Estimation of apple tree canopy height and area coverage using 3D LiDAR point clouds

( Rejaul Karim ),( Mohammod Ali ),( Shahriar Ahmed ),( Ashrafuzzaman Gulandaz ),( Nasim Reza ),( Justin Sung ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.1

Phenotyping characteristics (Tree height and canopy area) are crucial factors for growth and yield monitoring. This study aimed to use 3D LiDAR point clouds to estimate the apple tree canopy height and area using different digital models. The traditional method of manually measuring trees at various heights and canopies is time-consuming and labor-intensive, so the study aimed to automate the process using 3D LiDAR (VLP-16) to collect point clouds of orchard apple trees. The data was pre-processed using a 3D point cloud data processing software, and automatic segmentation methods were applied to calculate the canopy height and area for selected orchard tree samples. The processed 3D point cloud data was converted into raster images for visualization and estimation of orchard tree canopy height and area coverage using digital surface model (DSM), digital elevation model (DEM), and canopy height model (CHM). Python program was also used for visualization and reconstruction of trees from the preprocessed data. The accuracy of the sensor-based measuring method was compared to manually-acquired ground truth data, but the accuracy was worse by 15%. The study found that the proposed system could efficiently segment and measure tree canopy height and area coverage. The proposed models showed comparatively lower result than manual measurement, with an sensor based average tree canopy height and area of 2.1 m and 5.83 m², respectively, where as measured values were 2.4±0.2 m and 6.0±0.21 m², respectively. However, the findings of this study can still contribute to further horticultural crops research particularly for orchard fruits production and yield monitoring.

3D LiDAR-based Quantification of Phenotypic Traits and Land Characteristics in Rice Farming

( Md Rejaul Karim ),( Mohammod Ali ),( Shahriar Ahmed ),( Md Shaha Nur Kabir ),( Md Nasim Reza ),( Justin Sung ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.2

Phenotypic and land characteristics information plays a crucial role in effective management of rice farming. The utilization of LiDAR based object recognition as well as visualization provides a rapid and precise assessment of the phenotypic traits of rice plants. This study aimed to quantify the rice plant phenotypic and land characteristics using a 3D LiDAR. A data collection structure made of aluminum profile and a LiDAR sensor (i.e., VLP-16) mounted on the structure was used to collect 3D point cloud data from rice field. A rice field of RDA at Iksan in Korea was selected for data acquisition. Ten numbers of small plots considering the area of LiDAR data frames exhibiting diverse plant height, shapes, and sizes were randomly selected. From each LiDAR scanned data frame, a region of interest (RoI) segmented for sensor based processing and measurements. Commercial software utilized for segmentation and python-based programing codes also applied to process the collected data for visualization and measurements. The accuracy of the estimated outputs from the point cloud was evaluated by comparison with measured values collected randomly from ten spots remaining in the sensor-based data frame. The estimated plant heights from the point cloud were 0.84±0.03 m, while the measured heights were 0.77±0.03 m. The root mean square error (RMSE) for plant height estimation was 0.08 m, and the simple linear coefficient of determination (r²) was 0.88. Regarding the segment wise canopy volume, point cloud estimations were 1.01±0.06 m³, compared to the measured volume of 1.18±0.03 m³. The RMSE for canopy volume estimation was 0.18 m³, with r² of 0.87, indicating a high level of accuracy. For hill-to-hill distance and intra-row spacing, the point cloud measurements were 0.35±0.01 m, and 0.34±0.02 m, respectively, while the measurements were 0.30±0.03 m, and 0.30±0.03 m, respectively. The RMSE and r² for hill to hill distance were 0.04 m and 0.92, respectively, and for row distance, 0.03 m and 0.87, respectively. Despite minor differences, there was a strong relationship and close agreement between the estimation using point cloud data and measurements. The findings highlight the reliability and efficiency of the 3D LiDAR technology for accurately measuring phenotypic traits and land characteristics for maximizing rice cultivation.

