Patricia O'Byrne - Academia.edu (original) (raw)
Uploads
Papers by Patricia O'Byrne
In an era of climate change and global population growth, deep learning based multi-spectral imag... more In an era of climate change and global population growth, deep learning based multi-spectral imaging has the potential to significantly assist in production management across a wide range of agricultural and food production domains. A key challenge however in applying state-of-the-art methods is that they, unlike classical hand crafted methods, are usually thought of as being only useful when significant amounts of data are available. In this paper we investigate this hypothesis by examining the performance of state-ofthe-art deep learning methods when applied to a restricted data set that is not easily bootstrapped through pre-trained image processing networks. We demonstrate that significant result improvement can be obtained from deep residual networks over a baseline image processing model-even in the case where data collection is highly expensive and pre-trained networks cannot be easily built upon. Our work also constitutes a useful contribution to understanding the benefit of applying deep image multi-spectral processing techniques to the agri-food domain.
A series of experiments were conducted to measure and quantify the yield, dry matter content, sug... more A series of experiments were conducted to measure and quantify the yield, dry matter content, sugars content and nitrates content of grass intended for ensilement. These experiments took place in the East Midlands of Ireland during the Spring, Summer and Autumn of 2019. A bespoke sensor rig was constructed; included in this rig was a hyperspectral radiometer that measured a broad spectrum of reflected natural light from a circular spot approximately 1.2 metres in area. Grass inside a 50cm square quadrat was manually collected from the centre of the circular spot for ground truth estimation of the grass qualities. Up to 25 spots were recorded and sampled each day. The radiometer readings for each spot were automatically recorded onto a laptop that controlled the sensor rig, and ground truth measurements were made either on site or within 24 hours in a wet chemistry laboratory. The collected data was used to build Partial Least Squares Regression (PLSR) predictive models of grass qualities from the hyperspectral dataset and it was found that substantial relationships exist between the spectral reflectance from the grass and yield (r2 = 0.62), dry matter % (r2 = 0.54), sugar content (r2 = 0.54) and nitrates (r2 = 0.50). This shows that hyperspectral reflectance data contains substantial information about key grass qualities and can form part of a broader holistic data driven approach to provide accurate and rapid predictions to farmers, agronomists and agricultural contractors.
A dissertation submitted in partial fulfilment of the requirements of Staffordshire University fo... more A dissertation submitted in partial fulfilment of the requirements of Staffordshire University for the degree of M.Sc.
* Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for fu... more * Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for funding † on behalf of Tanco Autowrap TM ‡ Thanks to ADAPT Centre for Digital Content Technology which is funded under the SFI Research Centres Programme (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund for funding.
PROCEEDINGS OF 7th China Europe International Symposium on Software Industry oriented Education, 2011
A Tanzanian higher education institution (Institute of Finance Management) and an Irish higher ed... more A Tanzanian higher education institution (Institute of Finance Management) and an Irish higher education institutio (Dublin Institute of Technology) have worked in partnership to develop the final stage of a Computer Science programme to meet the specific needs of the Tanzanian Information and Communication Technology (ICT) sector. The designed stage sits above three stages of a programme designed for the Irish ICT sector which was transplanted to the Tanzanian context. The final stage is crucial as it represents the final ...
AgriEngineering, 2021
A series of experiments were conducted to measure and quantify the yield, dry matter content, sug... more A series of experiments were conducted to measure and quantify the yield, dry matter content, sugars content, and nitrates content of grass intended for ensilement. These experiments took place in the East Midlands of Ireland during the Spring, Summer, and Autumn of 2019. A bespoke sensor rig was constructed; included in this rig was a hyperspectral radiometer that measured a broad spectrum of reflected natural light from a circular spot approximately 1.2 m in area. Grass inside a 50 cm square quadrat was manually collected from the centre of the circular spot for ground truth estimation of the grass qualities. Up to 25 spots were recorded and sampled each day. The radiometer readings for each spot were automatically recorded onto a laptop that controlled the sensor rig, and ground truth measurements were made either on-site or within 24 h in a wet chemistry laboratory. The collected data was used to build Partial Least Squares Regression (PLSR) predictive models of grass qualities ...
2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021
* Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for fu... more * Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for funding † on behalf of Tanco Autowrap TM ‡ Thanks to ADAPT Centre for Digital Content Technology which is funded under the SFI Research Centres Programme (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund for funding.
A dissertation submitted in partial fulfilment of the requirements of Staffordshire University fo... more A dissertation submitted in partial fulfilment of the requirements of Staffordshire University for the degree of M.Sc.
