Adaptive three-dimensional reconstruction image acquisition method fusing multi-objective optimization and crop information proportion

Abstract

This invention discloses an adaptive 3D reconstruction image acquisition method that integrates multi-objective optimization and crop information proportion. A multi-factor coupled multi-view image acquisition experiment covering the entire growth period was designed to acquire reconstruction data under different morphological and shooting parameter combinations. An ant colony algorithm-optimized support vector regression model was constructed to predict reconstruction time and peak signal-to-noise ratio. Based on this, a dual-objective optimization model was established. A multi-objective particle swarm optimization algorithm was used to generate a non-dominated solution set, and an approximation-ideal solution sorting algorithm was combined to determine shooting parameters (coarse strategy) such as camera height, angle, and the number of reconstructed images. Furthermore, crop information proportion and a Top-k mechanism were introduced to filter reconstructed images (fine strategy) to construct an adaptive shooting model. This invention can provide image acquisition technology support for high-throughput crop phenotypic analysis, achieving synergistic optimization of 3D reconstruction efficiency and quality.

Images