Assessment of Precision and Accuracy of Cropcircle and MicaSense Sensors for Monitoring Crop Growth Dynamics During NDVI Calculation

( Asrakul Haque ),( Rejaul Karim ),( Shahriar Ahmed ),( Nasim Reza ),( Ka Young Lee ),( Yeong Ho Kang ),( Keong Do Lee ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.1

The Normalized Difference Vegetation Index (NDVI) is a widely used index for monitoring crop growth and diagnosing nitrogen status. This study aimed to evaluate the precision and accuracy of Cropcircle and MicaSense sensors for monitoring crop growth dynamics using NDVI. The assessment was conducted under varying vegetative conditions and growth stages of the crops. In this experiment, a GPS unit and a combination of an active (Cropcircle ACS-435) and a passive sensor (MicaSense RedEdge MX) mounted on a frame were used. The data was taken at a height of 1m to cover the canopy area of wheat crop. Two plots with low and high vegetation levels were examined during Feekes 1 and Feekes 2-3. The coefficients of determination (R2) between Cropcircle and MicaSense sensors for low and high vegetation plots were 1E-6 and 0.29, respectively, for Feekes 1 and increased to 0.43 and 0.73, respectively, in Feekes 2-3. For Feekes 1, the average mean for the crop-circle and MicaSense NDVI values were 0.13 ± 0.02 and 0.18 ± 0.02, respectively, for low vegetation and 0.23 ± 0.08 and 0.25 ± 0.04, respectively, for the high vegetation plots. For Feekes 2-3, the mean for the crop-circle and MicaSense NDVI values were 0.13 ± 0.04 and 0.19 ± 0.03, respectively, for low vegetation and 0.23 ± 0.01 and 0.27 ± 0.07, respectively, for the high vegetation plots. The results showed that the high vegetative plot had a significant impact on NDVI calculation using crop canopy sensors than the low vegetative plot. Therefore, it is suggested to consider vegetative state of crops for accurate and precise monitoring of crop growth using crop canopy sensors.

Estimation of pepper plant height and canopy area under field conditions using an image processing approach

( Mohammad Ali ),( Rejaul Karim ),( Ashrafuzzaman Gulandaz ),( Eliezel Habineza ),( Sazzadul Kabir ),( Ho-sung An ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.1

Regular crop phenotypic trait monitoring is a crucial agricultural field management technique for analyzing crop growth and predicting yield. The aim of this study was to estimate the height and canopy area of pepper plants using automatically captured images for monitoring plant growth under field conditions. Four Raspberry Pi cameras were mounted on a crop-scouting electric vehicle platform, and used in an experimental field to collect random images of pepper plants. The recorded images were processed with a commercial and open-source data-processing software. Automatic data collection and segmentation methods were tested on two rows with 83 pepper plants. The image used in the measuring method was compared to the measured ground truth data to evaluate the accuracy of the results. The average plant height and canopy area were 60.43±7.92 cm and 0.32±0.13 m2, respectively. The sensor data processing algorithm showed an RMSE of 3.11 to 3.62 cm and 0.04 to 0.07 m2 for plant height and canopy average estimations, respectively. The R2 values were 0.85 for the individual phenotype traits of plant height and 0.82 for canopy area coverage, respectively. The results showed that the proposed system could automatically segment and measure pepper plant height and canopy area under field conditions. The findings of this study would contribute to further research on upland crop growth, and yield monitoring.

Measurement of pepper plant height and canopy area using 3D LiDAR point cloud

알리모하마드 ( Mohammod Ali ),카림레자울 ( Rejaul Karim ),카비르사자둘 ( Sazzadul Kabir ),구란다즈아스라푸자만 ( Ashrafuzzaman Gulandaz ),레자나심 ( Nasim Reza ),선저스틴 ( Justsung ),정선옥 ( Sun-ok Chung ) 한국농업기계학회 2022 한국농업기계학회 학술발표논문집 Vol.27 No.2

Plant height and canopy area are crucial plant factors for growth and yield monitoring. A significant number of plants at different heights and canopies are required to evaluate the plant phenotyping characteristics, which is labor-intensive and time-consuming. Therefore, the aim of this study was to use 3D LiDAR point clouds to assess the height and canopy area of pepper plants. A LiDAR (VLP-16) was installed in the experimental field to collect the 3D point clouds of pepper plants. The collected data was preprocessed with 3D point cloud data processing software. The automatic segmentation methods were tested on 13 pepper plants to calculate the height and canopy area. The 3D LiDAR point clouds used in the measuring method were compared to the manually gathered ground truth to determine the accuracy of the results. The average plant height and canopy area were found to be 66±4.5 m and 0.48±0.11 m2, respectively, by manual measurement, whereas the 3D point cloud data processing algorithm showed less accuracy. The R2 values were found to be more than 0.89 for the individual phenotypic traits. The results showed that the proposed system could automatically segment and measure plant height and canopy area. The findings of this study would contribute to further research for upland crop growth and yield monitoring.

