Wei Xing Zhu

Individual Identification Method for Group-Housed Pigs Based on Optimal Feature Extraction

In order to reduce the workload of farmers on individual identification of pigs, reduce interference to pigs, and raise automation and intelligent level of farms, a Matlab-based automatic classification system is set up. Firstly, top view video sequences are captured for group-housed pigs from pig barn, and the adaptive Gaussian mixture model based on prediction mechanism is used to detect the foreground objects. Then with morphological processing and hole filling, the targets are extracted as completely as possible from the background. Secondly, after building the sample image databases for each pig, color, texture and shape features are extracted and combined as the feature vector. Next, LLE (Locally Linear Embedding) algorithm is adopted to reduce the data dimension. Finally, hybrid kernel SVM (Support Vector Machine) classifier is designed to identify the pigs in test frames. Experimental results show that the recognition rate of pigs is 95.2%. For improving the recognition rate, decreasing cost and interference to pigs, the above method is proposed. Moreover, it can lay a foundation for future researches on group-housed pigs in behaviors, emotions, breathings and so on, which has good practical significance.

Monitoring System in Pig Farm Based on Improved Zigbee Tree Routing Algorithm

In view of the fixed testing point and complex wiring in conventional piggery environment control, a new system based on wireless sensor network was designed. This system consisted of the control center with ARM-Linux, executing mechanism nodes and wireless sensor network containing the flexible sensor nodes which could be arbitrarily placed in piggery. In order to make reasonable use of the energy of Zigbee network and prolong the survival time, a improved zigbee tree routing algorithm was proposed. First of all, by analyzing the advantages and disadvantages of the Cluster-tree and AODVjr algorithm in the Zigbee protocol, the neighbor table was introduced into the improved algorithm. Secondly, the scope of the destination node was confirmed to control the radio range of the RREQ and prevented invalid RREQ flooding. Simulation results show that the improved algorithm optimized the overall energy consumption effectively, prolonged the time of the critical nodes, reduced the number of death nodes, balanced the network load and improved the overall performance of the network.

Image Segmentation of Pig Using Wavelet Modulus and Edge Growth

In order to segment the pig image accurately in the complex background, an image segmentation algorithm based on wavelet modulus maxima and edge growth is proposed. Firstly wavelet transformation to pig image was adopted to get local maximum value points in gray gradient, then local adaptive method was used to determine the segmentation threshold within local maximum points, and get the accurate edges. Secondly, the edge growth method based on gradient was used to link discontinuous edge points and complete the edges sealing repair. Finally, the mathematical morphology methods including connecting method and corrosion expansion operation were adopted to obtain the image area of a pig. The experimental results show that, in the complex background of piggery, the smooth and full pig image separated from background could be obtained by this algorithm. This method could improve the effectiveness of the image segmentation.

Gesture Recognition of Pigs Based on Wavelet Moment and Probabilistic Neural Network

For real-time monitoring the behavior of pigs in piggery, the method that combined the advantages of wavelet multi-scale analysis with invariant moments is proposed. Firstly, the original image is pre-processed by using ant colony algorithm to extract object contour. Then the target contour edge growth method and binary morphology are used, and the outlines of pigs are extracted by canny operator. Wavelet moment was used to get the global features of an image and increase the structural details of the image feature description. Finally, the neural network is applied to identify four behaviors including normal walking, walking down, looked up walking and lying of pigs. Experimental results show that the accuracy of the classification and identification of swine gesture reached more than 95%. This method has a better effect in the recognition of pigs and the noise resistance.

Piggery Environmental Control System Based on Wireless Sensor Networks with Energy Optimization

According to the characteristics of piggery, such as complex environment, numerous control objects, complex wiring, a piggery environment control system based on wireless sensor networks was designed. This system consisted of the wireless network composed of various devices and sensors in piggery, the control center with ARM-LINUX and the remote control center. The energy of sensor nodes was offered by batteries in wireless sensor networks so that the power source was limited. In order to strengthen the node energy management, the LEACH algorithm was combined with AODV routing protocol, and a new optimization algorithm called LEACH-A was proposed in this paper. The wireless sensor networks execution process was divided into two phases, including cluster building and data transmission. In the two phases, the node residual energy was calculated by a weighted function and the cluster head role was changed by the size of remaining energy. NS-2 simulation results show that the proposed algorithm prolonged the lifetime of the nodes and the networks, and realized the energy optimization management.

