Deep learning-based egg group image segmentation fertilization information detection device and method

Abstract

The invention provides a deep learning-based egg group image segmentation fertilization information detection device and method, and the method comprises the following steps: firstly collecting a deep egg group transmission image, segmenting the egg group transmission image into a single target region through a self-designed image segmentation method, building a deep hatching egg fertilization information detection model, and carrying out the recognition of the egg group transmission image. Image texture features are extracted through an optimized MobileNetV1 neural network model, fertilization information in the hatching egg hatching process is detected online, and nondestructive detection of the hatching egg hatching process of the group is achieved; the device is simple to operate, can detect a plurality of hatching eggs at one time, is high in production efficiency, can realize nondestructive detection of hatching egg activity, is suitable for forming large-scale rapid detection and meets the requirements of current automatic production; the establishment method of the model is simple to operate, the model can be used for realizing the nondestructive fertilization detection at the early hatching stage of hatching eggs, the production cost is reduced, and the detection efficiency is improved.

Description

technical field [0001] The invention belongs to the field of detection of fertilized eggs, in particular to a device and method for detecting fertilization information based on deep learning based on group egg image segmentation. Background technique [0002] It takes about 21 days for the eggs to hatch into chicks. In order to ensure the normal development of the chicken embryos, the incubation environment needs to maintain a certain temperature and humidity. Failure to remove infertile eggs or dead embryo eggs in time during the incubation process not only takes up space in the incubator and wastes incubation resources, but also harmful microorganisms such as mold and bacteria multiply rapidly in the incubator, and even infect other normally hatched eggs, resulting in hatching The rate is reduced, resulting in huge economic losses. Therefore, it is of great significance to promptly remove infertile eggs in the early hatching of breeding eggs. [0003] At present, the mos...

AI Technical Summary

Problems solved by technology

[0005] However, spectral equipment is too expensive for low-cost agricultural production, and has high requirements for the detection environment, and the objects are all single eggs, which limits the promotion of these methods; although the above methods can detect the hatching of early eggs, the operation Complicated and costly, it is not suitable for large-scale rapid detection and to meet the needs of current automated production

At present, the hatching quality detection method of breeding eggs is single egg detection, which is low in efficiency and high in cost, and there is little research on the group detection method of hatching quality

Images