Design of an ICT-based monitoring system of soil water content and irrigation control operation for orchards

( Shahriar Ahmed ),( Nasim Reza ),( Rejaul Karim ),( Kayoung Lee ),( Heetae Kim ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.1

Insufficient or excessive water can impede root growth and lead to the leaching of minerals from the root zone, resulting in a nutritional deficiency. In addition, excessive water supply encourages crown and collar rot. Irrigation scheduling can be regulated based on soil water stress. A malfunctioning irrigation scheduling system can also compromise environmental control and inhibit crop growth, so it is crucial that components used in the system work accurately. The goal of this project was to create an automatic irrigation control system, based on real-time soil water content monitoring in orchard soil, and could be accessible remotely via the internet, while also detecting control failures. Experiments were conducted inside a greenhouse, in a 9-m2 soil bin. A Python program was coded to operate the irrigation pump and solenoid valves through a micro-controller utilizing water content information from sensors, as well as to transfer sensor data to a database system to observe the operation through the internet. The experiment showed that all sensor data could be collected and sent to the database system. The average water content values for the two channels were 18.3% and 16.6% during the irrigation period and 24.6% and 27.1% during the non-irrigation period, respectively. During the control operation, it was found that the first solenoid valve consumed 20.2 watts in average for “ON” state and 11.12 watts for “OFF” state. However, there was no significant change in power consumption of the second solenoid valve between the “ON” and “OFF” states. Therefore, the actuator used in the system showed a faulty condition, which can be corrected by further investigation.

Soil water content distribution mapping using an automatic monitoring system

( Nasim Reza ),( Eliezel Habineza ),( Rejaul Karim ),( Mohammod Ali ),( Shaha Nur Kabir ),( Young Yoon Jang ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.1

Soil water content plays a crucial role in plant growth, irrigation scheduling, and soil erosion prediction. Automatic sensor-based monitoring systems have emerged as efficient tools to provide continuous soil water content mapping against the traditional method, which is time consuming and limited to single point measurement. This study aimed to develop a sensor-based monitoring system for real-time mapping of soil water content distribution. To assess the variability of soil water content in sandy soil, a soil test bin of 3 m by 3 m was constructed. The system consisted of a series of sensors (SEN0193) installed at different depths, ranging from 0 to 60 cm. The monitoring system was equipped with wireless transmission technology using Arduino Mega 2560 and Raspberry Pi 4B microcontroller. Water content sensors were placed at predetermined locations and the geographic coordinates were obtained using GPS. The microcontroller collected data from the sensors, which was then evaluated using GIS to prepare a map of the soil moisture distribution. The results suggested that the monitoring system had the potential to revolutionize soil water content mapping and monitoring. The system can provide valuable insights into the spatial and temporal variations of soil water content, which can inform irrigation scheduling, crop management, and soil conservation practices.

Evaluating the Accuracy of FOV Alignment for Micasense Multispectral Imagery in VI Calculation

( Md Asrakul Haque ),( Md Rejaul Karim ),( Md Razob Ali ),( Shaha Nur Kabir ),( Keong Do Lee ),( Yeong Ho Kang ),( Sun-ok Chung ) 한국농업기계학회 2023 한국농업기계학회 학술발표논문집 Vol.28 No.2

Multispectral imagery is pivotal for vegetation index (VI) analysis, shaping crop nutritional management strategies and advancing precision agriculture. Yet, the efficacy of image enhancement techniques in VI calculation remains a critical inquiry. This study addresses this gap by evaluating various image enhancement methods for multispectral imagery, focusing on the widely accepted Normalized Differential Vegetation Index (NDVI). We employed a multispectral sensor, the MicaSense RedEdge MX, alongside an active sensor, the Crop-circle ACS-435, to assess NDVI calculation performance. Our objective was to assess the accuracy of the Field of View (FOV) alignment of MicaSense with the active sensor. Data collection occurred across four distinct wheat growth stages (GS1, GS2, GS3, and GS4) utilizing a handheld structure equipped with Crop Circle ACS 435, MicaSense RedEdge MX, and a Topcon Hiper VR GNSS rover. This setup maintained a consistent 90cm canopy height based on the plot width. Python programming facilitated GPS location processing and image segmentation based on pixel coordinates, mirroring the Crop-circle FOV. We extracted reflectance data from the segmented portion of each band and calculated NDVI using Red and NIR reflectance data. Data enhancement techniques were assessed by comparing enhanced and raw image data against standardized data from the Crop-circle sensor. Regression analysis, including the coefficient of determination (R2) and root mean square error (RMSE), was utilized for evaluation. The application of the FOV enhancement technique to MicaSense images yielded significant improvements in regression metrics (R2 and RMSE) across GS1, GS2, GS3, and GS4. Notably, FOV enhancement resulted in R2 increases of 50%, 18%, 16%, and 4% and RMSE values of0.06, 0.05, 0.06, and 0.03, respectively. The most substantial accuracy enhancements were observed in GS1 (50%), indicating varying effectiveness based on vegetation growth stage and density. This study underscores the critical role of multispectral imagery and the efficacy of FOV alignment in improving NDVI calculation accuracy. These findings hold valuable implications for future research and precision agriculture practices.

