{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/101720"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/101720","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"An adaptive image processing algorithm for field plant population analysis based on UAV imaging system","abstract":"Plant population is among the most important items of agricultural data. In crop science, plant population determines canopy closure time, plant spacing, and weeds competition. For farmers, if plant population can be obtained in time, it can help in crop management decisions on replanting, fertilizer application, and estimation of yield. For agricultural manufacturing companies, they need plant population data to check the accuracy between planter setting and actual plant population. However, counting of plant population is difficult in agriculture. Manual counting is the main counting method to get plant population. For 30 inch rows, researchers or farmers measure 17 feet 5 inches and count the number of plants along the length to get plant population of 1/1,000 acre. Then the count is multiplied by 1,000 to get the per acre population. This method is a sampling estimation with lots of time cost. To improve the counting efficiency, image-processing methods have been used to identify individual plants and count plant numbers in small experimental plots with a ground camera. But those methods are limited to an experimental environment with fixed cameras, which is hard to apply in large fields. With advanced technologies, it is possible to collect field images by using an unmanned aerial vehicle (UAV). In the most recent research, low-altitude (under 10 m) drone images have been used to estimate plant density of wheat crops. New algorithms of plant counting have been proposed for low-altitude drone images. But the limitation is that low-altitude flights only cover small field area and may damage plants near the flight path. In this study, a novel image-processing algorithm is developed for measuring plant population from medium-altitude (25 m - 50 m) drone images. Those images are collected by UAV image system for an 80 acres field. Based on a large number of field images, the algorithm was developed with overall consideration of crop color, crop space, growth status, and plant row information. Then, drone images are processed by the algorithm to generate a plant population map of entire field. Finally, the population map was checked by real field counting results. The population results are generated from medium-altitude drone images. When compared real field check points with manually counted populations, the difference of their mean population is less than1,000 plants/acre. The R2 between the manually checked points and the population map is 0.82, which means the two datasets in these sample points are highly correlated. Then for the two groups of data, statistical analysis by paired-samples t test yielded a p value of 0.062. There is no significant difference between two groups’ data. For this proposed method, the UAV imaging system can cover an 80-acre field in 10 minutes, and the plant population map of the entire field can be generated by the algorithm within 120 minutes, which would cost hundreds of hours for manual population counting. For the experiment field, planter population setting was 31,000/acre, while the actual counted plant population is averaged at 27,000/acre, 12.9% less than setting.","abstract_html":"Plant population is among the most important items of agricultural data. In crop science, plant population determines canopy closure time, plant spacing, and weeds competition. For farmers, if plant population can be obtained in time, it can help in crop management decisions on replanting, fertilizer application, and estimation of yield. For agricultural manufacturing companies, they need plant population data to check the accuracy between planter setting and actual plant population. However, counting of plant population is difficult in agriculture. Manual counting is the main counting method to get plant population. For 30 inch rows, researchers or farmers measure 17 feet 5 inches and count the number of plants along the length to get plant population of 1/1,000 acre. Then the count is multiplied by 1,000 to get the per acre population. This method is a sampling estimation with lots of time cost. To improve the counting efficiency, image-processing methods have been used to identify individual plants and count plant numbers in small experimental plots with a ground camera. But those methods are limited to an experimental environment with fixed cameras, which is hard to apply in large fields. With advanced technologies, it is possible to collect field images by using an unmanned aerial vehicle (UAV). In the most recent research, low-altitude (under 10 m) drone images have been used to estimate plant density of wheat crops. New algorithms of plant counting have been proposed for low-altitude drone images. But the limitation is that low-altitude flights only cover small field area and may damage plants near the flight path. In this study, a novel image-processing algorithm is developed for measuring plant population from medium-altitude (25 m - 50 m) drone images. Those images are collected by UAV image system for an 80 acres field. Based on a large number of field images, the algorithm was developed with overall consideration of crop color, crop space, growth status, and plant row information. Then, drone images are processed by the algorithm to generate a plant population map of entire field. Finally, the population