The Fast Detection Method Of Pseudosciaena Crocea Morphological Parameters Based On Machine Vision

As essential and basic data, parameters of morphological characteristics about Pseudosciaena Crocea provide direct reference data for quantitative analyses of fish farming, breeding and characters classification. At present, the traditional measurement methods are mainly manual work. Based on the related principle of biological statistics, manual operation uses a ruler, vernier caliper and other tools to measure and record the morphological parameters of the fish. This method of data collecting often takes a long time, and each single time, it can't turn out large amounts of data needed. In addition, the preciseness of the data between different batches are difficult to meet consistently. During the breeding period, morphologically characteristic data requires multiple batches and continuous collection. But such manual method gets few and discrete data, which are unable to meet the biologically statistical demand to analyze morphological characteristics of Pseudosciaena Crocea. In the selection of broodstock breeding, Pseudosciaena Crocea's overall shape, color, quality and like can only be assessed by the sensory subjective experience, which is labor intensive, with low efficiency, and is unable to meet the efficiency of scientific farming and breeding requirements.Machine vision technology can realize the non-contact, accurate as well as rapid measurement of the subject. Besides, the technology is mature and widely used. Using machine vision technology to measure morphological characteristic parameters on Pseudosciaena Crocea can effectively avoid error happening in the traditional manual measurement. Meanwhile, the efficiency can be improved substantially. Based on OpenCV, this paper aims to combine machine vision technology with weighing sensor measurement technology, so as to design a detection device which can achieve the synchronical measurement about different indicators and parameters, such as length, weight, shape size. Hence to measure the morphological characteristic parameters becomes a rapid process, and the data and images can be processed correspondingly. After many test analyses, the operation of detecting device designed in this paper is simple, and can effectively improve the efficiency. The precision of data is high, such as, the relative error of weigh data is as low as 0.7%, and for the characteristic parameters data, the average relative error is only 0.28%, which can fully meet the requirements of precise measurement. Moreover, the device can be disassembled and assembled easily, making it portable and practical. It indeed provides a new effective way to study the morphological characteristics of the fish, and to construct the basic database of the growth cycle of the fish, as well as to improve the statistical classification method so as to evaluate the resource accurately.