{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/14615"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/14615","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"A support vector machine embedded weed identification system","abstract":"Over the past decades, the over-reliance on herbicides during corn production has caused severe environmental and biological problems such as pollution in the soil and underground water, and the emergence of the herbicide-resistant weed species. A potential solution to reduce the use of herbicides while maintaining adequate weed control lies in the combined use of chemical and mechanical weeding, in which weeds are controlled adaptively according to their reaction to herbicides. Accurate weed identification is a prerequisite for accomplishing such a control strategy. A machine vision system for weed identification, which utilized the morphological properties of weed leaves, was developed in this research. The system incorporated a new image segmentation algorithm, termed the ‘Pixelwise method’ to binarize the color weed images for subsequent image processing and feature extraction procedures. Subsequently, a Support Vector Machine (SVM) based classifier was constructed to distinguish various weed species using seven morphological features. 2,325 indoor images consisting of six weed species were acquired during the first five weeks after emergence of the plants. Among 1,006 test images, the SVM system achieved over 94% accuracy in crop (corn) versus weed discrimination and 95% in grass versus broadleaf weed discrimination. The average classification accuracy for individual weed species was approximately 86%. In addition, the system obtained the best classification result after the second week after plant emergence. In field tests, the SVM classifier based on the indoor image library was able to identify 71.1% of 270 weed plants in the field. With an adaptive median filter to enhance the image quality, the accuracy was raised to 75.9% at the expense of extra image processing time. Both of the laboratory and field tests showed that the SVM method with reasonable accuracy is feasible for weed identification during their early growth season.","abstract_html":"Over the past decades, the over-reliance on herbicides during corn production has caused severe environmental and biological problems such as pollution in the soil and underground water, and the emergence of the herbicide-resistant weed species. A potential solution to reduce the use of herbicides while maintaining adequate weed control lies in the combined use of chemical and mechanical weeding, in which weeds are controlled adaptively according to their reaction to herbicides. Accurate weed identification is a prerequisite for accomplishing such a control strategy. A machine vision system for weed identification, which utilized the morphological properties of weed leaves, was developed in this research. The system incorporated a new image segmentation algorithm, termed the ‘Pixelwise method’ to binarize the color weed images for subsequent image processing and feature extraction procedures. Subsequently, a Support Vector Machine (SVM) based classifier was constructed to distinguish various weed species using seven morphological features. 2,325 indoor images consisting of six weed species were acquired during the first five weeks after emergence of the plants. Among 1,006 test images, the SVM system achieved over 94% accuracy in crop (corn) versus weed discrimination and 95% in grass versus broadleaf weed discrimination. The average classification accuracy for individual weed species was approximately 86%. In addition, the system obtained the best classification result after the second week after plant emergence. In field tests, the SVM classifier based on the indoor image library was able to identify 71.1% of 270 weed plants in the field. With an adaptive median filter to enhance the image quality, the accuracy was raised to 75.9% at the expense of extra image processing time. Both of the laboratory and field tests showed that the SVM method with reasonable accuracy is feasible for weed identification during their early growth season.","abstract_has_math":false,"creators":["Lin, Chufan"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Agricultural Engineering","degree_department":null,"school":null,"contributors":["Grift, Tony E."