Crop yield prediction method and device based on multi-level variable analysis

A crop yield prediction device based on a multi-level variable analysis, applied to a control terminal, the device comprising:

• an image acquisition instruction transmitting processor, configured to transmit an image acquisition instruction to a preset aircraft terminal to make the aircraft terminal control a preset image acquisition device on a preset aircraft to perform image acquisition processing on a specified ground to acquire a ground image, wherein crops are planted on the specified ground in advance;
• a driving instruction transmitting processor, configured to transmit a driving instruction to a first rail vehicle terminal and a second rail vehicle terminal which are preset under the specified ground at the same time to make the first rail vehicle terminal control a first rail vehicle to move forward along a preset first rail and make the second rail vehicle terminal control a second rail vehicle to move forward along a preset second rail, wherein both the first rail vehicle and the second rail vehicle drive toward a first direction, a connecting line between a starting point of the first rail vehicle and a starting point of the second rail vehicle is vertical to the first direction, a distance between the first rail vehicle and the second rail vehicle is inversely proportional to the driving time, the first rail and the second rail are located in the same height position, a plurality of light channels are embedded between the first rail and the second rail, and a light transmittance of the light channels is able to be affected by the external environment to change;
• a light emitting instruction transmitting processor, configured to transmit a light emitting instruction to the first rail vehicle terminal to make the first rail vehicle terminal control a preset light generation device on the first rail vehicle to transmit multiple beams of light to the second rail vehicle respectively through the plurality of light channels, and record initial data when the multiple beams of light are emitted;
• a light receiving instruction transmitting processor, configured to transmit a light receiving instruction to the second rail vehicle terminal to make the second rail vehicle terminal control a preset light sensor on the second rail vehicle to acquire the multiple beams of light which are from the first rail vehicle and pass through the plurality of light channels respectively, and record final data when the multiple beams of light are received;
• a light data acquisition processor, configured to acquire the initial data when the multiple beams of light transmitted by the first rail vehicle terminal and the second rail vehicle terminal respectively are emitted, and the final data when the multiple beams of light are received;
• and
• a predicted yield acquisition processor, configured to input the initial data when the ground image and the multiple beams of light are emitted, and the final data when the multiple beams of light are received into a preset crop yield prediction model for processing so as to acquire a predicted yield output by the crop yield prediction model, wherein the crop yield prediction model is trained by training data based on a neural network model, the training data being composed of a ground picture of an area with a known yield, light data emitted by the light generation device on the first rail vehicle in the area and light data received by the light sensor on the second rail vehicle in the area.