Tofael Ahamed

Dr. Tofael Ahamed

Associate Professor, Faculty and Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennodai, 305-8572, Tsukuba, Japan

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 July 2021 in Sensors
Verified
Journal Article
Open access
PDF file
Addie Ira ParicoAddie Ira Parico
Addie Ira Parico
Tofael AhamedTofael Ahamed
Tofael Ahamed
This study aimed to produce a robust real-time pear fruit counter for mobile applications using only RGB data, the variants of the state-of-the-art object detection model YOLOv4, and the multiple object-tracking algorithm Deep SORT. This study also provided a systematic and pragmatic methodology for choosing the most suitable model for a desired application in agricultural sciences. In terms of accuracy, YOLOv4-CSP was observed as the optimal model, with an AP@0.50 of 98%. In terms of speed and computational cost, YOLOv4-tiny was found to be the ideal model, with a speed of more than 50 FPS and FLOPS of 6.8–14.5. If considering the balance in terms of accuracy, speed and computational cost, YOLOv4 was found to be most suitable and had the highest accuracy metrics while satisfying a real time speed of greater than or equal to 24 FPS. Between the two methods of counting with Deep SORT, the unique ID method was found to be more reliable, with an F1count of 87.85%. This was because YOLOv4 had a very low false negative in detecting pear fruits. The ROI line is more reliable because of its more restrictive nature, but due to flickering in detection it was not able to count some pears despite their being detected.

25 October 2019 in GeoJournal
Verified
Journal Article
Closed access
PDF file
MHMuhammad Iqbal Habibie
Muhammad Iqbal Habibie
RNRyozo Noguchi
Ryozo Noguchi
MSMatsushita Shusuke
Matsushita Shusuke
Tofael AhamedTofael Ahamed
Tofael Ahamed

23 February 2021 in Land
Verified
Journal Article
Open access
PDF file
RMRubaiya Mostafiz
Rubaiya Mostafiz
RNRyozo Noguchi
Ryozo Noguchi
Tofael AhamedTofael Ahamed
Tofael Ahamed
Satellite remote sensing technologies have a high potential in applications for evaluating land conditions and can facilitate optimized planning for agricultural sectors. However, misinformed land selection decisions limit crop yields and increase production-related costs to farmers. Therefore, the purpose of this research was to develop a land suitability assessment system using satellite remote sensing-derived soil-vegetation indicators. A multicriteria decision analysis was conducted by integrating weighted linear combinations and fuzzy multicriteria analyses in a GIS platform for suitability assessment using the following eight criteria: elevation, slope, and LST vegetation indices (SAVI, ARVI, SARVI, MSAVI, and OSAVI). The relative priorities of the indicators were identified using a fuzzy expert system. Furthermore, the results of the land suitability assessment were evaluated by ground truthed yield data. In addition, a yield estimation method was developed using indices representing influential factors. The analysis utilizing equal weights showed that 43% of the land (1832 km2) was highly suitable, 41% of the land (1747 km2) was moderately suitable, and 10% of the land (426 km2) was marginally suitable for improved yield productions. Alternatively, expert knowledge was also considered, along with references, when using the fuzzy membership function; as a result, 48% of the land (2045 km2) was identified as being highly suitable; 39% of the land (2045 km2) was identified as being moderately suitable, and 7% of the land (298 km2) was identified as being marginally suitable. Additionally, 6% (256 km2) of the land was described as not suitable by both methods. Moreover, the yield estimation using SAVI (R2 = 77.3%), ARVI (R2 = 68.9%), SARVI (R2 = 71.1%), MSAVI (R2 = 74.5%) and OSAVI (R2 = 81.2%) showed a good predictive ability. Furthermore, the combined model using these five indices reported the highest accuracy (R2 = 0.839); this model was then applied to develop yield prediction maps for the corresponding years (2017–2020). This research suggests that satellite remote sensing methods in GIS platforms are an effective and convenient way for agricultural land-use planners and land policy makers to select suitable cultivable land areas with potential for increased agricultural production.

16 May 2023 in Sensors
Verified
Journal Article
Open access
PDF file
AJAilian Jiang
Ailian Jiang
Tofael AhamedTofael Ahamed
Tofael Ahamed
Traditional Japanese orchards control the growth height of fruit trees for the convenience of farmers, which is unfavorable to the operation of medium- and large-sized machinery. A compact, safe, and stable spraying system could offer a solution for orchard automation. Due to the complex orchard environment, the dense tree canopy not only obstructs the GNSS signal but also has effects due to low light, which may impact the recognition of objects by ordinary RGB cameras. To overcome these disadvantages, this study selected LiDAR as a single sensor to achieve a prototype robot navigation system. In this study, density-based spatial clustering of applications with noise (DBSCAN) and K-means and random sample consensus (RANSAC) machine learning algorithms were used to plan the robot navigation path in a facilitated artificial-tree-based orchard system. Pure pursuit tracking and an incremental proportional–integral–derivative (PID) strategy were used to calculate the vehicle steering angle. In field tests on a concrete road, grass field, and a facilitated artificial-tree-based orchard, as indicated by the test data results for several formations of left turns and right turns separately, the position root mean square error (RMSE) of this vehicle was as follows: on the concrete road, the right turn was 12.0 cm and the left turn was 11.6 cm, on grass, the right turn was 12.6 cm and the left turn was 15.5 cm, and in the facilitated artificial-tree-based orchard, the right turn was 13.8 cm and the left turn was 11.4 cm. The vehicle was able to calculate the path in real time based on the position of the objects, operate safely, and complete the task of pesticide spraying.

