3D reconstruction is the process of creating a computer model of the 3D image of an object from twodimensional images. 3D object reconstruction finds various applications in the real world. Reconstruction allows to gain information about qualitative features of the object, which cannot be deduced from a single plane of sight. 3D reconstruction can be used in art analysis and restoration, phenotype analysis, film, and game graphics, design, and criminal investigations, among others. 3D reconstruction techniques enable users to design product at each stage of its development, eventually reducing the overall production cost by a considerable margin. Such wide range of applications across the end-user industries and the commercial application projections have been the prominent factors prompting vendors to invest in technology developments. The advent of deep learning techniques and the increasing availability of large training data sets is leading to a new generation of methods that can recover the 3D geometry and structure of objects, either from one or several RGB images, without using any complex camera calibration process. The image-based 3D reconstruction is an impor tant problem and a building block to several applications, ranging from autonomous navigation and robotics to graphics and enter tainment. Some of the applications deal with generic objects, while some deal with objects that belong to specific classes, such as the human body parts (e.g. faces and hands), animals, and automobile. The recent advancements in geometrically correct 3D intravascular ultrasound (IVUS) reconstruction enabled the depiction of the true coronary anatomy, with the combination of IVUS data with biplane angiographic images. Additionally, fur ther advancements of the 3D IVUS reconstruction software, in conjunction with the advancement in hardware capabilities, are expected to enable the implementation of real-time 3D IVUS reconstruction within the catheterization laboratories. This is expected to aid the market’s growth.

产业在线数据，2020 年8 月家用空调总销量1055.9 万台，YOY+2.8%，内销735.3 万台，YOY+0.9%；出口320.6 万台，YOY+7.5%。1-8 月累计家用空调销量10101.6 万台，YOY-10.8%，内销5684.4 万台，YOY-17.4%；出口4417.2 万台，YOY-0.6%。

产业在线披露8 月空调产销数据：行业总产量1059 万台，同比+3.32%，总销量1056万台，同比+2.78%，其中内销735 万台，同比+0.86%，出口321 万台，同比+7.49%，期末库存1605 万台，同比+36.86%；具体到品牌来看：格力当月+5.26%（内销+6.45%，出口基本持平）；美的当月+15.79%（内销+10.58%，出口+29.87%）；海尔当月+8.03%（内销+6.56%，出口+20.00%）；奥克斯当月-58.86%（内销-59.82%，出口-56.52%）

投资建议。建议关注：设备厂商：露笑科技、三安光电、晶盛机电；衬底厂商：露笑科技、三安光电、天科合达、山东天岳等；外延厂商：瀚天天成和东莞天域等；器件厂商：三安光电、华润微、斯达半导、扬杰科技等。

2020 年9 月16 日凌晨1 点（北京时间），苹果于线上召开名为“Time fora Recap”的秋季新品发布会，对外发布了新款iPad 平板电脑系列、AppleWatch 智能手表系列以及Apple One 等软件服务升级，并首次发布5nm A14仿生芯片，而在本次发布会上并没有发布iPhone 新产品。

重点公司：苹果产业链重点公司立讯精密、歌尔股份、信维通信、鹏鼎控股、蓝思科技、领益智造、环旭电子。

本周电子指数(0.69%)整体表现强于沪深300 指数(-1.53%)，其中表现较好的为LED 板块。从细分子板块来看，消费电子加权平均涨跌幅为-0.95%，跑赢沪深300 指数0.58 个百分点；被动元件加权平均涨跌幅为2.70%，跑赢沪深300 指数4.23 个百分点；显示面板加权平均涨跌幅为3.30%，跑赢沪深300 指数4.83 个百分点；LED 板块加权平均涨跌幅为6.10%，跑赢沪深300 指数7.62 个百分点；安防板块加权平均涨跌幅为-0.79%，跑赢沪深300指数0.74 个百分点；电子化学品加权平均涨跌幅为1.98%，跑赢沪深300指数3.5 个百分点。

近期产业趋势：2020H1，全球半导体销售额2082亿美元，同比+6.77%，中国半导体销售额714亿美元，同比+5.47%，全球占比35.52%。受疫情影响，产业复苏有所放缓。

投资建议：维持推荐评级。海外疫情对电子板块业绩冲击影响有限，电子行业上市公司2020年第二季度收入和利润增速均有回升，行业上行趋势不变。我们继续看好5G周期和国产替代趋势下电子行业龙头的业绩表现，建议重点关注苹果5G手机、智能可穿戴设备、第三代半导体、液晶面板和Mini LED等细分板块的投资机遇。