Multi-view processing method of self-storage water in cattle drinking trough based on artificial intelligence and bim

Abstract

The present invention provides a multi-view processing method based on artificial intelligence and BIM for the self-storage water of cattle drinking troughs. Define a region of interest ROI in , combine the ROI to obtain the position information of the cow, match the key points in the key point heat map, and perform judgment operations to obtain the orientation information of the cow, and determine the key of the channel corresponding to the center point of the cow's hip bone The point heat map is superimposed to obtain the moving track of the cow, and the moving information of the cow is obtained according to the moving track, and a water level sensor is placed in the drinking tank to obtain the water level information; combined with the above information, the drinking tank is self-storing; this method only needs to get the key The above-mentioned information can be obtained by further processing the points, the operation is simple, the calculation speed is fast, and the problem of inaccurate water addition when animals block the drinking trough in the prior art is solved.

Description

Technical field [0001] The present invention relates to artificial intelligence, livestock, and in particular is based on artificial intelligence and bovine BIM water from drinking troughs multi-view processing method. Background technique [0002] Livestock industry has shown a large-scale, intelligent trend, but for a wide range of farming area homes, livestock feeding part is still done by artificial means, for breeding large number of livestock farms, the amount of work required and the method long excessive work. [0003] The method of adding water for drinking groove barn, currently used method information obtained by the water level sensor, the water tank that is located near the cow by the pressure sensor, and then automatically add the cattle can drink water tank. The disadvantage of this method is that the pressure sensor is susceptible to temperature, other factors such as the impact manure, fodder and the like. Drinking water in the tank and positioned too long may ca...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that the pressure sensor is susceptible to temperature, other factors such as manure, forage, etc.

Moreover, if the water in the drinking tank is left for too long, the water quality may be damaged. For example, parasites, dust, etc. are mixed in the drinking water, causing physical discomfort after drinking.

[0004] Based on the problems of the above-mentioned pressure sensor, the existing technology proposes a processing method based on computer vision: image acquisition is performed on the area near the trough, and the amount of food in the trough and the number of animals facing the trough are judged based on the neural network. The amount of food, the number of animals, and the proportion of the area of ​​the trough covered by animals are used as the basis for judging feed delivery. The defect of this method is that it is impossible to judge the amount of food in the trough when the animal covers a large area of ​​the trough. It is difficult to achieve accurate delivery of feed; and this method cannot accurately detect the number of animals facing the trough when the animals block each other, and the inaccurate number of animals obtained will directly lead to inaccurate food delivery. Too much will cause food to deteriorate and cause waste, and too little will affect the growth of animals, and in severe cases, animals will be injured or even killed due to competition for food

Images