7 April 2023 in Sensors
Verified
Journal Article
Open access
PDF file
RM Rasika D AbeyrathnaRM Rasika D Abeyrathna
RM Rasika D Abeyrathna
Victor Massaki NakaguchiVictor Massaki Nakaguchi
Victor Massaki Nakaguchi
AMArkar Minn
Arkar Minn
Tofael AhamedTofael Ahamed
Tofael Ahamed
Recognition and 3D positional estimation of apples during harvesting from a robotic platform in a moving vehicle are still challenging. Fruit clusters, branches, foliage, low resolution, and different illuminations are unavoidable and cause errors in different environmental conditions. Therefore, this research aimed to develop a recognition system based on training datasets from an augmented, complex apple orchard. The recognition system was evaluated using deep learning algorithms established from a convolutional neural network (CNN). The dynamic accuracy of the modern artificial neural networks involving 3D coordinates for deploying robotic arms at different forward-moving speeds from an experimental vehicle was investigated to compare the recognition and tracking localization accuracy. In this study, a Realsense D455 RGB-D camera was selected to acquire 3D coordinates of each detected and counted apple attached to artificial trees placed in the field to propose a specially designed structure for ease of robotic harvesting. A 3D camera, YOLO (You Only Look Once), YOLOv4, YOLOv5, YOLOv7, and EfficienDet state-of-the-art models were utilized for object detection. The Deep SORT algorithm was employed for tracking and counting detected apples using perpendicular, 15°, and 30° orientations. The 3D coordinates were obtained for each tracked apple when the on-board camera in the vehicle passed the reference line and was set in the middle of the image frame. To optimize harvesting at three different speeds (0.052 ms−1, 0.069 ms−1, and 0.098 ms−1), the accuracy of 3D coordinates was compared for three forward-moving speeds and three camera angles (15°, 30°, and 90°). The mean average precision (mAP@0.5) values of YOLOv4, YOLOv5, YOLOv7, and EfficientDet were 0.84, 0.86, 0.905, and 0.775, respectively. The lowest root mean square error (RMSE) was 1.54 cm for the apples detected by EfficientDet at a 15° orientation and a speed of 0.098 ms−1. In terms of counting apples, YOLOv5 and YOLOv7 showed a higher number of detections in outdoor dynamic conditions, achieving a counting accuracy of 86.6%. We concluded that the EfficientDet deep learning algorithm at a 15° orientation in 3D coordinates can be employed for further robotic arm development while harvesting apples in a specially designed orchard.

4 January 2023 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Md. ShamsuzzohaMd. Shamsuzzoha
Md. Shamsuzzoha
Tofael AhamedTofael Ahamed
Tofael Ahamed
Bangladesh is a global south hotspot due to climate change and sea level rise concerns. It is a highly disaster-prone country in the world with active deltaic shorelines. The shorelines are quickly changing to coastal accretion and erosion. Erosion is one of the water hazards to landmass sinking, and accretion relates to land level rises due to sediment load deposition on the Bay of Bengal continental shelf. Therefore, this study aimed to explore shoreline status with change assessment for the three study years 1991, 2006, and 2021 using satellite remote sensing and geographical information system (GIS) approaches. Landsat 5, 7 ETM+, and 8 OLI satellite imageries were employed for onshore tasseled cap transformation (TCT) and land and sea classification calculations to create shore boundaries, baseline assessment, land accretion, erosion, point distance, and near feature analysis. We converted 16,550 baseline vertices to points as the study ground reference points (GRPs) and validated those points using the country datasheet collected from the Survey of Bangladesh (SoB). We observed that the delta’s shorelines were changed, and the overall lands were accredited for the land-increasing characteristics analysis. The total accredited lands in the coastal areas observed during the time periods from 1991 to 2006 were 825.15 km2, from 2006 to 2021 was 756.69 km2, and from 1991 to 2021 was 1223.94 km2 for the 30-year period. Similarly, coastal erosion assessment analysis indicated that the results gained for the period 1991 to 2006 and 2006 to 2021 were 475.87 km2 and 682.75 km2, respectively. Therefore, the total coastal erosion was 800.72 km2 from 1991 to 2021. Neat accretion was 73.94 km2 for the 30-year period from 1991 to 2021. This research indicates the changes in shorelines, referring to the evidence for the delta’s active formation through accretion and erosion processes of ‘climate change’ and ‘sea level rise’. This research projects the erosion process and threatens land use changes toward agriculture and settlements in the coastal regions of Bangladesh.