Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Zafar Iqbal has a total of 34 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Handheld UV-C Fluorescence Imaging with Cascaded Denoising and Deep Learning for Escherichia coli Concentration Classification on Diverse Food-Contact Surfaces

    by Zafar Iqbal, Thomas F. Burks, Sriram Thota, Akash Singh, Snehit Vaddi, Pappu Yadav, Jianwei Qin, Moon S. Kim, Mark A. Ritenour, Jiuxu Zhang, et al.

    Rapid, non-destructive assessment of Escherichia coli contamination levels on food-contact surfaces is important for food safety monitoring and sanitation verification. This study developed a deep-learning framework integrating UV-C fluorescence imaging and cascaded denoising to classify predefined E. coli concentration levels on four food-contact surfaces. A SafetySpect CSI-D+ handheld system acquired fluorescence images at four concentration levels: Control (0), 106, 107, and 108 CFU/mL, on stainless steel, white rubber, white plastic, and black plastic. Five image-denoising operators (wavelet thresholding, NLM, unsharp masking, Wiener filtering, and RL-TV) were evaluated in sequential and parallel configurations. Eight deep-learning classifiers, including five YOLO11-cls variants, EfficientNetB7, ConvNeXtBase, and ViT, were benchmarked using consistent evaluation protocols. Classification performance varied with surface type and denoising strategy. The sequential Wavelet → Wiener → RL-TV cascade achieved the highest observed mean test accuracy (94.18%), with surface-specific accuracies of 97.57%, 97.57%, 94.44%, and 87.15% for stainless steel, white rubber, black plastic, and white plastic, respectively. On the most challenging surface (white plastic), the cascade improved accuracy by 11.46 percentage points over the best single operator. YOLO11-based models generally outperformed ViT, although the highest-performing architecture varied by surface. Most misclassifications occurred primarily between 106 and 107 CFU/mL, while the 108 CFU/mL class was consistently classified accurately. Overall, the ordered cascaded denoising can improve robust E. coli concentration classification across diverse food-contact surfaces, providing a promising foundation for rapid, non-destructive food safety inspection. The results also highlight the importance of surface-specific preprocessing and deep learning architectures for robust performance across diverse materials.

  • Open Access

    Detecting Escherichia coli on Conventional Food Processing Surfaces Using UV-C Fluorescence Imaging and Deep Learning

    by Zafar Iqbal, Thomas F. Burks, Snehit Vaddi, Pappu Kumar Yadav, Quentin Frederick, Satya Aakash Chowdary Obellaneni, Jianwei Qin, Moon Kim, Mark A. Ritenour, Jiuxu Zhang, et al.

    Detecting Escherichia coli on food preparation and processing surfaces is critical for ensuring food safety and preventing foodborne illness. This study focuses on detecting E. coli contamination on common food processing surfaces using UV-C fluorescence imaging and deep learning. Four concentrations of E. coli (0, 105, 107, and 108 colony forming units (CFU)/mL) and two egg solutions (white and yolk) were applied to stainless steel and white rubber to simulate realistic contamination with organic interference. For each concentration level, 256 droplets were inoculated in 16 groups, and fluorescence videos were captured. Droplet regions were extracted from the video frames, subdivided into quadrants, and augmented to generate a robust dataset, ensuring 3–4 droplets per sample. Wavelet-based denoising further improved image quality, with Haar wavelets producing the highest Peak Signal-to-Noise Ratio (PSNR) values, up to 51.0 dB on white rubber and 48.2 dB on stainless steel. Using this dataset, multiple deep learning (DL) models, including ConvNeXtBase, EfficientNetV2L, and five YOLO11-cls variants, were trained to classify E. coli concentration levels. Additionally, Eigen-CAM heatmaps were used to visualize model attention to bacterial fluorescence regions. Across four dataset groupings, YOLO11-cls models achieved consistently high performance, with peak test accuracies of 100% on white rubber and 99.60% on stainless steel, even in the presence of egg substances. YOLO11s-cls provided the best balance of accuracy (up to 98.88%) and inference speed (4–5 ms) whilst having a compact size (11 MB), outperforming larger models such as EfficientNetV2L. Classical machine learning models lagged significantly behind, with Random Forest reaching 89.65% accuracy and SVM only 67.62%. Overall, the results highlight the potential of combining UV-C fluorescence imaging with deep learning for rapid and reliable detection of E. coli on stainless steel and rubber conveyor belt surfaces. Additionally, this approach could support the design of effective interventions to remove E. coli from food processing environments.


  • Open Access

    Detecting Escherichia coli Contamination on Plant Leaf Surfaces Using UV-C Fluorescence Imaging and Deep Learning

    by Snehit Vaddi, Thomas F. Burks, Zafar Iqbal, Pappu Kumar Yadav, Quentin Frederick, Satya Aakash Chowdary Obellaneni, Jianwei Qin, Moon Kim, Mark A. Ritenour, Jiuxu Zhang, et al.

