diff --git a/README.md b/README.md
index 7a09aaf..dd5011c 100644
--- a/README.md
+++ b/README.md
@@ -667,12 +667,12 @@ Status DBImpl::Delete(const WriteOptions& options, const Slice& key) {
 
 ### **3.不同数据量的情况下索引数据更新/删除对原始数据写入性能影响：**
 
-| ![](.\Report\png\data_a.png) | ![](.\Report\png\data_b.png) | ![](.\Report\png\data_c.png) |
+| ![](./Report/png/data_a.png) | ![](./Report/png/data_b.png) | ![](./Report/png/data_c.png) |
 |------------------------------|------------------------------|------------------------------|
 
-| ![](.\Report\png\average_latency.png) | ![](.\Report\png\index_time.png) |
+| ![](./Report/png/average_latency.png) | ![](./Report/png/index_time.png) |
 |---------------------------------------|----------------------------------|
-| ![](.\Report\png\latency_percentiles.png) | ![](.\Report\png\write_throughput.png) |
+| ![](./Report/png/latency_percentiles.png) | ![](./Report/png/write_throughput.png) |
 
 
 
@@ -759,9 +759,9 @@ Status DBImpl::Delete(const WriteOptions& options, const Slice& key) {
 
 利用多个线程并发写入10^6的数据：
 
-![](.\Report\png\tread_8.png)
+![](./Report/png/tread_8.png)
 
-![](.\Report\png\concurrent_benchmark_results.png)
+![](./Report/png/concurrent_benchmark_results.png)
 
 ### **单线程性能分析**
 
@@ -790,10 +790,10 @@ Status DBImpl::Delete(const WriteOptions& options, const Slice& key) {
 
 ### **5.不同数据量下二级索引查询和直接遍历的性能比较**
 
-| ![](.\Report\png\data_a1.png) | ![](.\Report\png\data_b1.png) | ![](.\Report\png\data_c1.png) |
+| ![](./Report/png/data_a1.png) | ![](./Report/png/data_b1.png) | ![](./Report/png/data_c1.png) |
 |------------------------------|------------------------------|------------------------------|
 
-![](D:\programing\database_sys_building\secondary_index\levelDB_secondary-index\Report\png\query_performance_comparison.png)
+![](./Report/png/query_performance_comparison.png)
 
 从数据中可以观察到二级索引查询和直接遍历在不同数据量下的性能表现差异：