Multi-stage model joint detection method for ultra-complex class defects of grading ring of power grid

• Main Authors: LI PENG, ZHOU RUIYE, XI WEI, CHEN JIAJIE, HUANG WENQI, WU YANG, ZHENG HUA
• Format: Patent
• Language: Chinese; English
• Published: 26.04.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIRCHEMICAL OR PHYSICAL PROPERTIES; MEASURING; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS; TESTING

The invention relates to a multi-stage model joint detection method and device for ultra-complex class defects of a power grid grading ring, computer equipment, a storage medium and a computer program product. The method comprises the steps of performing joint reasoning on a first model, a second model and a third model to obtain a first training result, a second training result and a third training result. And determining a first optimized first model based on the first training result and the image tag corresponding to the first training result. And determining a second optimized first model and a second optimized second model based on the second training result and the image label corresponding to the second training result. And determining an optimized first model, an optimized second model and an optimized third model based on the third training results and the image tags corresponding to the third training results. Through multiple iteration training, the first target model, the second target model and