    The transmission of Escherichia coli through contaminated fruits and vegetables poses serious public health risks and has led to several national outbreaks in the USA. To enhance food safety, rapid and reliable detection of E. coli on produce is essential. This study evaluated the performance of the CSI-D+ system combined with deep learning for detecting varying concentrations of E. coli on citrus and spinach leaves. Eight levels of E. coli contamination, ranging from 0 to 108 colony-forming units (CFU)/mL, were inoculated onto the leaf surfaces. For each concentration level, 10 droplets were applied to 8 citrus and 12 spinach leaf samples (2 cm in diameter), and fluorescence images were captured. The images were then subdivided into quadrants, and several post-processing operations were applied to generate the final dataset, ensuring that each sample contained at least 2–3 droplets. Using this dataset, multiple deep learning (DL) models, including EfficientNetB7, ConvNeXtBase, and five YOLO11 variants (n, s, m, l, x), were trained to classify E. coli concentration levels. Additionally, Eigen-CAM heatmaps were used to visualize the spatial responses of the models to bacterial presence. All YOLO11 models outperformed EfficientNetB7 and ConvNeXtBase. In particular, YOLO11s-cls was identified as the best-performing model, achieving average validation accuracies of 88.43% (citrus) and 92.03% (spinach), and average test accuracies of 85.93% (citrus) and 92.00% (spinach) at a 0.5 confidence threshold. This model demonstrated an inference speed of 0.011 s per image with a size of 11 MB. These findings indicate that fluorescence-based imaging combined with deep learning for rapid E. coli detection could support timely interventions to prevent contaminated produce from reaching consumers.


  • A Robust Framework Combining Hyperspectral Imaging and Machine Learning for Assessing Sudden Death Syndrome (SDS) Severity in Soybean Foliage

    by Zafar Iqbal, Thomas Burks, Pappu Yadav, Satya Aakash Chowdary Obellaneni, Inayat Rasool, Quentin Frederick, Jianwei Qin, Moon Kim


  • Open Access

    Spatial and Temporal Variability of Environmental Variables in Chinese Solar Greenhouses in the Summer Season

    by Nafiul Islam, Nasim Reza, Zafar Iqbal, Kyu-Ho Lee, Moon-Ki Jang, Sun-Ok Chung

    To ensure a high crop profit in Chinese solar greenhouses (CSGs), it is crucial to effectively manage major environmental variables such as temperature, humidity, and CO2 concentrations, among others, to mitigate harmful effects on crop growth. The objectives of this study were to assess the spatial, vertical, and temporal variability of major environmental variables in CSGs during summer, and to provide fundamental information that could facilitate the monitoring and control of environmental factors in CSGs. The experiments were conducted in two CSGs: one with crops and another without crops. The measured environmental variables included air temperature, humidity, CO2 concentration, light intensity, and wind conditions. Significant variations in the spatial, vertical, and temporal distribution of environmental factors were observed in both greenhouses. The results revealed significant diurnal patterns in temperature and humidity, with higher daytime temperatures and lower humidity levels. The greenhouse with crops exhibited warmer bottom layers due to restricted air mobility. CO2 concentrations peaked at night, aligning with plants’ respiration and photosynthesis cycles, whereas light intensity showed substantial daytime peaks, slightly affected by the presence of crops. The study emphasized the necessity of stratified control of the environment and dynamic management of CO2. The deployment of a wireless sensor network (WSN) and placement of an error-based sensor ensured precise monitoring, highlighting the importance of continuous data collection and adaptive management for optimal greenhouse conditions.


  • Open Access

    Spatial, Temporal, and Vertical Variability of Ambient Environmental Conditions in Chinese Solar Greenhouses during Winter

    by Nasim Reza, Nafiul Islam, Zafar Iqbal, Shaha Nur Kabir, Milon Chowdhury, Ashrafuzzaman Gulandaz, Mohammod Ali, Moon-Ki Jang, Sun-Ok Chung

    The monitoring and control of environmental conditions are crucial as they influence crop quality and yield in Chinese solar greenhouses (CSGs). The objectives of this study were to assess the spatial, temporal, and vertical variability of major environmental parameters in CSGs during winter and to provide greenhouse climate/microclimate characteristics in order to facilitate the monitoring and control of greenhouse environmental conditions. A wireless sensor network (WSN) was deployed in two CSGs: one with crops and one without. Sensors were placed at different locations inside and outside the greenhouses, and the air temperature, humidity, CO2 concentration, light intensity, solar radiation, and wind conditions were measured and analyzed. Significant variability in the spatial, temporal, and vertical distribution of environmental factors was observed in both greenhouses. The average minimum and maximum temperatures and humidity inside the CSG with crops were 9.96 °C (4:00 h) and 24.5 °C (12:00 h), and 32.6% (12:00 h) and 92.1% (5:00 h), respectively. The temperature difference was 2.2 °C between layers in the CSG without crops and 1.4 °C between layers in the CSG with crops. The CO2 concentration in the different layers inside the CSG with crops was highest at night. The average maximum light intensity inside the CSG with crops was 32,660.19 lx, 36,618.12 lx, and 40,660.48 lx (12:00 h to 13:00 h) in the bottom, middle, and top layers, respectively. Sensor positioning in the greenhouse was evaluated by considering the sensors’ data variability. The findings of this study could aid in the development of a better monitoring and control system for CSG’s microclimate during winter. More research is needed on greenhouse microclimate control systems based on this variability analysis, which could improve crop quality and yield in greenhouses.