Real-time sound monitoring using 2D convolutional neural network (CNN) for pig diseases symptoms detection in pig farm

레자나심 ( Nasim Reza ),하케아스라쿨 ( Asrakul Haque ),카림레자울 ( Rejaul Karim ),송민호 ( Minho Song ),김국환 ( Gookhwan Kim ),정선옥 ( Sun-ok Chung ) 한국농업기계학회 2022 한국농업기계학회 학술발표논문집 Vol.27 No.2

Monitoring and preventing diseases in livestock is essential for modern farming, and an early warning system may significantly reduce the economic impact of diseased events. Manual monitoring of pigs in a pig farm is time consuming and labor intensive. Automatic monitoring for pig diseases may help to control the spread of infections. The purpose of this research was to identify the signs of sickness in pigs using acoustic monitoring in real-time. Two microphones were installed in the pig farm for automatic sound acquisition. The sound signals were converted into spectrograms by fast fourier transform (FFT) and mel-frequency cepstral coefficients (MFCC) as a characteristic parameter. Using a 2D convolutional neural network (CNN) and features extracted from the spectrogram, we presented a classification approach for real-time application. The conversion of sound inputs into spectrograms made it possible to recognize by the use of CNN. For the real-time detection, the proposed algorithm showed the ability to recognize the sounds of cough, sneezing, scream, and crushing with an overall recognition accuracy of 75.8%, 69.5%, 72.4%, and 71.6%, respectively, and an average F1-score of 81.2%. Future work is needed to enhance sound detection robustness.

Thermal imaging and computer vision technologies for the enhancement of pig husbandry: a review

Md Nasim Reza,Md Razob Ali,Samsuzzaman,Md Shaha Nur Kabir,Md Rejaul Karim,Shahriar Ahmed,Hyunjin Kyoung,김국환,Sun-Ok Chung 한국축산학회 2024 한국축산학회지 Vol.66 No.1

Pig farming, a vital industry, necessitates proactive measures for early disease detection and crush symptom monitoring to ensure optimum pig health and safety. This review explores advanced thermal sensing technologies and computer vision-based thermal imaging techniques employed for pig disease and piglet crush symptom monitoring on pig farms. Infrared thermography (IRT) is a non-invasive and efficient technology for measuring pig body temperature, providing advantages such as non-destructive, long-distance, and high-sensitivity measurements. Unlike traditional methods, IRT offers a quick and labor-saving approach to acquiring physiological data impacted by environmental temperature, crucial for understanding pig body physiology and metabolism. IRT aids in early disease detection, respiratory health monitoring, and evaluating vaccination effectiveness. Challenges include body surface emissivity variations affecting measurement accuracy. Thermal imaging and deep learning algorithms are used for pig behavior recognition, with the dorsal plane effective for stress detection. Remote health monitoring through thermal imaging, deep learning, and wearable devices facilitates non-invasive assessment of pig health, minimizing medication use. Integration of advanced sensors, thermal imaging, and deep learning shows potential for disease detection and improvement in pig farming, but challenges and ethical considerations must be addressed for successful implementation. This review summarizes the state-of-the-art technologies used in the pig farming industry, including computer vision algorithms such as object detection, image segmentation, and deep learning techniques. It also discusses the benefits and limitations of IRT technology, providing an overview of the current research field. This study provides valuable insights for researchers and farmers regarding IRT application in pig production, highlighting notable approaches and the latest research findings in this field.