In an era of climate change and global population growth, deep learning based multi-spectral imag... more In an era of climate change and global population growth, deep learning based multi-spectral imaging has the potential to significantly assist in production management across a wide range of agricultural and food production domains. A key challenge however in applying state-of-the-art methods is that they, unlike classical hand crafted methods, are usually thought of as being only useful when significant amounts of data are available. In this paper we investigate this hypothesis by examining the performance of state-ofthe-art deep learning methods when applied to a restricted data set that is not easily bootstrapped through pre-trained image processing networks. We demonstrate that significant result improvement can be obtained from deep residual networks over a baseline image processing model-even in the case where data collection is highly expensive and pre-trained networks cannot be easily built upon. Our work also constitutes a useful contribution to understanding the benefit of applying deep image multi-spectral processing techniques to the agri-food domain.
A series of experiments were conducted to measure and quantify the yield, dry matter content, sug... more A series of experiments were conducted to measure and quantify the yield, dry matter content, sugars content and nitrates content of grass intended for ensilement. These experiments took place in the East Midlands of Ireland during the Spring, Summer and Autumn of 2019. A bespoke sensor rig was constructed; included in this rig was a hyperspectral radiometer that measured a broad spectrum of reflected natural light from a circular spot approximately 1.2 metres in area. Grass inside a 50cm square quadrat was manually collected from the centre of the circular spot for ground truth estimation of the grass qualities. Up to 25 spots were recorded and sampled each day. The radiometer readings for each spot were automatically recorded onto a laptop that controlled the sensor rig, and ground truth measurements were made either on site or within 24 hours in a wet chemistry laboratory. The collected data was used to build Partial Least Squares Regression (PLSR) predictive models of grass qualities from the hyperspectral dataset and it was found that substantial relationships exist between the spectral reflectance from the grass and yield (r2 = 0.62), dry matter % (r2 = 0.54), sugar content (r2 = 0.54) and nitrates (r2 = 0.50). This shows that hyperspectral reflectance data contains substantial information about key grass qualities and can form part of a broader holistic data driven approach to provide accurate and rapid predictions to farmers, agronomists and agricultural contractors.
A dissertation submitted in partial fulfilment of the requirements of Staffordshire University fo... more A dissertation submitted in partial fulfilment of the requirements of Staffordshire University for the degree of M.Sc.
* Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for fu... more * Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for funding † on behalf of Tanco Autowrap TM ‡ Thanks to ADAPT Centre for Digital Content Technology which is funded under the SFI Research Centres Programme (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund for funding.
PROCEEDINGS OF 7th China Europe International Symposium on Software Industry oriented Education, 2011
A Tanzanian higher education institution (Institute of Finance Management) and an Irish higher ed... more A Tanzanian higher education institution (Institute of Finance Management) and an Irish higher education institutio (Dublin Institute of Technology) have worked in partnership to develop the final stage of a Computer Science programme to meet the specific needs of the Tanzanian Information and Communication Technology (ICT) sector. The designed stage sits above three stages of a programme designed for the Irish ICT sector which was transplanted to the Tanzanian context. The final stage is crucial as it represents the final ...
AgriEngineering, 2021
A series of experiments were conducted to measure and quantify the yield, dry matter content, sug... more A series of experiments were conducted to measure and quantify the yield, dry matter content, sugars content, and nitrates content of grass intended for ensilement. These experiments took place in the East Midlands of Ireland during the Spring, Summer, and Autumn of 2019. A bespoke sensor rig was constructed; included in this rig was a hyperspectral radiometer that measured a broad spectrum of reflected natural light from a circular spot approximately 1.2 m in area. Grass inside a 50 cm square quadrat was manually collected from the centre of the circular spot for ground truth estimation of the grass qualities. Up to 25 spots were recorded and sampled each day. The radiometer readings for each spot were automatically recorded onto a laptop that controlled the sensor rig, and ground truth measurements were made either on-site or within 24 h in a wet chemistry laboratory. The collected data was used to build Partial Least Squares Regression (PLSR) predictive models of grass qualities ...
2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS, 2021
* Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for fu... more * Thanks to Enterprise Ireland Innovation Partnership IP 2018 0728 and Tanco Autowrap Ltd. for funding † on behalf of Tanco Autowrap TM ‡ Thanks to ADAPT Centre for Digital Content Technology which is funded under the SFI Research Centres Programme (Grant 13/RC/2106) and is co-funded under the European Regional Development Fund for funding.
A dissertation submitted in partial fulfilment of the requirements of Staffordshire University fo... more A dissertation submitted in partial fulfilment of the requirements of Staffordshire University for the degree of M.Sc.