  • Open Access

    Evaluation of a 0.7 kW Suspension-Type Dehumidifier Module in a Closed Chamber and in a Small Greenhouse

    by Nafiul Islam, Zafar Iqbal, Mohammod Ali, Ashrafuzzaman Gulandaz, Shaha Nur Kabir, Seung-Ho Jang, Sun-Ok Chung

    Controlling humidity inside greenhouses is crucial for optimum plant growth and controlling physiological disorders and diseases. The humidity response and uniformity depend extensively on the evaluation of the dehumidifier. The objective of this research was to evaluate a low-powered suspension-type dehumidifier module in terms of humidity changes and spatial and vertical variability in a closed chamber and in a small greenhouse. A wireless sensor network including 27 sensor nodes was used to collect the data during the humidity changes from 80% to 70% and 90% to 70%. The humidity response results showed that the times required for dehumidification from 80% to 70% and 90% to 70% were 13.75 and 21.51 min, respectively, for the closed-chamber operation. Similarly, for the small greenhouse, 18 and 35 min were required to reduce the humidity levels from 80% to 70% and 90% to 70%, respectively. The spatial and variability results indicated that the changes in humidity at the rear and bottom layers were slower than those in the other layers of both experimental areas. The findings of this study would aid in the development of dehumidification strategies and sustainable agriculture for monitoring and controlling humidity in greenhouses using low-powered dehumidifiers.


  • Open Access

    Stress and Fatigue Analysis of Picking Device Gears for a 2.6 kW Automatic Pepper Transplanter

    by Nafiul Islam, Zafar Iqbal, Milon Chowdhury, Mohammod Ali, Kiraga Shafik, Shaha Nur Kabir, Dae-Hyun Lee, Sun-Ok Chung

    A seedling picking device is an essential component for an automatic transplanter to automatically convey the seedling to the dibbling part. It is necessary to find the appropriate material and dimensions for the picking device gears to avoid mechanical damage and increase their durability. Therefore, the objectives of this research were to analyze the stress of a picking device gear mechanism in order to select suitable materials and dimensions, and to predict the fatigue life by considering the damage level. The picking device gear shaft divided the input power into two categories, i.e., crank and cam gear sets. Finite element analysis simulation and American Gear Manufacturers Association standard stress analysis theory tests were conducted on both of the crank and cam gear sets for different materials and dimensions. A test bench was fabricated to collect the load (torque) data at different gear operating speeds. The torque data were analyzed using the load duration distribution method to observe the cyclic load patterns. The Palmgren–Miner cumulative damage rule was used to determine the damage level of the picking mechanism gears with respect to the operating speed. The desired lifespan of the transplanter was 255 h to meet the real field service life requirement. Predicted fatigue life range of the picking mechanism gears was recorded as from 436.65 to 4635.97 h, making it higher (by approximately 2 to 18 times) than the lifespan of the transplanter. According to the analyses, the “Steel Composite Material 420H carbon steel” material with a 5 mm face width gear was suitable to operate the picking device for a 10-year transplanter service life. The analysis of stress and fatigue presented in this study will guide the design of picking device gears with effective material properties to maintain the recommended service life of the pepper transplanter.


  • Open Access

    Working Speed Analysis of the Gear-Driven Dibbling Mechanism of a 2.6 kW Walking-Type Automatic Pepper Transplanter

    by Zafar Iqbal, Nafiul Islam, Milon Chowdhury, Sumaiya Islam, Tusan Park, Yong-Joo Kim, Sun-Ok Chung

    The development of an automatic walking-type pepper transplanter could be effective in improving the mechanization rate in pepper cultivation, where the dibbling mechanism plays a vital role and determines planting performance and efficiency. The objective of this research was to determine a suitable working speed for a gear-driven dibbling mechanism appropriate for a pepper transplanter, while considering agronomic transplanting requirements. The proposed dibbling mechanism consisted of two dibbling hoppers that simultaneously collected free-falling seedlings from the supply mechanism and dibbled them into soil. To enable the smooth collection and plantation of pepper seedlings, analysis was carried out via a mathematical working trajectory model of the dibbling mechanism, virtual prototype simulation, and validation tests, using a physical prototype. In the mathematical model analysis and simulation, a 300 mm/s forward speed of the transplanter and a 60 rpm rotational speed of the dibbling mechanism were preferable in terms of seedling uprightness and low mulch film damage. During the field test, transplanting was conducted at a 40 mm planting depth, using different forward speed levels. Seedlings were freely supplied to the hopper from a distance of 80 mm, and the success rate for deposition was 96.79%. A forward speed of 300 mm/s with transplanting speed of 120 seedlings/min was preferable in terms of achieving a high degree of seedling uprightness (90 ± 3.26), a low rate of misplanting (8.19%), a low damage area on mulch film (2341.95 ± 2.89 mm2), high uniformity of planting depth (39.74 ± 0.48 mm), and low power consumption (40.91 ± 0.97 W).