[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121304-en":3,"doc-seo-121304-105":30,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},121304,1099514067438,"River Wang","https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542",8,"Research & Report","Machine Learning to Aid Pollinator Monitoring of Endangered Plants - Research Overview","Rare plants often rely on pollinators, creating high conservation value for both species and pollination interactions. The study focuses on UCSB’s North Campus Open Space, where the federally endangered Salt Marsh Bird’s Beak is pollinated by Crotch’s Bumblebee and other insects. Camera deployments captured thousands of images, but many were discarded for framing and color, and early timing limited bumblebee observations. A synthetic image generator was developed and a preliminary PyTorch model distinguishes empty versus occupied images with limited accuracy, with plans to expand training data and improve reliability.","UC Santa Barbara  \nPosters  \nTitle  \nMachine Learning to Aid Pollinator Monitoring of Endangered Plants  \nPermalink  \n[https://escholarship.org/uc/item/17n6f098](https://escholarship.org/uc/item/17n6f098)  \nAuthors  \nRosillo, Ethan Seltmann, Katja Evelyn, Chris  \nPublication Date  \n2025-05-12  \n[eScholarship.org](eScholarship.org) Powered by the California Digital Library  \nUniversity of California  \nMachine Learning to Aid Pollinator Monitoring of Endangered Plants  \nEthan Rosillo, Katja Seltmann, & Chris Evelyn  \nCheadle Center for Biodiversity and Ecological Restoration, University of California, Santa Barbara  \nIntroduction  \nRare plants are often dependent on pollinators to maintain their populations. In some cases, the pollinators themselves are of high conservation value, resulting in a system that has inherently high value for conservation. The North Campus Open Space (NCOS) at UCSB is home to such a system, where the Federally Endangered Salt Marsh Bird’s Beak (Chloropyron maritimum) is pollinated by recently listed Crotch’s Bumblebee (Bombus crotchii) along with other bumblebee, and insect species.  \nMonitoring pollination events is time-consuming, and researchers are looking to automated cameras to capture images of pollination events and machine learning to help extract pollinator identifications from the captured images.  \nResults  \nThe cameras successfully captured thousands of images, but many photos had to be discarded due to framing and color issues. No bumblebees were observed in any photos due to the first deployment being too late in the year.  \nThe synthetic image-generating script can create thousand of synthetic photos if given background images and cutout images of bees.  \nA preliminary version of the machine learning model was able to split the images (training vs. testing), train itself, then test itself for accuracy when given a folder of empty and occupied images. The predictions were often unreliable (\u003C50%), but this result is expected given the preliminary model was trained using a small image pool.  \nMethods  \nCamera Deployment  \nFour stakes were placed throughout a patch of Salt Marsh Bird’s Beak in NCOS. Throughout August and September 2024, custom cameras were mounted on these stakes and set to capture images every four seconds. Each camera captured thousands of images during eight deployments.  \nImage Sorting  \nImages were sorted using the application LabelMe. Images were labeled with the name of any invertebrate present and labeled‘Empty’ if none were visible.  \nSynthetic Image Generation  \nUsing Python, a script was written to take a cutout image of a bumblebee and paste it onto an ‘Empty’ image from one of the cameras. The script was looped to generated thousands of synthetic images using randomly selected background images and bee cutouts. Additionally, bee size, position, and orientation was randomized within selected parameters.  \nCreation of Machine Learning Model (in progress)  \nUtilizing PyTorch, a machine learning library, a Python script was created to train a model to identify empty images and images with bees. Synthetic images were used to train the model on images with bees.  \nReferences  \nKoch, Jonathan, et al. Bumble Bees of the Western United States. 2012 Spiesman, Brian J. , Claudio Gratton, Richard G. Hatfield, et al.  \n“Assessing the Potential for Deep Learning and Computer Vision to Identify Bumble Bee Species from Images.” Scientific Reports , vol. 11, no. 1, Apr. 2021, [p. 7580.](p. 7580. DOI.org)[ DOI.org](p. 7580. DOI.org) (Crossref),  \n[https://doi.org/10.1038/s41598-021-87210-1](https://doi.org/10.1038/s41598-021-87210-1) . Spiesman, Brian J. , Claudio Gratton, Elena Gratton, et al.“Deep  \nLearning for Identifying Bee Species from Images of Wings and Pinned Specimens.” PLOS ONE, edited by Alessandro Cini, vol.  \n19, no. 5, May 2024, [p. e0303383.](p. e0303383. DOI.org)[ DOI.org](p. e0303383. DOI.org) (Crossref), [https://doi.org/10.1371/journal.pone.0303383](https:/","cbCaiuSbuUhoxOpA","https://ap.wps.com/l/cbCaiuSbuUhoxOpA","pdf",503081,1,2,"English","en",105,"# Introduction\n# Methods\n## Camera Deployment\n## Image Sorting\n## Synthetic Image Generation\n## Creation of Machine Learning Model\n# Results\n# Discussion","[{\"question\":\"Why are pollinator monitoring efforts important for endangered plants in this study?\",\"answer\":\"Many rare plants depend on pollinators to maintain their populations, and in some systems the pollinators themselves are also conservation-relevant. Monitoring supports understanding and conservation of both plant and pollinator interactions.\"},{\"question\":\"What approach was used to collect and label images of pollination activity?\",\"answer\":\"Custom cameras captured images every four seconds during deployments. Images were sorted and labeled using LabelMe as either containing invertebrates or marked as 'Empty' when none were visible.\"},{\"question\":\"How does the synthetic image pipeline support the machine learning model?\",\"answer\":\"A Python script pastes bee cutouts onto background 'Empty' images to generate thousands of synthetic training examples. Bee size, position, and orientation are randomized within selected parameters to broaden variability for training.\"}]","Machine Learning to Aid Pollinator Monitoring of Endangered Plants - Research Overview | PDF",1785734989,5,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":28},"machine-learning-to-aid-pollinator-monitoring-of-endangered-plants-research-overview","",{"@graph":36,"@context":84},[37,53,67],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":21},"https://docshare.wps.com/document/","Document",{"item":48,"name":12,"@type":43,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/machine-learning-to-aid-pollinator-monitoring-of-endangered-plants-research-overview/121304/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why are pollinator monitoring efforts important for endangered plants in this study?","Question",{"text":74,"@type":75},"Many rare plants depend on pollinators to maintain their populations, and in some systems the pollinators themselves are also conservation-relevant. Monitoring supports understanding and conservation of both plant and pollinator interactions.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"What approach was used to collect and label images of pollination activity?",{"text":79,"@type":75},"Custom cameras captured images every four seconds during deployments. Images were sorted and labeled using LabelMe as either containing invertebrates or marked as 'Empty' when none were visible.",{"name":81,"@type":72,"acceptedAnswer":82},"How does the synthetic image pipeline support the machine learning model?",{"text":83,"@type":75},"A Python script pastes bee cutouts onto background 'Empty' images to generate thousands of synthetic training examples. Bee size, position, and orientation are randomized within selected parameters to broaden variability for training.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,108,113,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":29,"doc_module":4,"doc_module_name":46,"category_name":105,"show_sort_weight":106,"slug":107},"Comic",60,"comic",{"id":109,"doc_module":4,"doc_module_name":46,"category_name":110,"show_sort_weight":111,"slug":112},6,"Technology",50,"technology",{"id":114,"doc_module":4,"doc_module_name":46,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":46,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":46,"category_name":135,"show_sort_weight":29,"slug":136},19,"General","general"]