Automated detection and measurement of individual sorghum panicles using density-based clustering of terrestrial lidar data

Malambo, Lonesome; Popescu, Sorin C.; Horne, David W.; Pugh, N. Ace; Rooney, William L.

doi:10.1016/j.isprsjprs.2018.12.015

Cited by 49 publications

(27 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on the two performance metrics, the estimated panicle count compared very well with the measured panicle count in each plot (Figure 8), which shows promise for the application of deep learning approaches for such a plant phenotyping task. The level of counting accuracy (94%) achieved in this study is an improvement from our previous study [5] (89.3%) that applied terrestrial lidar data and a density-based clustering approach. While different input data were used, we attribute the increased performance in this study to the deep learning model applied.…”

Section: Panicle Counting Performance and Field-level Panicle Mappingmentioning

confidence: 43%

“…Reference Estimated The level of counting accuracy (94%) achieved in this study is an improvement from our previous study [5] (89.3%) that applied terrestrial lidar data and a density-based clustering approach. While different input data were used, we attribute the increased performance in this study to the deep learning model applied.…”

Section: Plot Nomentioning

confidence: 58%

“…To get the total number of correctly classified panicles, N t , we matched detected panicle locations to measured x-y panicle locations. We considered a detected panicle correctly matched if it fell in a 5 cm buffer zone around a reference panicle, selecting the 5 cm radius based on average panicle width data [5]. Where multiple panicles fell in the buffer zone, we considered only the nearest panicle to the buffer center.…”

Section: Validating Detected Panicle Countsmentioning

confidence: 99%

“…Recent years have seen unmanned aerial systems (UAS) emerge as effective means for field-relevant phenotyping activities by enabling efficient and more affordable collection of aerial crop images over entire crop growth cycles. Traditional image analysis and machine learning have played key roles in transforming these image data into targeted phenotypic information such as plant height [1,2] and plant population counts [3][4][5]. However, the massive image data being collected by UAS and other high throughput platforms pose challenges to traditional methods, whose performance tends to level off and fall short of the high accuracy required for fully automated systems [6].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Deep Learning Semantic Segmentation-Based Approach for Field-Level Sorghum Panicle Counting

Malambo

Popescu

et al. 2019

Remote Sensing

View full text Add to dashboard Cite

Small unmanned aerial systems (UAS) have emerged as high-throughput platforms for the collection of high-resolution image data over large crop fields to support precision agriculture and plant breeding research. At the same time, the improved efficiency in image capture is leading to massive datasets, which pose analysis challenges in providing needed phenotypic data. To complement these high-throughput platforms, there is an increasing need in crop improvement to develop robust image analysis methods to analyze large amount of image data. Analysis approaches based on deep learning models are currently the most promising and show unparalleled performance in analyzing large image datasets. This study developed and applied an image analysis approach based on a SegNet deep learning semantic segmentation model to estimate sorghum panicles counts, which are critical phenotypic data in sorghum crop improvement, from UAS images over selected sorghum experimental plots. The SegNet model was trained to semantically segment UAS images into sorghum panicles, foliage and the exposed ground using 462, 250 × 250 labeled images, which was then applied to field orthomosaic to generate a field-level semantic segmentation. Individual panicle locations were obtained after post-processing the segmentation output to remove small objects and split merged panicles. A comparison between model panicle count estimates and manually digitized panicle locations in 60 randomly selected plots showed an overall detection accuracy of 94%. A per-plot panicle count comparison also showed high agreement between estimated and reference panicle counts (Spearman correlation ρ = 0.88, mean bias = 0.65). Misclassifications of panicles during the semantic segmentation step and mosaicking errors in the field orthomosaic contributed mainly to panicle detection errors. Overall, the approach based on deep learning semantic segmentation showed good promise and with a larger labeled dataset and extensive hyper-parameter tuning, should provide even more robust and effective characterization of sorghum panicle counts.

show abstract

Section: Panicle Counting Performance and Field-level Panicle Mappingmentioning

confidence: 43%

Section: Plot Nomentioning

confidence: 58%

Section: Validating Detected Panicle Countsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Deep Learning Semantic Segmentation-Based Approach for Field-Level Sorghum Panicle Counting

