Automatic Plant Counting and Location Based on a Few-Shot Learning Technique

Karami, Alimohammad; Crawford, Melba M.; Delp, Edward J.

doi:10.1109/jstars.2020.3025790

Cited by 47 publications

(34 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In CenterNet, objects are represented as a single point, and heatmap is used to predict the centers of objects. Heatmap is created using a Gaussian kernel and an FCN, estimated centers are derived from the peak values in the heatmap [22]. Based on the center localization, object properties such as size and dimension can be regressed directly without any prior anchor [23].…”

Section: ) Training and Testingmentioning

confidence: 99%

Weed Identification Using Deep Learning and Image Processing in Vegetable Plantation

Che²,

2021

View full text Add to dashboard Cite

Weed identification in vegetable plantation is more challenging than crop weed identification due to their random plant spacing. So far, little work has been found on identifying weeds in vegetable plantation. Traditional methods of crop weed identification used to be mainly focused on identifying weed directly; however, there is a large variation in weed species. This paper proposes a new method in a contrary way, which combines deep learning and image processing technology. Firstly, a trained CenterNet model was used to detect vegetables and draw bounding boxes around them. Afterwards, the remaining green objects falling out of bounding boxes were considered as weeds. In this way, the model focuses on identifying only the vegetables and thus avoid handling various weed species. Furthermore, this strategy can largely reduce the size of training image dataset as well as the complexity of weed detection, thereby enhancing the weed identification performance and accuracy. To extract weeds from the background, a color index-based segmentation was performed utilizing image processing. The employed color index was determined and evaluated through Genetic Algorithms (GAs) according to Bayesian classification error. During the field test, the trained CenterNet model achieved a precision of 95.6%, a recall of 95.0%, and a " score of 0.953, respectively. The proposed index-19R + 24G-2B ≥ 862 yields high segmentation quality with a much lower computational cost compared to the wildly used ExG index. These experiment results demonstrate the feasibility of using the proposed method for the ground-based weed identification in vegetable plantation.

show abstract

Section: ) Training and Testingmentioning

confidence: 99%

Weed Identification Using Deep Learning and Image Processing in Vegetable Plantation

Che²,

2021

View full text Add to dashboard Cite

show abstract

“…[24,25]). The last category of methods widely used now is based on deep learning algorithms for automatic object detection [26][27][28].…”

Section: Introductionmentioning

confidence: 99%

“…An optimal altitude should therefore be selected to compromise between the acquisition throughput and the image GSD. Previous studies reporting early-stage maize plant detection from UAVs from deep learning methods did not addressed specifically this important scaling issue [ 20 , 26 , 27 ]. One way to address this scaling issue is to transform the low-resolution images into higher resolution ones using super resolution techniques.…”

Section: Introductionmentioning

confidence: 99%

Estimates of Maize Plant Density from UAV RGB Images Using Faster-RCNN Detection Model: Impact of the Spatial Resolution

Velumani

López-Lozano

Madec³

et al. 2021

Plant Phenomics

View full text Add to dashboard Cite

Early-stage plant density is an essential trait that determines the fate of a genotype under given environmental conditions and management practices. The use of RGB images taken from UAVs may replace the traditional visual counting in fields with improved throughput, accuracy, and access to plant localization. However, high-resolution images are required to detect the small plants present at the early stages. This study explores the impact of image ground sampling distance (GSD) on the performances of maize plant detection at three-to-five leaves stage using Faster-RCNN object detection algorithm. Data collected at high resolution (GSD≈0.3 cm) over six contrasted sites were used for model training. Two additional sites with images acquired both at high and low (GSD≈0.6 cm) resolutions were used to evaluate the model performances. Results show that Faster-RCNN achieved very good plant detection and counting (rRMSE=0.08) performances when native high-resolution images are used both for training and validation. Similarly, good performances were observed (rRMSE=0.11) when the model is trained over synthetic low-resolution images obtained by downsampling the native training high-resolution images and applied to the synthetic low-resolution validation images. Conversely, poor performances are obtained when the model is trained on a given spatial resolution and applied to another spatial resolution. Training on a mix of high- and low-resolution images allows to get very good performances on the native high-resolution (rRMSE=0.06) and synthetic low-resolution (rRMSE=0.10) images. However, very low performances are still observed over the native low-resolution images (rRMSE=0.48), mainly due to the poor quality of the native low-resolution images. Finally, an advanced super resolution method based on GAN (generative adversarial network) that introduces additional textural information derived from the native high-resolution images was applied to the native low-resolution validation images. Results show some significant improvement (rRMSE=0.22) compared to bicubic upsampling approach, while still far below the performances achieved over the native high-resolution images.

