Research And Application Of Acoustic Signal In Baiting System Of Shrimp Pond

As a nutrient rich aquatic product,the demand of Penaeus vannamei is increasing day by day.In order to meet the growth of demand,improve the production of breeding,reduce the cost of breeding,etc.,it has become an urgent problem to be solved.At present,the main way to reduce the feeding cost is to reduce the waste of feed.The observation of feeding status of Penaeus vannamei in China is still at the stage of observing the feed consumption in the culture pond through the hopper.Most of the feeding amount is obtained through the experience of shrimp farmers,and the observation value of the hopper is only a reference value,which has a large error.In this paper,aiming at the problem that it is difficult to observe the feeding condition of the shrimp colony,the system design of using the sound signal to control the baiter in the shrimp pond is realized by combining the underwater acoustic signal acquisition technology,signal analysis technology,neural network algorithm and PLC technology.In the shrimp pond feeding system based on acoustic signal,the acoustic signal of Penaeus vannamei was extracted from the audio by collecting underwater audio,the time-frequency characteristic value of the signal was extracted by continuous wavelet transform and image processing,and the acoustic signal of Penaeus vannamei was classified by principal component method and K-means clustering method.The acoustic signal of Penaeus vannamei was divided into eating signal and non eating signal The feeding quantity is calculated by establishing the model of feeding signal and phonation times,and the feeding quantity is predicted by combining the neural network algorithm.Finally,the communication between MATLAB and PLC controller is established,and PLC controls the feeding machine to complete the feeding.The main contents of this paper are as follows:(1)In order to study the frequency range and pulse duration of the underwater sound of Penaeus vannamei,the experiment of audio acquisition before and after baiting was designed in the culture environment,and the experiment of audio acquisition before and after baiting was designed in the laboratory environment to reduce the noise interference.Select acquisition equipment,establish sound acquisition system,and send acquisition signal to maltab software for analysis.(2)Build a platform in MATLAB to read data and write analysis program.In order to analyze the sound signal of Penaeus vannamei,the audio was preprocessed and the sound signal of Penaeus vannamei was extracted.Through the short-time zero crossing rate and short-time energy method in the signal endpoint detection to find the possible signal frame,and combined with the experimental data without noise interference,the correlation method is used to eliminate the noise frame and screen the effective frame.Secondly,the twodimensional waveform signal is transformed into the time-frequency information by continuous wavelet transform.In order to facilitate observation and analysis,the timefrequency information is drawn.Thirdly,due to the large number of signals and long data,the data matrix after wavelet transform is not conducive to the observation and extraction of eigenvalues,so the method of image recognition is used to extract the eigenvalues of signal time-frequency images in batches.Finally,in order to analyze the changes of signals before and after eating,the signal eigenvalues were analyzed and classified by principal component analysis and K-means cluster analysis,and the signals were divided into eating signals and non eating signals.(3)Because of the feeding signal produced by Penaeus vannamei during feeding,we designed the experiment of the number of times of the feeding signal to obtain the number of times of the feeding sound of one feed,and established the model of the number of times of eating sound and the amount of eating.In order to replace the operation of shrimp farmers to judge the feed consumption of shrimp pond by observing the hopper and to get the feed quantity by using experience,this paper uses BP neural network to predict the next feed quantity,and takes the three parameters of the current feed quantity,the current calculated feed quantity and the feed speed as the input quantity and the next feed quantity as the output quantity to establish BP neural network.Through training network It can predict the amount of baiting and convert it into the time of baiting.(4)At last,the communication between MATLAB and PLC controller is realized by OPC configuration component in MATLAB.The output of Matlab can read and write the data of PLC input directly,and the PLC program can be written to control the relay in the baiter and complete baiting operation.