Research On The Key Techniques Of Fish Resource Assessment Based On Imaging Sonar

With the implementation of "Yangtze River Protection" policy,the storage of biological resources and the stability of the ecosystem in our inland waters represented by the rivers and lakes in Yangtze River Basin have gradually attracted widespread attention from all walks of life.Fish resources have an important function of measuring whether the ecological environment in the water area is stable and healthy.It is an effective and accurate way to obtain information on the quantity,growth status,and spatial distribution of fish resources without damaging their living conditions and surrounding environment.Acoustic survey of fish resources is widely used in the current ocean and inland waters due to its many advantages,such as no damage to organisms,no damage to the environment,wide survey scope,high efficiency,and low human and material resources.Acoustic survey is inseparable from the support of sonar instruments.As a high-definition multi-beam imaging sonar,the dual-frequency identification sonar DIDSON can be used for the investigation and monitoring of fish resources in inland waters.Based on the above background,this paper focuses on the key technologies involved in the application of imaging sonar DIDSON to the investigation and evaluation of fish resources,mainly focusing on three aspects: acoustic image processing method,sonar data processing model based on Echoview,and acoustic investigation and evaluation of fish resources in Qingcaosha Reservoir.Firstly,the data storage format of the sonar image is introduced,and the construction of the sonar image is completed through coordinate conversion and interpolation.A variety of commonly used image enhancement and denoising methods are used in sonar images,and the processing effects of each method in sonar images are compared through simulation experiments.A target extraction method based on morphological operations and graphic thresholding and a moving target extraction and extraction based on KNN algorithm are proposed.The former has a better target recognition effect in a single sonar image,and the latter extracts moving targets in a sonar video.Finally,the classic Kalman filter algorithm is used to track and count fish.Secondly,fish count alone cannot meet the needs of actual resource assessment.A semi-automatic data processing model for imaging sonar is established in the commonly used acoustic data post-processing system Echoview.This model directly processes the sonar echo information and mainly contains data.Three sections: preprocessing,fish target recognition and information extraction,image stitching and target tracking and counting.In the model,noise removal,trajectory tracking,information export and other parts require certain manual intervention and correction,so as to extract the relevant information needed for resource assessment such as detection depth,route length,number of fish,individual size,spatial location,and echo intensity.The model counting is verified through experiments.The results show that compared with manual counting,the deviation of this method is within 10%,and it has a higher accuracy.Finally,based on the above model,the Qingcaosha Reservoir located at the mouth of the Yangtze River was surveyed and assessed for fish resources in different seasons using imaging sonar combined with net catching.Qingcaosha Reservoir has always maintained water quality with non-classical biological manipulation counts such as farming fish.It is of great significance to understand the changes in fish resources in the reservoir.The assessment results show that the fish are mainly Carassius auratus auratus,Parabramis pekinensis,Hypophthalmichthysmolitrix,and Aristichthys nobilis.The average number of fish in the reservoir in 2019 is about 2.9 million,and the total resource is about 2,400 tons.Taking into account the reservoir fishing and instrument errors,etc.In terms of spatial distribution,fish are mainly distributed in the middle water layer(5-15 m)of the reservoir and the waters on the west side of the lake island,and this distribution is subject to obvious seasonal images..This chapter shows a relatively complete imaging sonar resource assessment process and discusses the factors affecting the assessment,which can provide a reference for subsequent acoustic surveys of fish resources in rivers and lakes in the Yangtze River Basin.