A handheld device for measuring the diameter at breast height of individual trees using laser ranging and deep-learning based image recognition

Song, Chuangye; Yang, Bin; Zhang, Lin; Wu, Dongxiu

doi:10.1186/s13007-021-00748-z

“…In recent years, the application of deep learning algorithms, such as convolutional neural networks, has improved the convenience and accuracy of image recognition and segmentation in different fields, such as face mask detection [ 33 ], classification of magnetic resonance images [ 34 ], X-ray images [ 35 ], and tree trunk identification [ 36 ]. The performance of deep learning has proven to be superior to other classical methods of computer vision [ 37 ].…”

Section: Discussionmentioning

confidence: 99%

Adaptiveness of RGB-image derived algorithms in the measurement of fractional vegetation coverage

Song

¹

,

Sang

²

,

Zhang

³

et al. 2022

BMC Bioinformatics

Self Cite

2

0

View full text Add to dashboard Cite

Background Fractional vegetation coverage (FVC) is a crucial parameter in determining vegetation structure. Automatic measurement of FVC using digital images captured by mobile smart devices is a potential direction for future research on field survey methods in plant ecology, and this algorithm is crucial for accurate FVC measurement. However, there is a lack of insight into the influence of illumination on the accuracy of FVC measurements. Therefore, the main objective of this research is to assess the adaptiveness and performance of different algorithms under varying light conditions for FVC measurements and to deepen our understanding of the influence of illumination on FVC measurement. Methods and results Based on a literature survey, we selected four algorithms that have been reported to have high accuracy in automatic FVC measurements. The first algorithm (Fun01) identifies green plants based on the combination of $$R/G$$ R / G , $$B/G$$ B / G , and $$ExG$$ ExG ($$R$$ R , $$G$$ G , and $$B$$ B are the actual pixel digital numbers from the images based on each RGB channel, $$ExG$$ ExG is the abbreviation of the Excess Green index), the second algorithm (Fun02) is a decision tree that uses color properties to discriminate plants from the background, the third algorithm (Fun03) uses $$ExG-ExR$$ E x G - E x R ($$ExR$$ ExR is the abbreviation of the Excess Red index) to recognize plants in the image, and the fourth algorithm (Fun04) uses $$ExG$$ ExG and $$O{\text{tsu}}$$ O tsu to separate the plants from the background. $$Otsu$$ Otsu is an algorithm used to determine a threshold to transform the image into a binary image for the vegetation and background. We measured the FVC of several surveyed quadrats using these four algorithms under three scenarios, namely overcast sky, solar forenoon, and solar noon. FVC values obtained using the Photoshop-assisted manual identification method were used as a reference to assess the accuracy of the four algorithms selected. Results indicate that under the overcast sky scenario, Fun01 was more accurate than the other algorithms and the MAPE (mean absolute percentage error), BIAS, relBIAS (relative BIAS), RMSE (root mean square error), and relRMSE (relative RMSE) are 8.68%, 1.3, 3.97, 3.13, and 12.33%, respectively. Under the scenario of the solar forenoon, Fun02 (decision tree) was more accurate than other algorithms, and the MAPE, BIAS, relBIAS, RMSE, and relRMSE are 22.70%, − 2.86, − 7.70, 5.00, and 41.23%. Under the solar noon scenario, Fun02 was also more accurate than the other algorithms, and the MAPE, BIAS, relBIAS, RMSE, and relRMSE are 20.60%, − 6.39, − 20.67, 7.30, and 24.49%, respectively. Conclusions Given that each algorithm has its own optimal application scenario, among the four algorithms selected, Fun01 (the combination of $$R/G$$ R / G , $$B/G$$ B / G , and $$ExG$$ ExG ) can be recommended for measuring FVC on cloudy days. Fun02 (decision tree) is more suitable for measuring the FVC on sunny days. However, it considerably underestimates the FVC in most cases. We expect the findings of this study to serve as a useful reference for automatic vegetation cover measurements.

show abstract

“…It yields results which are close to state of the art or better in many areas [27]. These technologies are used in image recognition, real-time translations, voice recognition services such as Siri from Apple, Alexa from Amazon, and Cortana from Microsoft [28,29]. In addition, multiple studies have shown that DL algorithms are state of the art in breast histopathology analysis, skin cancer classification, risk of cardiovascular disease prediction, the detection of lung cancer, and several applications in the ophthalmic area with the potential to revolutionize the diagnosis of eye diseases [30][31][32].…”