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2010,"date_issued":"2010-01-06T16:19:46Z","date_published":"2010-01-06T16:19:46Z","updated_at":"2026-07-22T22:25:07Z","subjects":["machine vision","Pixelwise method","SVM","Weed identification"],"languages":["eng"],"rights":["Copyright 2009 Chufan Lin"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/14615","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Grift, Tony E."]},{"key":"dc:creator","label":"Author","values":["Lin, Chufan"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2010-01-06T16:19:46Z","2009-12"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Agricultural Engineering"]},{"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":["machine vision","Pixelwise method","SVM","Weed identification"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2009 Chufan Lin"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/14615"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Over the past decades, the over-reliance on herbicides during corn production has caused severe environmental and biological problems such as pollution in the soil and underground water, and the emergence of the herbicide-resistant weed species. A potential solution to reduce the use of herbicides while maintaining adequate weed control lies in the combined use of chemical and mechanical weeding, in which weeds are controlled adaptively according to their reaction to herbicides. Accurate weed identification is a prerequisite for accomplishing such a control strategy. A machine vision system for weed identification, which utilized the morphological properties of weed leaves, was developed in this research. The system incorporated a new image segmentation algorithm, termed the ‘Pixelwise method’ to binarize the color weed images for subsequent image processing and feature extraction procedures. Subsequently, a Support Vector Machine (SVM) based classifier was constructed to distinguish various weed species using seven morphological features. 2,325 indoor images consisting of six weed species were acquired during the first five weeks after emergence of the plants. Among 1,006 test images, the SVM system achieved over 94% accuracy in crop (corn) versus weed discrimination and 95% in grass versus broadleaf weed discrimination. The average classification accuracy for individual weed species was approximately 86%. In addition, the system obtained the best classification result after the second week after plant emergence. In field tests, the SVM classifier based on the indoor image library was able to identify 71.1% of 270 weed plants in the field. With an adaptive median filter to enhance the image quality, the accuracy was raised to 75.9% at the expense of extra image processing time. Both of the laboratory and field tests showed that the SVM method with reasonable accuracy is feasible for weed identification during their early growth season.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2009-12-08T22:11:32Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 9 FieldWeedIdentify.m: 8678 bytes, checksum: 083b7f175b50cb6d5bbbb4edd6ee0d00 (MD5) FieldImgAcq.m: 2504 bytes, checksum: 879c333d91c992c6024f3ce1fd3834c6 (MD5) PixelwiseSeg.m: 1832 bytes, checksum: f65388ee3025d4569a8f5fe0b17735a9 (MD5) RGBrelation.m: 7444 bytes, checksum: 67c476ef0d22a9456f4bd3b85f03c94b (MD5) ModifiedHue.m: 3916 bytes, checksum: 2c5879aaeab8a74b4a99004d9053e257 (MD5) NExG.m: 2082 bytes, checksum: 3f4b48799af2a4f734aa73e04dee1df3 (MD5) AcquireImage.m: 3695 bytes, checksum: c07b750bbfff20b85ca6245ab292a683 (MD5) Lin_Chufan.pdf: 1118470 bytes, checksum: 2ab2bfbf731ce0a278defcb9a9124801 (MD5) Lin_Chufan.doc: 7027712 bytes, checksum: 650ce4afe7d9739f54b89cd9f4e1d8ad (MD5)","Made available in DSpace on 2010-01-06T16:19:46Z (GMT). No. of bitstreams: 12 license.txt: 4057 bytes, checksum: 14acf691a3f9ac4c5f3b840b825f75fc (MD5) Lin_Chufan.pdf: 1118350 bytes, checksum: b70261d3b595cb7fd62750742ba0713a (MD5) FieldWeedIdentify.m: 8678 bytes, checksum: 083b7f175b50cb6d5bbbb4edd6ee0d00 (MD5) FieldImgAcq.m: 2504 bytes, checksum: 879c333d91c992c6024f3ce1fd3834c6 (MD5) PixelwiseSeg.m: 1832 bytes, checksum: f65388ee3025d4569a8f5fe0b17735a9 (MD5) RGBrelation.m: 7444 bytes, checksum: 67c476ef0d22a9456f4bd3b85f03c94b (MD5) ModifiedHue.m: 3916 bytes, checksum: 2c5879aaeab8a74b4a99004d9053e257 (MD5) NExG.m: 2082 bytes, checksum: 3f4b48799af2a4f734aa73e04dee1df3 (MD5) AcquireImage.m: 3695 bytes, checksum: c07b750bbfff20b85ca6245ab292a683 (MD5) 1_Lin_Chufan.pdf: 1120325 bytes, checksum: f3ac24e66789f94753447c9c805f72e5 (MD5) 2_Lin_Chufan.pdf: 1118470 bytes, checksum: 2ab2bfbf731ce0a278defcb9a9124801 (MD5) Lin_Chufan.doc: 7027712 bytes, checksum: 650ce4afe7d9739f54b89cd9f4e1d8ad (MD5)"]},{"key":"dc:title","label":"Title","values":["A support vector machine embedded weed identification system"]}]}],"canonical_facts":{"dc:contributor":["Grift, Tony E."],"dc:creator":["Lin, Chufan"],"dc:date":["2010-01-06T16:19:46Z","2009-12"],"dc:description":["Over the past decades, the over-reliance on herbicides during corn production has caused severe environmental and biological problems such as pollution in the soil and underground water, and the emergence of the herbicide-resistant weed species. A potential solution to reduce the use of herbicides while maintaining adequate weed control lies in the combined use of chemical and mechanical weeding, in which weeds are controlled adaptively according to their reaction to herbicides. Accurate weed identification is a prerequisite for accomplishing such a control strategy. A machine vision system for weed identification, which utilized the morphological properties of weed leaves, was developed in this research. The system incorporated a new image segmentation algorithm, termed the ‘Pixelwise method’ to binarize the color weed images for subsequent image processing and feature extraction procedures. Subsequently, a Support Vector Machine (SVM) based classifier was constructed to distinguish various weed species using seven morphological features. 2,325 indoor images consisting of six weed species were acquired during the first five weeks after emergence of the plants. Among 1,006 test images, the SVM system achieved over 94% accuracy in crop (corn) versus weed discrimination and 95% in grass versus broadleaf weed discrimination. The average classification accuracy for individual weed species was approximately 86%. In addition, the system obtained the best classification result after the second week after plant emergence. In field tests, the SVM classifier based on the indoor image library was able to identify 71.1% of 270 weed plants in the field. With an adaptive median filter to enhance the image quality, the accuracy was raised to 75.9% at the expense of extra image processing time. Both of the laboratory and field tests showed that the SVM method with reasonable accuracy is feasible for weed identification during their early growth season.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2009-12-08T22:11:32Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 9 FieldWeedIdentify.m: 8678 bytes, checksum: 083b7f175b50cb6d5bbbb4edd6ee0d00 (MD5) FieldImgAcq.m: 2504 bytes, checksum: 879c333d91c992c6024f3ce1fd3834c6 (MD5) PixelwiseSeg.m: 1832 bytes, checksum: f65388ee3025d4569a8f5fe0b17735a9 (MD5) RGBrelation.m: 7444 bytes, checksum: 67c476ef0d22a9456f4bd3b85f03c94b (MD5) ModifiedHue.m: 3916 bytes, checksum: 2c5879aaeab8a74b4a99004d9053e257 (MD5) NExG.m: 2082 bytes, checksum: 3f4b48799af2a4f734aa73e04dee1df3 (MD5) AcquireImage.m: 3695 bytes, checksum: c07b750bbfff20b85ca6245ab292a683 (MD5) Lin_Chufan.pdf: 1118470 bytes, checksum: 2ab2bfbf731ce0a278defcb9a9124801 (MD5) Lin_Chufan.doc: 7027712 bytes, checksum: 650ce4afe7d9739f54b89cd9f4e1d8ad (MD5)","Made available in DSpace on 2010-01-06T16:19:46Z (GMT). No. of bitstreams: 12 license.txt: 4057 bytes, checksum: 14acf691a3f9ac4c5f3b840b825f75fc (MD5) Lin_Chufan.pdf: 1118350 bytes, checksum: b70261d3b595cb7fd62750742ba0713a (MD5) FieldWeedIdentify.m: 8678 bytes, checksum: 083b7f175b50cb6d5bbbb4edd6ee0d00 (MD5) FieldImgAcq.m: 2504 bytes, checksum: 879c333d91c992c6024f3ce1fd3834c6 (MD5) PixelwiseSeg.m: 1832 bytes, checksum: f65388ee3025d4569a8f5fe0b17735a9 (MD5) RGBrelation.m: 7444 bytes, checksum: 67c476ef0d22a9456f4bd3b85f03c94b (MD5) ModifiedHue.m: 3916 bytes, checksum: 2c5879aaeab8a74b4a99004d9053e257 (MD5) NExG.m: 2082 bytes, checksum: 3f4b48799af2a4f734aa73e04dee1df3 (MD5) AcquireImage.m: 3695 bytes, checksum: c07b750bbfff20b85ca6245ab292a683 (MD5) 1_Lin_Chufan.pdf: 1120325 bytes, checksum: f3ac24e66789f94753447c9c805f72e5 (MD5) 2_Lin_Chufan.pdf: 1118470 bytes, checksum: 2ab2bfbf731ce0a278defcb9a9124801 (MD5) Lin_Chufan.doc: 7027712 bytes, checksum: 650ce4afe7d9739f54b89cd9f4e1d8ad (MD5)"],"dc:identifier":["http://hdl.handle.net/2142/14615"],"dc:language":["eng"],"dc:rights":["Copyright 2009 Chufan Lin"],"dc:subject":["machine vision","Pixelwise method","SVM","Weed identification"],"dc:title":["A support vector machine embedded weed identification system"],"thesis:degree_discipline":["Agricultural Engineering"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:07Z"}