Rapid and automated identification of blight disease in potato will help farmers to apply timely remedies to protect their produce. Manual detection of blight disease can be cumbersome and may require trained experts. To overcome these issues, we present an automated system using the Mask Region-based convolutional neural network (Mask R-CNN) architecture, with residual network as the backbone network for detecting blight disease patches on potato leaves in field conditions. The approach uses transfer learning, which can generate good results even with small datasets. The model was trained on a dataset of 1423 images of potato leaves obtained from fields in different geographical locations and at different times of the day. The images were manually annotated to create over 6200 labeled patches covering diseased and healthy portions of the leaf. The Mask R-CNN model was able to correctly differentiate between the diseased patch on the potato leaf and the similar-looking background soil patches, which can confound the outcome of binary classification. To improve the detection performance, the original RGB dataset was then converted to HSL, HSV, LAB, XYZ, and YCrCb color spaces. A separate model was created for each color space and tested on 417 field-based test images. This yielded 81.4% mean average precision on the LAB model and 56.9% mean average recall on the HSL model, slightly outperforming the original RGB color space model. Manual analysis of the detection performance indicates an overall precision of 98% on leaf images in a field environment containing complex backgrounds.

No abstract

No abstract

A robot has been designed and fabricated which can autonomously navigate to desired direction in unknown environment; which can scan the environment to generate SONAR map plus vision system for obstacle detection. Robot can be accessed and controlled manually using wireless module for specific situation. In the strategy for robot navigation in unknown environment, the fuzzy-rule-based algorithm has been employed and implemented in controller itself for fast response.Design of the fuzzy behavior controller has been done by multi sensor integration and fuzzy rule implementation. Through a GUI interfaced wireless control panel, collaborative navigation control can be shared between the human and robot at some specific situations. A low-cost and robust platform has been developed for mobile robot for strength and stability. This mobile robot has been tested in unknown indoor environment with different obstacle course. All of obstacle course was finished successfully autonomously within satisfactory limit of 70 seconds.It was observed from correlation values that left control outputs observed from simulated and actual controller are more similar than others. Percentage error analysis for different outputs for simulated and actual controller values has been calculated and fidelity is found to be satisfactory.

No abstract