Search NASA⌕ Search

DOE OSTI · 2205564

Real-time stored product insect detection and identification using deep learning: System integration and extensibility to mobile platforms

Abstract

Not Available

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Badgujar, Chetan M. (ORCID:0000000276982843), Armstrong, Paul R., Gerken, Alison R., Pordesimo, Lester O., Campbell, James F.. 2023-12-01. Real-time stored product insect detection and identification using deep learning: System integration and extensibility to mobile platforms. https://doi.org/10.1016/j.jspr.2023.102196

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

Community composition and abundance of wild bees at row crop-grassland interfaces in west central Nebraska

Abstract Perennial mixed forb and grassland habitats are crucial to conservation of pollinators and connectivity of habitats in intensely farmed landscapes. This study aims to understand the effects of land use on the pollinator community by describing bee abundance, species richness and community composition in perennial conservation grasslands and adjacent annual row crops located in west central Nebraska. In 2022 and 2023, we collected and identified bees via sticky traps at 4 locations (center and edge of adjacent grasslands and crop fields) at 6 replicated sites. We collected 1,768 specimens from sticky traps, resulting in 70 species within 28 genera. Halictidae accounted for 84% of the specimens collected. Bee abundance was influenced by the simple effects of land use (grassland vs. crops), edge adjacency, and the month and year of collection. Differences in bee abundance within a collection date were found mostly in early 2022 (May and June) and late 2023 (September), when the crop center location was generally the lowest, with some evidence for spillover of bees from the grassland into the crop edge during the early summer months. Bee species richness was affected only by month and was not significantly different by land use and edge adjacency. Bee community composition overlapped across the 4 locations, although there were significant dissimilarities between crop fields and grasslands. Surveys of the plant community revealed very low abundance of blooming stems and plant taxonomic richness at crop locations for all sampling periods, while grassland locations were comparatively high and varied over time. Plant communities showed no overlap between crop field and grassland locations. Overall, we found that conservation grasslands, while not seeded specifically with pollinator-attractive forbs, provide diverse resources to support wild bee communities in west central Nebraska; crop edges may also provide non-plant resources such as nesting sites and irrigation water. Going forward, better understanding pollinator species composition and resource utilization relative to land use characteristics and drought conditions will allow for better tailoring of conservation efforts and management strategies in Nebraska and across the larger region.

Entomology↗

Exploiting Trap Type and Color for Monitoring Macadamia Felted Coccid Acanthococcus ironsidei (Williams) and Associated Parasitic Wasps in Macadamia Orchards in Hawai’i

Acanthococcus ironsidei (Williams) (Hemiptera: Eriococcidae) is an invasive pest of macadamia, Macadamia integrifolia, in Hawai’i, causing death to macadamia trees and decreased nut productivity. Monitoring relies on wrapping double-sided sticky tapes over tree branches to trap dispersing crawlers (i.e., mobile immature stage), but this is tedious for growers, especially in large orchards. From September to November 2022 and December 2022 to February 2023, at two commercial macadamia orchards on Hawai’i Island, the use of colored sticky cards was assessed for improving the monitoring of A. ironsidei and to investigate the Hymenopteran parasitoid complex that inhabits macadamia canopies. At each study site, four different colored sticky cards (yellow, lime green, dark green, and white) were placed on the lower canopy of five trees, and on each tree, a transparent double-sided sticky tape was deployed. At bi-weekly intervals, the sticky cards were replaced and re-randomized on each tree, and the double-sided sticky tapes were replaced. The results showed that the sticky cards captured both A. ironsidei crawlers and (winged) male adults, while the double-sided sticky tapes captured only crawlers. The trap color did not have significant effects on the captures of A. ironsidei male adults at the sites, while the captures of crawlers on sticky cards were lowest on the dark green sticky traps at one site. The captures of A. ironsidei adult males on white sticky traps were generally correlated with the number of crawlers captured on the double-sided sticky tapes. The parasitoid complex captured had disparities in the attraction to color; however, the yellow, lime green and dark green colors were seemingly more effective for monitoring Encarsia lounsburyi (Berlese & Paoli), a reported parasitoid of A. ironsidei. These results have useful practical implications for improved monitoring of A. ironsidei crawlers, male adults and associated natural enemies.

Entomology↗

Predicted range expansion of Prostephanus truncatus (Coleoptera: Bostrichidae) under projected climate change scenarios

Abstract The larger grain borer (Prostephanus truncatus [Horn] [Coleoptera: Bostrichidae]) is a wood-boring insect native to Central America and adapted to stored maize and cassava. It was accidentally introduced to Tanzania and became a pest across central Africa. Unlike many grain pests, P. truncatus populations can establish and move within forests. Consequently, novel infestations can occur without human influence. The objectives of our study were to (i) develop an updated current suitability projection for P. truncatus, (ii) assess its potential future distribution under different climate change scenarios, and (iii) identify climate variables that best inform the model. We used WALLACE and MaxEnt to predict potential global distribution by incorporating bioclimatic variables and occurrence records. Future models were projected for 2050 and 2070 with Representative Concentration Pathways (RCPs) 2.6 (low change) and 8.5 (high change). Distribution was most limited by high precipitation and cold temperatures. Globally, highly suitable areas (> 75%) primarily occurred along coastal and equatorial regions with novel areas in northern South America, India, southeastern Asia, Indonesia, and the Philippines, totaling 7% under current conditions. Highly suitable areas at RCPs 2.6 and 8.5 are estimated to increase to 12% and 15%, respectively, by 2050 and increase to 19% in 2070 under RCP 8.5. Centroids of highly suitable areas show distribution centers moving more inshore and away from the equator. Notably, the result is a range expansion, not a shift. Results can be used to decrease biosecurity risks through more spatially explicit and timely surveillance programs for targeting the exclusion of this pest.

Entomology↗