Real-time and accurate channel information is essential for improving the spectral efficiency and reliability of communication systems. In mobile communication environments, especially in Internet of Vehicles (IoV) scenarios, both mobile base stations (mBSs) and user equipment (UE) are continuously moving. This mobility causes the wireless channel to change rapidly. As a result, the channel feedback received from the UE is often outdated and cannot accurately reflect current conditions. In the absence of real-time feedback, the mBS must determine the coding rate prior to transmission and cannot rely on confirmation from the UE. To address this issue, we propose a Heterogeneous Environment-Aware Link Adaptation (HEALA) framework for IoV that operates without feedback. The framework utilizes data from event cameras and RGB cameras to mitigate the spectral-efficiency degradation caused by stale or inaccurate channel feedback. Considering the limited computational and storage resources of mBSs in IoV systems, we design a lightweight deep estimation algorithm. The algorithm fuses multi-source heterogeneous data with spatio-temporal dependencies to estimate the real-time distance between mBS and UE. Based on this estimation, the coding rate is adjusted to better match the actual channel conditions. Experimental results show that the proposed framework achieves stable convergence and low computational overhead. HEALA improves link adaptation performance by 64.1% and 17.5% compared to traditional methods and the latest deep reinforcement learning-based methods, respectively. It also yields more stable coding-rate decisions under rapidly varying channels. In addition to practical benefits, this work demonstrates the potential of environmental information as a valuable external factor for enhancing communication system performance.