基于YOLO7目标检测的Planogram落地实现技术咨询

我正在开展一项面向商超的个人项目，采用YOLO7进行目标检测，核心功能之一是实现Planogram（货架图）对比。目前已完成货架及商品的检测与分类，输出的分析数据如下：

analysis = {
    'Shelves': 4,
    'Shelves data': {1: [439.0, 533.0], 2: [315.0, 435.0], 3: [141.0, 322.0], 4: [0.0, 138.0]},
    'Products': {'Product1': [2, 1, 1, 2, 4, 1, 4, 2, 2, 2, 2, 3, 4], 'Product2': [2, 2]},
    'Product count': {'Product1': 13, 'Product2': 2, 'Total': 15},
    'Products per shelf': {2: {'Product1': 6, 'Product2': 2}, 1: {'Product1': 3}, 4: {'Product1': 3}, 3: {'Product1': 1}},
    'Percentage': {
        1: {'Product1': '100.00', 'Total percentage in relation to all detected products': '86.67', 'Total percentage of the same product ': '23.08'},
        2: {'Product1': '75.00', 'Total percentage in relation to all detected products': '86.67', 'Total percentage of the same product ': '46.15', 'Product2': ('25.00', {'Total percentage in relation to all detected products': '13.33', 'Total percentage of the same product ': '100.00'})},
        3: {'Product1': '100.00', 'Total percentage in relation to all detected products': '86.67', 'Total percentage of the same product ': '7.69'},
        4: {'Product1': '100.00', 'Total percentage in relation to all detected products': '86.67', 'Total percentage of the same product ': '23.08'}
    },
    'Versus Competition': {2: [{'Product2': {'Product1': '7'}}, False]}
}

我曾考虑通过处理货架图图片并对比.txt结果来实现，但不确定这是否为最优方案，恳请专业指导如何落地Planogram功能。

内容的提问来源于stack exchange，提问作者MoroccanTea