一种基于神经网络的多目标检测系统及方法
By using a neural network-based multi-target detection system, which processes images with an industrial camera and a convolutional neural network to calculate placement angles and construct feature matrices, the system solves the problems of state diversification analysis and light dependence in multi-target detection, and achieves high-precision product defect identification and screening.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 南京海汇装备科技有限公司
- Filing Date
- 2025-10-10
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot achieve diversified analysis of products under different states in multi-target detection, and image matching methods are affected by ambient light and cannot accurately handle small targets, making it impossible to determine whether a product can be further processed.
A multi-target detection system based on neural networks is adopted. It acquires images through industrial cameras, calculates the placement angle, processes feature maps using convolutional neural networks, constructs feature matrices and screening prediction models, and realizes intelligent screening of products.
It enables diversified analysis of products in different states, improves detection accuracy, reduces dependence on ambient light, and can quickly identify product defects and determine their processability.
Smart Images
Figure CN121169903B_ABST