A processing method for multi-claw transplanting and sorting of plug seedlings based on depth camera

Abstract

The invention discloses a processing method for multi-claw transplanting-sorting of plug seedlings based on a depth camera, and relates to the field of agricultural robots. A fast method for removing damage seedlings and pointer interference information in selection operations. For the automatic transplanting-sorting operation in industrial seedling cultivation, the vertical and horizontal viewing angles are effectively compressed by the depth camera, and the depth images of single seedlings in the tray seedlings are transported in rows in order to be acquired frame by frame. The single seedling-tub depth image is extracted from the multi-sapling-tub continuous depth data stream, and then the three-dimensional depth template is used for rapid matching processing to realize the rapid elimination of pointer interference information and the efficient detection of seedling damage, so as to effectively meet the needs of plate replacement and transplanting. Continuous non-pause detection of multi-claw seedlings is required.

Description

technical field [0001] The invention relates to the field of agricultural robots, in particular to a depth camera-based multi-claw transplanting-sorting processing method for plug seedlings. Background technique [0002] In the transplanting of seedlings in plug trays, we often face the situation that the seedlings in the seedling trays grow unevenly, and the lack of seedlings and unhealthy seedlings generally occur. The lack of seedlings will lead to "catch empty" and "missed planting", while the lack of seedlings, unhealthy seedlings and damage during transplanting will lead to uneven emergence and poor growth of plug seedlings after transplanting, which will affect the yield and cause waste . Existing machine vision-based detection of missing seedlings and unhealthy seedlings is mainly carried out statically in the target tray, which is only suitable for very small seedlings with no cross-border branches and leaves, and the detection effect on the common occurrence of le...

AI Technical Summary

Problems solved by technology

[0002] In the transplanting of plug seedlings, it is often faced with the situation that the seedlings in the seedling plugs grow unevenly, and the lack of seedlings and unhealthy seedlings generally occur. At the same time, the transplanting will cause damage to the seedlings and seedlings. The lack of seedlings will lead to "catch empty" and "missed planting", while the lack of seedlings, unhealthy seedlings and damage during transplanting will lead to uneven emergence and poor growth of plug seedlings after transplanting, which will affect the yield and cause waste

Existing machine vision-based detection of missing seedlings and unhealthy seedlings is mainly carried out statically in the target tray, which is only suitable for very small seedlings with no cross-border branches and leaves, and the detection effect on commonly occurring leaf cross-border is extremely poor. Its application scope and operation The effect is greatly limited

At the same time, it is impossible to feed back the key "not taken out" and "damaged" operation information, which leads to complex systems and seriously affects the efficiency and quality of transplanting operations

Images