the use of Ground Control Points (GCPs), also at 0.10 m resolution. The overall RMSE values of UAS DSMs ranged from ±0.03 to ±0.06 m compared to 12 checkpoints. Next, DSMs of Difference (DoDs) compared UAS DSMs to the LiDAR DSM, with results ranging from ±1.97 m to ±2.09 m overall. Upon further investigation over respective land covers, high discrepancies occurred over vegetated terrain and in areas outside the extent of GCPs. This indicated LiDAR’s superiority in mapping complex vegetation surfaces and stressed the importance of a complete GCP network spanning the entirety of the study area. While UAS DSMs and LiDAR DSM were of comparable high quality when evaluated based on checkpoints, further examination of the DoDs exposed critical discrepancies across the study site, namely in vegetated areas. Each of the four test UAS performed consistently well, with P4P as the clear front runner in overall ranking.

On November 17, 2015 07:10 UTC an earthquake of Mw 6.5 induced slope failures mainly at the western part of the island of Lefkada, Ionian Sea, Greece. The most characteristic one is the Okeanos site deep-seated landslide (near village Athani) that affected the man-made environment; part of the site, including a dirt road, moved downwards along the coastal cliff and a luxury hotel suffered heavy damage. The goal of this study is to map in detail the Okeanos landslide and to provide relevant quantitative data i.e. structural data and volume of both the accumulated and removed material. In order to achieve this, a comparison between pre-earthquake and post-earthquake point clouds generated from aerial imagery and UAV-acquired photos took place. Using Structure from Motion, the steep coastal cliff site was surveyed in detail, even in areas that are difficult to approach. Results from this study show that the total landslide area covers a surface of ~36,000 m2, while a removed volume of ~93.000 m3 of landslide material was evaluated using change detection techniques. In addition, fieldwork and survey data revealed that the sliding surface was a pre-existing fault surface upon which the geological material moved downwards the coastal cliff. The sliding surface dips west and is exposed along a 115 m long and 25–30 m high section scarp. As a conclusion, this study underlines the importance of a thorough geological study in order to identify landslide prone areas prior to any building activity, and shows that a UAV survey is capable to provide adequate information during a post-earthquake field survey for co-seismic landslide mapping and modeling.

In the Republic of Korea, one of the biggest threats to amphibians is habitat modification such as urbanisation and land conversion. With the loss of natural habitats, rice paddies play an important role as substitute habitats for amphibians that originally inhabited wetlands. However, since the 70’s, traditional rice agriculture has been modernised, leading to an increase in the number of concrete ditches and roads bordering rice paddies. This modernisation could have affected the distribution and density of amphibians. In this study, we investigated the preferred position, based on the advertisement calls for two treefrog species (Dryophytes japonicus and D. suweonensis), in relation to different types of borders such as natural ditch, concrete ditch, one-lane dirt road and vegetation. The results show that treefrogs seem to avoid rice paddies with concrete ditches, and with no ditch, which provided no resting microhabitat. The sides of the paddies preferred by the two treefrog species were the ones with vegetation of 30 cm wider or higher, while the two species seemed to avoid the side of paddies with roads. Our results are important for the conservation of anuran species in rice paddies in general as it highlights the need for vegetated areas, preferentially along natural ditches.