show abstract

“…In other words, the seamlines are controlled to only pass through mid-row-to-row and midplant-to-plant separation (refer to Figure 13c for a graphical illustration of the impact of using such a constraint). In this study, plant locations are derived from early-season UAV orthophotos through the approach proposed by Karami et al [50]. The proposed seamline control strategy is based on defining a uv local coordinate system where the v axis is aligned along the row direction.…”

Section: Orthophoto Quality Assessmentmentioning

confidence: 99%

“…The Average elevation within a row segment DSM is generated using the UAV-A1 LiDAR dataset. Plant locations were detected using early-season UAV RGB orthophoto through the approach described in Karami et al [50]. Figure 20 shows portions of the resulting orthophotos, with superimposed seamlines in yellow.…”

Section: Quality Verification Of Generated Orthophotos Using Ground Fmentioning

confidence: 99%

New Orthophoto Generation Strategies from UAV and Ground Remote Sensing Platforms for High-Throughput Phenotyping

et al. 2021

Self Cite

View full text Add to dashboard Cite

Remote sensing platforms have become an effective data acquisition tool for digital agriculture. Imaging sensors onboard unmanned aerial vehicles (UAVs) and tractors are providing unprecedented high-geometric-resolution data for several crop phenotyping activities (e.g., canopy cover estimation, plant localization, and flowering date identification). Among potential products, orthophotos play an important role in agricultural management. Traditional orthophoto generation strategies suffer from several artifacts (e.g., double mapping, excessive pixilation, and seamline distortions). The above problems are more pronounced when dealing with mid- to late-season imagery, which is often used for establishing flowering date (e.g., tassel and panicle detection for maize and sorghum crops, respectively). In response to these challenges, this paper introduces new strategies for generating orthophotos that are conducive to the straightforward detection of tassels and panicles. The orthophoto generation strategies are valid for both frame and push-broom imaging systems. The target function of these strategies is striking a balance between the improved visual appearance of tassels/panicles and their geolocation accuracy. The new strategies are based on generating a smooth digital surface model (DSM) that maintains the geolocation quality along the plant rows while reducing double mapping and pixilation artifacts. Moreover, seamline control strategies are applied to avoid having seamline distortions at locations where the tassels and panicles are expected. The quality of generated orthophotos is evaluated through visual inspection as well as quantitative assessment of the degree of similarity between the generated orthophotos and original images. Several experimental results from both UAV and ground platforms show that the proposed strategies do improve the visual quality of derived orthophotos while maintaining the geolocation accuracy at tassel/panicle locations.

show abstract

Automatic Plant Counting and Location Based on a Few-Shot Learning Technique

Cited by 47 publications

References 60 publications

Weed Identification Using Deep Learning and Image Processing in Vegetable Plantation

Weed Identification Using Deep Learning and Image Processing in Vegetable Plantation

Estimates of Maize Plant Density from UAV RGB Images Using Faster-RCNN Detection Model: Impact of the Spatial Resolution

New Orthophoto Generation Strategies from UAV and Ground Remote Sensing Platforms for High-Throughput Phenotyping

Contact Info

Product

Resources

About