Automatic weight measurement of pigs based on 3D images and regression network

Weighing pigs with scales takes a lot of time and makes pigs stressful easily. It can be avoided by measuring weight based on the depth images. We design a series of preprocessing algorithms including instance segmentation, distance independence, noise reduction, rotation correction. These algorithms are used to eliminate the influence of the surrounding environment and other factors. In order to predict the weight precisely, we build a regression network based on BotNet. Dual branch of 3 x 3 convolution and MHSA replaces single 3 x 3 convolution of the fourth block in ResNet. At the end of the network, we use multiple fully connection layers in parallel, then concat them to predict weight. After training on 122736 images, MAE of our network is 6.366 on 5326 test images.

Automated summary

Measuring pig weight based on depth images can avoid the time-consuming and stressful process of weighing with scales. By designing preprocessing algorithms and a regression network based on BotNet, we achieve good accuracy in predicting pig weight.