Walking Posture Analysis of Pigs Based on Star Skeleton Model

In order to make better use of video image processing technology to monitor and analyze behaviors of animals, a new walking posture analysis method based on star skeleton is proposed. Firstly, a pig target is detected and its boundary is extracted from the images. Secondly, according to the target’s boundary the center of mass is computed and the curve of distance between contour points and center of mass is draw. The extremums of the above curve are computed to obtain the key contour points from the curve. The stick model of pigs is built by connecting the key contour points with its center of mass. Finally, the gaits of pigs are recognized by analyzing cyclic motion of the key contour points in the stick model. More specifically, the foreleg walking frequency of pigs could be calculated using the spectrum analysis method. The experiments results indicate that this method has higher recognition accuracy for the walking gaits of pigs. The relative error of the results is only within 10%. Furthermore, compared with manual monitoring method, it is more objective and convenient. Thus, this algorithm provides a new solution for animal behavior analysis based on video surveillance and image processing.

Respiratory Frequency Detection of Pigs Based on Area Feature Operator

In order to detect and identify the abnormal respiratory of pig and provide real-time warning, in this paper we propose to use machine vision that apply the feature of area operator to detect the porcine respiratory frequency .The structure of the paper is as follows .Firstly the videos of pig standing in the piggery are captured through the use of camera and transmitted to the computer. Then on the MATLAB platform, the changing abdominal area of pig is extracted. Secondly the characteristics of adaptive and multi-resolution analysis of wavelet analysis allow us to remove the burrs of the area signal. Moreover, we use the peak point detection algorithm to acquire pig’s respiratory frequency during the monitoring time, which finally is transformed into breath rate. The experimental results show that the detection accuracy of respiratory frequency is higher than 92% for the abnormal breathing pigs.

Gait Abnormality Detected of Pigs Based on Machine Vision

In order to detect and identify the abnormal behavior of pigs and provide real-time warning, the detection method of pig’s gait abnormalities using the gait information and machine vision technology is proposed. First of all, the single frame image was extracted from the objective videos, which was preprocessed to obtain contour of pigs. Secondly, the distance curve of center – edge point of the contour was drawn, hence, seven key contour points, ears, nose, limbs, and tail, were selected from the curve. So the star skeleton model could be established according to the key contour points of pigs. Thirdly, by taking advantage of three key contour points to get a angle. In this paper, to choosing ear, centroid and tail of pig to structure a angle, the data about angle changing of pigs both lame and normal walk gaits are calculated from the model respectively, which can be used to detect lame walk. This presented study provides a new method for real-time surveillance in animal behaviors.

Design of Precision Breeding System for Pigs Based on DDE

The pig farming industry in China has realized a rapid growth in recent years, but the breeding facilities and technologies are still relatively backward compared to those in developed countries. The intelligent precision breeding system proposed in this study are used to identify commercial pigs , collect the data of feed intake and weight, obtain feed conversion rate and increase efficiency of farming enterprises. Besides, the system is designed with an electronic pounds module and radio frequency identification (RFID) module. The software of King View with strong function is chosen to practice managing the collected data and monitoring in time in the upper computer monitoring system. Because there is no driver for RFID module in King View, the study has focused on the dynamic data exchange (DDE) technology, which is applied to accomplish the communication between RFID site and the data of King View. The practical application shows that the system operates steadily and the application scope of King View is extended.

Synchronous Control System for Envelope Machine Using DSP

In order to meet the requirements of paper envelope production automation, a synchronous control method for multiple AC servo motors is proposed and the four AC servo motors synchronous control system for paper envelope production base on a DSP is designed. In this system, DSP works as core of control, and four HCTL_2020 counters are used as decoding circuit to realize signal feedback. Feedback signals are the output signal of optical encoder on the AC servo motors, which connect to the corresponding decoder respectively. The output signals of four decoders connect to the GPIO ports of DSP in parallel. Real-time speed and position information of motors are obtained by computing unit of DSP. The PID control algorithm is designed to realize synchronous precision of four motors in speed and position. The field experiment results show that the control method is practical and the system has good stability, accuracy and efficiency in paper envelope production.