Mobile 3D river crab real-time detection method for 2D images

Abstract

The invention discloses a mobile 3D river crab real-time detection method for 2D images, which comprises the following steps of: performing high-angle shooting through a CMOS (Complementary Metal-Oxide-Semiconductor Transistor) camera arranged below an automatic bait casting ship, enabling the camera to move along with the ship, and continuously acquiring underwater two-dimensional RGB (Red GreenBlue) images capable of reflecting the natural living state of river crabs; utilizing an annotation tool and the augmented reality session data to jointly establish a river crab data set with annotation postures and variable shapes; inspired by an anchor-frame-free mechanism, adopting Gaussian center distribution to start from a single RGB image, and after features are extracted by an encoder decoder architecture, predicting a 2D bounding box of an unknown river crab by adopting a shape, detection and regression assisted multi-task joint learning mode; and through a perfect attitude estimationalgorithm EPnP, expanding 2D prediction to a 3D bounding box to estimate the attitude and physical size of the river crab, so as to construct an ultra-light single-order 3D river crab detection model. According to the method, the variable feeding efficiency and the feeding effect of the automatic feeding ship can be improved.

Description

technical field [0001] The invention relates to the fields of machine vision and pattern recognition, in particular to a three-dimensional detection and size measurement method based on two-dimensional machine vision. [0002] technical background [0003] River crabs are the most productive freshwater crabs in my country. They have rich nutritional and medicinal values ​​and bring huge economic benefits to fishermen. However, current river crab farming is faced with the big problem of variable and uniform bait feeding, and its bait utilization rate and feeding efficiency are relatively low. Consider the huge difference in bait required in different ponds due to the different size, density distribution and water environment of river crabs in ponds. However, it is difficult to estimate the size and density distribution of pond crabs without human intervention because crabs are sensitive and move freely in underwater environments where visibility and light cannot be controlled...

AI Technical Summary

Problems solved by technology

Although there are many potential advantages in fish / crab classification and marine life monitoring, such methods are disadvantaged due to their high price, computational complexity and slow operation speed (compared to 2D methods) and the support of additional hardware. Not yet popularized in mobile devices (such as automated bait-casting boats)

Images