[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86106-en":3,"doc-seo-86106-105":30,"detail-sidebar-cat-0-en-105":91},{"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":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},86106,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Unsupervised Detection of Entry and Exit Regions for Camera-Agnostic Turning Movement Counts","Turning movement counts are essential for intersection-level traffic management, yet collection is largely manual because per-camera region annotation is costly. This paper introduces an unsupervised pipeline that derives entry and exit regions from raw vehicle trajectories obtained via object detection and multi-object tracking, with no manual annotation, calibration, or intersection-geometry prior. Instead of repeatedly clustering trajectories by similarity, it clusters initial and terminal points to form persistent region polygons and classifies new trajectories by point-in-polygon containment. Across 25 cameras and 10 UA-DETRAC sequences, parameter testing identifies three significant settings and achieves ~3% median classification error. Extended evaluation shows improved estimation with 60+ minute calibration clips and peak-traffic selection, outperforming two clustering baselines in stability and computational cost while trading higher median error.","Unsupervised Detection of Entry and Exit Regions from Vehicle Trajectories for Camera-Agnostic Turning Movement Counts  \nParikshit Singh Rathore*, Vishwajeet Pattanaik* and Punit Rathore   \nRobert Bosch Centre for Cyber-Physical Systems (RBCCPS), Centre for infrastructure, Sustainable Transportation and Urban Planning (CiSTUP) Indian Institute of Science (IISc), Bengaluru, India  \n{parikshits, vishwajeetp, [prathore](prathore}@iisc.ac.in)[}](prathore}@iisc.ac.in)[@iisc.ac.in](prathore}@iisc.ac.in)  \narXiv :2607 . 10949v1 [ cs .CV] 12 Jul 2026  \nAbstract—Turning movement counts are essential for intersection-level traffic management, yet their collection remains predominantly manual due to the cost of per-camera region annotation. This paper presents an unsupervised pipeline that identifies entry and exit regions directly from raw vehicle trajectories extracted via object detection and multi-object tracking, requiring no manual annotation, camera calibration, or prior knowledge of intersection geometry. Unlike trajectory clustering methods that classify individual trajectories using pairwise similarity and must be re-executed on every new batch, the proposed pipeline clusters initial and terminal point locations to produce persistent spatial region polygons that classify future trajectories by point-in-polygon containment at linear cost. The pipeline comprises six sequential steps, each with configurable parameters evaluated through a systematic statistical analysis spanning 19,152 pipeline executions across 25 surveillance cameras capturing dense heterogeneous traffic in Bengaluru, India, and 10 sequences from the UA-DETRAC benchmark dataset. Both parametric and nonparametric testing frameworks identify three consistently significant parameters and yield an empirically grounded recommended configuration. Under this configuration, the pipeline achieves a median classification error of approximately 3% across all 25 cameras, including 16 heldout locations, with GEH values within accepted engineering thresholds. Compared with two trajectory clustering baselines, the proposed pipeline exhibits greater stability across camera views and lower computational cost, at the expense of higher median error. Extended evaluation demonstrates that calibration clips of at least 60 minutes and peak-traffic selection further improve region estimation quality.  \nIndex Terms—Turning pattern, Transportation planning and design, Trajectories, Traffic networks, Smart cities, Computer vision  \nI. INTRODUCTION  \nTurning movement counts (TMCs) classify vehicles by their entry and exit paths at intersections and are among the most critical inputs for urban traffic management. Accurate TMCs underpin signal timing, capacity planning, and safety analysis [1] . Yet TMC data collection remains dominated by manual methods originating in the 1950s. The 2012 U.S. National Traffic Signal Report Card assigned a failing grade to traffic monitoring practices nationwide [2], and TMC data is typically collected for only one to two days per year per intersection due to prohibitive labor costs [3] . The resulting data scarcity limits the ability of transportation agencies to  \n*Both authors contributed equally to this work.  \nrespond to evolving traffic patterns, particularly in rapidly growing cities where intersection demand changes faster than manual surveys can track.  \nThe growing deployment of surveillance cameras in metropolitan areas worldwide presents an opportunity to close this data gap, though not through purpose-built traffic infrastructure. India’s Safe City project, for instance, is deploying tens of thousands of cameras across eight metropolitan cities for public safety rather than traffic monitoring. Bengaluru alone operates over 7,000 such cameras at junctions, markets, and transport hubs [4], while Delhi is expected to deploy approximately 10,000 under the same programme [5] . Similar surveillance expansions are underway across Southeast Asia, the Mid","cbCaijx5CZUWe9dh","https://ap.wps.com/l/cbCaijx5CZUWe9dh","pdf",14868753,2,1,21,"English","en",105,"# Introduction\n## Problem context and motivation\n## Limitations of manual data collection\n## Using surveillance-derived trajectories\n## Challenges in unsupervised, camera-agnostic region discovery","[{\"question\":\"What problem does the paper address in turning movement counts data collection?\",\"answer\":\"It addresses the high cost and bottleneck of manually delineating entry and exit regions per camera or intersection, which limits network-wide deployment and response to changing traffic patterns.\"},{\"question\":\"How does the proposed method avoid manual annotation and camera calibration?\",\"answer\":\"It builds entry and exit regions directly from vehicle trajectories extracted by object detection and multi-object tracking, without requiring manual region annotation, camera calibration, or prior knowledge of intersection geometry.\"},{\"question\":\"What is the core idea behind the pipeline’s clustering and classification strategy?\",\"answer\":\"Instead of clustering entire trajectories repeatedly by pairwise similarity, it clusters initial and terminal point locations to produce persistent spatial region polygons, then classifies future trajectories by point-in-polygon containment.\"}]",1784208557,53,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"unsupervised-detection-of-entry-and-exit-regions-for-camera-agnostic-turning-movement-counts","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":20},"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/unsupervised-detection-of-entry-and-exit-regions-for-camera-agnostic-turning-movement-counts/86106/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-25","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address in turning movement counts data collection?","Question",{"text":75,"@type":76},"It addresses the high cost and bottleneck of manually delineating entry and exit regions per camera or intersection, which limits network-wide deployment and response to changing traffic patterns.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the proposed method avoid manual annotation and camera calibration?",{"text":80,"@type":76},"It builds entry and exit regions directly from vehicle trajectories extracted by object detection and multi-object tracking, without requiring manual region annotation, camera calibration, or prior knowledge of intersection geometry.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the core idea behind the pipeline’s clustering and classification strategy?",{"text":84,"@type":76},"Instead of clustering entire trajectories repeatedly by pairwise similarity, it clusters initial and terminal point locations to produce persistent spatial region polygons, then classifies future trajectories by point-in-polygon containment.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]