Malambo

Popescu

et al. 2019

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In the last 5 years, different approaches have been excogitated with promising capabilities of recording sorghum phenology in field environments. Some of these include various UAS platforms [17,18], field-based robotic phenotyping system [19], unmanned aerial system [20], ultrasonic sensors [21], the light detection and ranging (LiDAR) [22], the time of flight cameras [23], tomography imaging [24], Kinect v2 camera [21], RGB and NIR imaging [25], and Phenobot 1.0 [26]. The next-generation phenomics tools generate enormous amount of data that are being translated via machine-learning statistical approaches into trait descriptions, relevant to sorghum breeders [27].…”

Section: Analyzing Sorghum Biomass Potential 21 Phenotyping Biomass mentioning

confidence: 99%

Sorghum an Important Annual Feedstock for Bioenergy

Sadia¹,

Awan²,

Saleem³

et al. 2019

Biomass for Bioenergy - Recent Trends and Future Challenges

View full text Add to dashboard Cite

Plant-based renewable biofuels guarantee sustainable solutions to food and energy demands. High-biomass C4 grasses including sugarcane, corn, and sorghum are potential candidates for bioenergy. Among these, sorghum enjoys the status of a highly diverse food, feed, and biofuel source worldwide. The natural attributes like abiotic stress tolerance, diverse genetic base, viable seed industry, and sound breeding system make sorghum a perfect candidate for establishing an efficient and low-cost biofuel industry. Scientists are exploring ways to exploit forage, sweet, and biomass sorghums as climate-smart energy crops. In this context, conventional breeding has played a significant role in developing high-yielding sorghum varieties. For biomass sorghum, stem compositional analysis helps screen low lignin and high polysaccharide types as feedstocks for biofuels. Recent tools of phenomics, genomics, proteomics, and genome editing are key players of designing eco-friendly bioenergy sorghum. Here, we report stem compositional analysis and proteomicsbased evaluation of USDA sorghum germplasm as a baseline to develop sorghum as a biofuel feedstock.

show abstract

High‐throughput phenotyping of canopy height in cool‐season crops using sensing techniques

et al. 2021

View full text Add to dashboard Cite

Plant breeders are interested in plant height data, which is an important agronomic data associated with lodging and mechanical harvest. Manual measurement of plant height with limited samples per plot and data acquisition frequency remains the standard method in breeding programs. To overcome such limitations, this study focuses on plant height estimation in canola (Brassica napus L./winter canola, and B. napus L. and B. rapa L./spring canola), pea (Pisum sativum L.), chickpea (Cicer arietinum L.), and camelina (Camelina sativa L.) breeding trials using sensors. Plant height data were collected using a light detection and ranging (LiDAR) sensor system mounted on a tractor (for pea and chickpea) and an unmanned aerial system (UAS) integrated with a Red–Green–Blue (RGB) camera (for four crops). The LiDAR data and UAS‐based images were processed to extract six plant height features. Significant (P < .0001) correlations between LiDAR estimated and manually measured plant height data were observed with correlation coefficient (r) of .74 and .91 in chickpea and pea, respectively. Image‐based plant height estimations were also correlated (P < .0001) with manually measurement in the four crops (r = .57 – .98). This study demonstrated that the plant height of four cool‐season crops can be estimated using either proximal or remote sensing techniques even if the canopy architectures of these crops pose challenges. Such high throughput phenotyping technologies can be applied in plant breeding and crop production to monitor plant height and associated traits such as lodging in an efficient and timely manner.

show abstract

Automated detection and measurement of individual sorghum panicles using density-based clustering of terrestrial lidar data

Cited by 49 publications

References 48 publications

A Deep Learning Semantic Segmentation-Based Approach for Field-Level Sorghum Panicle Counting

A Deep Learning Semantic Segmentation-Based Approach for Field-Level Sorghum Panicle Counting

Sorghum an Important Annual Feedstock for Bioenergy

High‐throughput phenotyping of canopy height in cool‐season crops using sensing techniques

Contact Info

Product

Resources

About