map was checked by real field counting results. The population results are generated from medium-altitude drone images. When compared real field check points with manually counted populations, the difference of their mean population is less than1,000 plants/acre. The R2 between the manually checked points and the population map is 0.82, which means the two datasets in these sample points are highly correlated. Then for the two groups of data, statistical analysis by paired-samples t test yielded a p value of 0.062. There is no significant difference between two groups’ data. For this proposed method, the UAV imaging system can cover an 80-acre field in 10 minutes, and the plant population map of the entire field can be generated by the algorithm within 120 minutes, which would cost hundreds of hours for manual population counting. For the experiment field, planter population setting was 31,000/acre, while the actual counted plant population is averaged at 27,000/acre, 12.9% less than setting.","abstract_has_math":false,"creators":["Fan, Youheng"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Agricultural & Biological Engr","degree_department":null,"school":null,"contributors":["Tian, Lei"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2018,"date_issued":"2018-09-27T16:34:21Z","date_published":"2018-09-27T16:34:21Z","updated_at":"2026-07-22T22:24:40Z","subjects":["plant population, image processing, unmanned aerial vehicle, remote sensing, population map, precision agriculture"],"languages":["en"],"rights":["Copyright 2018 Youheng Fan"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/101720","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Tian, Lei"]},{"key":"dc:creator","label":"Author","values":["Fan, Youheng"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2018-09-27T16:34:21Z","2020-09-28T09:15:22Z","2018-07-18","2018-08"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Agricultural & Biological Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["plant population, image processing, unmanned aerial vehicle, remote sensing, population map, precision agriculture"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2018 Youheng Fan"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/101720"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Plant population is among the most important items of agricultural data. In crop science, plant population determines canopy closure time, plant spacing, and weeds competition. For farmers, if plant population can be obtained in time, it can help in crop management decisions on replanting, fertilizer application, and estimation of yield. For agricultural manufacturing companies, they need plant population data to check the accuracy between planter setting and actual plant population. However, counting of plant population is difficult in agriculture. Manual counting is the main counting method to get plant population. For 30 inch rows, researchers or farmers measure 17 feet 5 inches and count the number of plants along the length to get plant population of 1/1,000 acre. Then the count is multiplied by 1,000 to get the per acre population. This method is a sampling estimation with lots of time cost. To improve the counting efficiency, image-processing methods have been used to identify individual plants and count plant numbers in small experimental plots with a ground camera. But those methods are limited to an experimental environment with fixed cameras, which is hard to apply in large fields. With advanced technologies, it is possible to collect field images by using an unmanned aerial vehicle (UAV). In the most recent research, low-altitude (under 10 m) drone images have been used to estimate plant density of wheat crops. New algorithms of plant counting have been proposed for low-altitude drone images. But the limitation is that low-altitude flights only cover small field area and may damage plants near the flight path. In this study, a novel image-processing algorithm is developed for measuring plant population from medium-altitude (25 m - 50 m) drone images. Those images are collected by UAV image system for an 80 acres field. Based on a large number of field images, the algorithm was developed with overall consideration of crop color, crop space, growth status, and plant row information. Then, drone images are processed by the algorithm to generate a plant population map of entire field. Finally, the population map was checked by real field counting results. The population results are generated from medium-altitude drone images. When compared real field check points with manually counted populations, the difference of their mean population is less than1,000 plants/acre. The R2 between the manually checked points and the population map is 0.82, which means the two datasets in these sample points are highly correlated. Then for the two groups of data, statistical analysis by paired-samples t test yielded a p value of 0.062. There is no significant difference between two groups’ data. For this proposed method, the UAV imaging system can cover an 80-acre field in 10 minutes, and the plant population map of the entire field can be generated by the algorithm within 120 minutes, which would cost hundreds of hours for manual population counting. For the experiment field, planter population setting was 31,000/acre, while the actual counted plant population is averaged at 27,000/acre, 12.9% less than setting.