Joyeta Saha

Md. Mamunur Rashid, Al-Akhir Nayan, Md. Obaidur Rahman et al.

Traditional fish farming faces several challenges, including water pollution, temperature imbalance, feed, space, cost, etc. Biofloc technology in aquaculture transforms the manual into an advanced system that allows the reuse of unused feed by converting them into microbial protein. The objective of the research is to propose an IoT-based solution to aquaculture that increases efficiency and productivity. The article presented a system that collects data using sensors, analyzes them using a machine learning model, generates decisions with the help of Artificial Intelligence (AI), and sends notifications to the user. The proposed system has been implemented and tested to validate and achieve a satisfactory result.

Al-Akhir Nayan, Ahamad Nokib Mozumder, Joyeta Saha et al.

Early detection of fish diseases and identifying the underlying causes are crucial for farmers to take necessary steps to mitigate the potential outbreak and thus to avert financial losses with apparent negative implications to the national economy. Typically, fish diseases are caused by viruses and bacteria; according to biochemical studies, the presence of certain bacteria and viruses may affect the level of pH, DO, BOD, COD, TSS, TDS, EC, PO43-, NO3-N, and NH3-N in water, resulting in the death of fishes. Besides, natural processes, e.g., photosynthesis, respiration, and decomposition, also contribute to the alteration of water quality that adversely affects fish health. Being motivated by the recent successes of machine learning techniques, a state-of-art machine learning algorithm has been adopted in this paper to detect and predict the degradation of water quality timely and accurately. Thus, it helps to take preemptive steps against potential fish diseases. The experimental results show high accuracy in detecting fish diseases specific to water quality based on the algorithm with real datasets.

Al-Akhir Nayan, Joyeta Saha, Ahamad Nokib Mozumder et al.

The aim of this research is to detect small objects with low resolution and noise. The existing real time object detection algorithm is based on the deep neural network of convolution need to perform multilevel convolution and pooling operations on the entire image to extract a deep semantic characteristic of the image. The detection models perform better for large objects. The features of existing models do not fully represent the essential features of small objects after repeated convolution operations. We have introduced a novel real time detection algorithm which employs upsampling and skip connection to extract multiscale features at different convolution levels in a learning task resulting a remarkable performance in detecting small objects. The detection precision of the model is shown to be higher and faster than that of the state-of-the-art models.