Section: Introductionmentioning

confidence: 88%

Literature Review on Artificial Intelligence Methods for Glaucoma Screening, Segmentation, and Classification

Camara

¹

,

Neto

²

,

Pires

³

et al. 2022

View full text Add to dashboard Cite

Artificial intelligence techniques are now being applied in different medical solutions ranging from disease screening to activity recognition and computer-aided diagnosis. The combination of computer science methods and medical knowledge facilitates and improves the accuracy of the different processes and tools. Inspired by these advances, this paper performs a literature review focused on state-of-the-art glaucoma screening, segmentation, and classification based on images of the papilla and excavation using deep learning techniques. These techniques have been shown to have high sensitivity and specificity in glaucoma screening based on papilla and excavation images. The automatic segmentation of the contours of the optic disc and the excavation then allows the identification and assessment of the glaucomatous disease’s progression. As a result, we verified whether deep learning techniques may be helpful in performing accurate and low-cost measurements related to glaucoma, which may promote patient empowerment and help medical doctors better monitor patients.

show abstract

“…We developed a device based on image sensor and laser module to estimate SD of individual tree and evaluate the accuracy of measurements by comparing their corresponding field measurements. Different from the work of Fan et al [44] and Song et al [37], the collimated beam emitted by the laser module keeps the spot shape constant in natural scenes and can be used as a reference and anchor point. Image sensor can retain texture and shape features in images.…”

Section: Characteristics Of the Devicementioning

confidence: 98%

“…The measured DBH produced RSM E is 2.17 mm and M AE is 15.10 mm. Song et al [37]. constructed an integrated device of digital camera and LiDAR and RM SE of measured DBH is 3.07 mm and BIAS is 0.06 mm.…”

Section: The Analysis Of Major Error Sourcesmentioning

confidence: 99%

“…The price of LiDAR tends to be proportional to the measurement accuracy and range, usually above 23,000.00 dollars. The total cost of the device for Song et al [37] was about 2,000.00 dollars. The economic cost of passive measurement methods using smartphones, on the other hand, is between about 300.00 dollars and 3,000.00 dollars [36,53].…”

Section: Application Scenarios Of the Devicementioning

confidence: 99%

See 1 more Smart Citation

An Automated Method For Stem Diameter Measurement Based on Laser Module and Deep Learning

Wang

¹

,

Li

²

,

Li

³

et al. 2022

Preprint

0

View full text Add to dashboard Cite

Background: Stem diameter (SD) measuring is a crucial foundation for forest resource management. SD measuring in various scenarios necessitates the involvement of expert personnel. As a result, the adoption of automatic measurement methods is greatly expected. In this study, we built a device with an integrated image sensor and laser module to measure the SD of trees of varied diameters. The spot generated by the laser module on the tree stem can be utilized as reference information for measuring height and SD with this device. An end-to-end model is performed to identify the trunk in the panchromatic image, then SD is calculated from the linear relationship between the trunk contour and the spot diameter in pixels. Results: We measured in three scenarios with different land cover types, i.e., transitional woodland/shrub, mixed forest, and green urban area. The overall SD ranges from 2.00 cm to 89.00 cm. To investigate the differences in measurements in different natural scenarios, we analyzed the accuracy of the device for measuring trees of different diameters. The results showed a high consistency between our method and the tape measure measurements. The absolute mean error was 0.36 cm and the root mean square error was 0.45 cm. Compared to most studies, our method demonstrates better versatility and exhibits higher performance. Our integrated device is low cost, portable, and without the assistance of a tripod. Conclusion: Our method achieves the objective of automated measurements and is experimentally proven to possess the potential of efficient and accurate measurement of SD in natural scenarios. Although some limitations remain at present, the device can be employed as a means of forest inventory. In the future, the device can be further explored to be integrated into autonomous mobile robots for more scenarios.

show abstract

A handheld device for measuring the diameter at breast height of individual trees using laser ranging and deep-learning based image recognition

Cited by 15 publications

References 25 publications

Adaptiveness of RGB-image derived algorithms in the measurement of fractional vegetation coverage

Adaptiveness of RGB-image derived algorithms in the measurement of fractional vegetation coverage

Literature Review on Artificial Intelligence Methods for Glaucoma Screening, Segmentation, and Classification

An Automated Method For Stem Diameter Measurement Based on Laser Module and Deep Learning

Contact Info

Product

Resources

About