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2020-08-01","The student, Youheng Fan, accepted the attached license on 2018-07-17 at 15:21.","The student, Youheng Fan, submitted this Thesis for approval on 2018-07-17 at 15:34.","This Thesis was approved for publication on 2018-07-18 at 09:15.","DSpace SAF Submission Ingestion Package generated from Vireo submission #12904 on 2018-09-27 at 11:19:22","Made available in DSpace on 2018-09-27T16:34:21Z (GMT). No. of bitstreams: 2 FAN-THESIS-2018.pdf: 5396405 bytes, checksum: f5bc8437282e46c1c33f12b9f699bcfb (MD5) LICENSE.txt: 4208 bytes, checksum: d70f8de86e9b4230fc5712b5cb4cb99e (MD5) Previous issue date: 2018-07-18","Embargo set by: Seth Robbins for item 107820 Lift date: 2020-09-27T16:34:29Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only Restriction Lifted for Item 107820 on 2020-09-28T09:15:22Z."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["An adaptive image processing algorithm for field plant population analysis based on UAV imaging system"]}]}],"canonical_facts":{"dc:contributor":["Tian, Lei"],"dc:creator":["Fan, Youheng"],"dc:date":["2018-09-27T16:34:21Z","2020-09-28T09:15:22Z","2018-07-18","2018-08"],"dc:description":["Plant population is among the most important items of agricultural data. In crop science, plant population determines canopy closure time, plant spacing, and weeds competition. For farmers, if plant population can be obtained in time, it can help in crop management decisions on replanting, fertilizer application, and estimation of yield. For agricultural manufacturing companies, they need plant population data to check the accuracy between planter setting and actual plant population. However, counting of plant population is difficult in agriculture. Manual counting is the main counting method to get plant population. For 30 inch rows, researchers or farmers measure 17 feet 5 inches and count the number of plants along the length to get plant population of 1/1,000 acre. Then the count is multiplied by 1,000 to get the per acre population. This method is a sampling estimation with lots of time cost. To improve the counting efficiency, image-processing methods have been used to identify individual plants and count plant numbers in small experimental plots with a ground camera. But those methods are limited to an experimental environment with fixed cameras, which is hard to apply in large fields. With advanced technologies, it is possible to collect field images by using an unmanned aerial vehicle (UAV). In the most recent research, low-altitude (under 10 m) drone images have been used to estimate plant density of wheat crops. New algorithms of plant counting have been proposed for low-altitude drone images. But the limitation is that low-altitude flights only cover small field area and may damage plants near the flight path. In this study, a novel image-processing algorithm is developed for measuring plant population from medium-altitude (25 m - 50 m) drone images. Those images are collected by UAV image system for an 80 acres field. Based on a large number of field images, the algorithm was developed with overall consideration of crop color, crop space, growth status, and plant row information. Then, drone images are processed by the algorithm to generate a plant population map of entire field. Finally, the population map was checked by real field counting results. The population results are generated from medium-altitude drone images. When compared real field check points with manually counted populations, the difference of their mean population is less than1,000 plants/acre. The R2 between the manually checked points and the population map is 0.82, which means the two datasets in these sample points are highly correlated. Then for the two groups of data, statistical analysis by paired-samples t test yielded a p value of 0.062. There is no significant difference between two groups’ data. For this proposed method, the UAV imaging system can cover an 80-acre field in 10 minutes, and the plant population map of the entire field can be generated by the algorithm within 120 minutes, which would cost hundreds of hours for manual population counting. For the experiment field, planter population setting was 31,000/acre, while the actual counted plant population is averaged at 27,000/acre, 12.9% less than setting.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2020-08-01","The student, Youheng Fan, accepted the attached license on 2018-07-17 at 15:21.","The student, Youheng Fan, submitted this Thesis for approval on 2018-07-17 at 15:34.","This Thesis was approved for publication on 2018-07-18 at 09:15.","DSpace SAF Submission Ingestion Package generated from Vireo submission #12904 on 2018-09-27 at 11:19:22","Made available in DSpace on 2018-09-27T16:34:21Z (GMT). No. of bitstreams: 2 FAN-THESIS-2018.pdf: 5396405 bytes, checksum: f5bc8437282e46c1c33f12b9f699bcfb (MD5) LICENSE.txt: 4208 bytes, checksum: d70f8de86e9b4230fc5712b5cb4cb99e (MD5) Previous issue date: 2018-07-18","Embargo set by: Seth Robbins for item 107820 Lift date: 2020-09-27T16:34:29Z Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD system","U of I Only Restriction Lifted for Item 107820 on 2020-09-28T09:15:22Z."],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/101720"],"dc:language":["en"],"dc:rights":["Copyright 2018 Youheng Fan"],"dc:subject":["plant population, image processing, unmanned aerial vehicle, remote sensing, population map, precision agriculture"],"dc:title":["An adaptive image processing algorithm for field plant population analysis based on UAV imaging system"],"dc:type":["text"],"thesis:degree_discipline":["Agricultural & Biological Engr"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:40Z"}