Model-driven data acquisition is one of the strategies utilized to save sensor node energy in Wireless Sensor Networks (WSNs), which suppresses data transmission by running one synchronized prediction model at both the sensor and sink node, and only when the predicted value deviates far from the real value should the sensor node transmit the sensed data to the sink node. In this paper, we propose a novel online model-driven data acquisition method which runs two prediction models on the sensor node simultaneously. Specifically, one model is updated online using stochastic gradient descent (SGD) learning algorithm once a new sensor data is available and the other is used to predict sensor value and updated with the former one when it is time for model re-training. The collaborative working of these two models, together with the SGD learning algorithm, solve two main problems of existing methods: data transmission during off-line model re-training and high resource requirement for model update. Extensive experiments are performed to verify the benefits of our method over two existing methods based on more than 20000 data records from three data sets. The experiment results demonstrate that up to 96% of data transmission are reduced by our method while remaining user defined data accuracy, which outperforms the compared methods in terms of energy usage and data accuracy.

In recent years, the blue-green algae blooms occur frequently in the worldwide lakes and rivers. The harmful algae blooms destroy the aquatic ecosystem balance, and their produced toxins threat humans and animals. In this paper, we propose a new method that can estimate the intensity of the algae blooms and predict their trend in the future based on water quality parameter prediction, which can be used to effectively dispatch the blooms salvage resources. Based on the proposed method, we design, develop and deploy a algae blooms monitoring system on Lake Tai in China. As a under development project, we present the preliminary results of water quality parameter prediction, based on extensive experiments which apply three kinds of time serial prediction methods to the water quality parameter data collected by our deployed system. The experiment results show that the water quality parameters in the future can be accurately predicted with the historical sensor data, which provides the basis for predicting the trend of the intensity of the algae blooms.

Binarization of document images with poor contrast, high noise and variable modalities remains a challenging problem. This paper proposes a new binarization method that adopts the use of seeded region growing and character's topographic feature. It consists of three steps: first, seed pixels are selected automatically according to their topographic features; second, regions are grown controlled by new weighted priority until all pixels are labeled black or white; third, noisy regions are removed based on the average stroke width feature. Our method overcomes the difficulty of global binarization to find a single value to fit all. It also avoids the common problem in most local thresholding technique of finding a suitable window size. The proposed method performed well in binarization and the experimental results of evaluation showed significant improvement compared to several other methods.